U.S. patent application number 16/194956 was filed with the patent office on 2019-07-04 for methods for treating urothelial carcinoma using genotypic and phenotypic biomarkers.
This patent application is currently assigned to Pacific Edge Limited. The applicant listed for this patent is Pacific Edge Limited. Invention is credited to Mark Dalphin, David Darling, Laimonis Kavalieris, Satish Kumar, Paul O'Sullivan, James Miller Suttie.
Application Number | 20190203300 16/194956 |
Document ID | / |
Family ID | 53180146 |
Filed Date | 2019-07-04 |
View All Diagrams
United States Patent
Application |
20190203300 |
Kind Code |
A1 |
Darling; David ; et
al. |
July 4, 2019 |
Methods for Treating Urothelial Carcinoma Using Genotypic and
Phenotypic Biomarkers
Abstract
New methods for treating patents for urothelial cancer (UC)
include combining selected phenotypic variables with levels of
genotypic expression into a metric, the "G+P INDEX." The G+P INDEX
combines age, sex, smoking history, presence of hematuria, and
frequency of hematuria with genotypic expression of the genetic
markers, MDK, CDC2, HOXA13, IGFBP5, and optionally IL8Rb, then
determining of the G+P INDEX value obtained for a patient is within
one of three groups, either: (1) at High Risk of UC, (2) at Risk of
UC, or (3) at Low Risk of UC. For groups 1 and 2, further clinical
and laboratory work up or treatment is indicated, and patients in
group 3 are monitored periodically to determine the need for
further clinical workup. Using the G+P INDEX can save substantial
time, effort, and funds by avoiding unnecessary medical diagnostic
procedures for patients having or are at risk for developing
UC.
Inventors: |
Darling; David; (Dunedin,
NZ) ; Suttie; James Miller; (Dunedin, NZ) ;
Dalphin; Mark; (Dunedin, NZ) ; Kavalieris;
Laimonis; (Dunedin, NZ) ; O'Sullivan; Paul;
(Dunedin, NZ) ; Kumar; Satish; (Havelock North,
NZ) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Pacific Edge Limited |
Dunedin |
|
NZ |
|
|
Assignee: |
Pacific Edge Limited
Dunedin
NZ
|
Family ID: |
53180146 |
Appl. No.: |
16/194956 |
Filed: |
November 19, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
15159359 |
May 19, 2016 |
10131955 |
|
|
16194956 |
|
|
|
|
PCT/US2014/066678 |
Nov 20, 2014 |
|
|
|
15159359 |
|
|
|
|
61907013 |
Nov 21, 2013 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
C12Q 2600/158 20130101;
C12Q 2600/16 20130101; G01N 2800/60 20130101; C12Q 1/6886 20130101;
C12Q 2600/118 20130101; G01N 33/57407 20130101; G01N 33/57484
20130101; G01N 33/50 20130101 |
International
Class: |
C12Q 1/6886 20060101
C12Q001/6886; G01N 33/574 20060101 G01N033/574; G01N 33/50 20060101
G01N033/50 |
Claims
1. A method for treating a patient for urothelial carcinoma
comprising the steps: a) providing a sample of urine from a
patient; b) quantifying a value, M1, comprising detecting and
quantifying the levels of expression of the human genotypic
markers, midkine (MDK) using a forward primer having the sequence
of SEQ ID NO.3, cyclin dependent kinase 1 (CDC2), homeobox A13
(HOXA13), insulin like growth factor binding protein (IGFBP5) in
said sample where MI=[IGFBP5]-[HOXA13]+[MDK]+[CDC2] in said sample,
where the square brackets "[ ]" are defined as the log of
concentrations in the sample of urine of each of said genotypic
markers; c) detecting the log concentration of IL8Rb in said
sample; d) assessing the phenotypic variables: detecting in the
urine of 3 or more red blood cells per high power field in a
6-month period (HFREQ), subject's age greater than 50 years
AgeGT50), gender, smoking history (SMK), and detecting the red
blood cell count (RBC) of said patient and e) quantifying a value
of G+P INDEX according to either: G+P
INDEX=(1*HFREQ+3*Gender+4*SMK)+(5*M1+2*IL8Rb), or formula (i), G+P
INDEX=(w1*HFREQ+w2*AgeGT50+w3*Gender+w4*SMK+w5*RBC)+(w6*M1+w7*IL8Rb),
or formula (ii), G+P INDEX=-8.46+0.79 IGF-1.60 HOXA+2.10 MDK+0.95
CDC-0.38 IL8Rb+0.98 SNS+0.56 Hfreq+1.11 Gender+0.64 Age; where the
terms w1-w7 are respectively the weights assigned to each of the
variables; where formula (iii), HFREQ means the frequency of
finding 3 or more red blood cells per high power field in a 6-month
period; if frequency is low, then HFREQ is set to 0, and if higher
than 3 red blood cells per high power field, then HFREQ is set to
1; AgeGT50 refers to subject's age, if greater than 50 years then
AgeGT50 is set to 1, and if less than 50 years, then AgeGT50 is set
to 0; Gender is assigned a value of 1 for male, and 0 for female;
SMK means whether the subject is a current or ex-smoker; if
non-smoker then SMK is set to 0 and if a smoker, then SMK is set to
1; RBC means red blood cell count: if 25 or more then RBC is set to
1, and if less than 25, then RBS is set to 0; if M1>4.5 then M1
is set to 1, if M1 is less than 4.5, then M1 is set to 0; if
IL8Rb>2.5 then IL8Rb is set to 1, if IL8Rb is less than 2.5,
IL8Rb is set to 0; the symbols "*" means the multiplication
operator, and weighting factors, w1-w7 are respectively the weights
assigned to each of the variables listed in the G+P INDEX; and if
the G+P INDEX has value of from 6-10, said patient undergoes
additional clinical or laboratory tests, cytology, uretoscopy,
and/or CT scan.
2. The method of claim 1, where if the G+P INDEX has a value of
from 11-15, said patient is tested for flexibile cystoscopy,
abdonimal ultrasound, or is treated immediately for urothelial
carcinoma.
3. The method of claim 1, wherein if the G+P INDEX has a value from
0 to 5, said patient receives the normal standard of care and be
placed on a waiting list.
4. The method of claim 1, wherein expression of MDK is determined
using a reverse primer having the sequence of SEQ ID No. 4.
5. The method of claim 1, wherein expression of MDK is determined
using a probe having the sequence of SEQ ID NO.5.
6. The method of claim 1, wherein expression of CDC2 is determined
using a forward primer having the sequence of SEQ ID NO.9.
7. The method of claim 1, wherein expression of CDC2 is determined
using a reverse primer having the sequence of SEQ ID NO.10.
8. The method of claim 1, wherein expression of CDC2 is determined
using a probe having the sequence of SEQ ID NO.11.
9. The method of claim 1, wherein expression of CDC2 is determined
using a forward primer having the sequence of SEQ ID NO.9.
10. The method of claim 1, wherein expression of HOXA13 is
determined using a forward primer having the sequence of SEQ ID
NO.12.
11. The method of claim 1, wherein expression of HOXA13 is
determined using a reverse primer having the sequence of SEQ ID
NO.13.
12. The method of claim 1, wherein expression of HOXA13 is
determined using a probe having the sequence of SEQ ID NO.14.
13. The method of claim 1, wherein expression of IGFBP5 is
determined using a forward primer having the sequence of SEQ ID
NO.6.
14. The method of claim 1, wherein expression of IGFBP5 is
determined using a reverse primer having the sequence of SEQ ID
NO.7.
15. The method of claim 1, wherein expression of IGFBP5 is
determined using a probe having the sequence of SEQ ID NO.8.
16. The method of claim 1, wherein expression of IL8Rb is
determined using a forward primer having the sequence of SEQ ID
NO.15.
17. The method of claim 1, wherein expression of IL8Rb is
determined using a reverse primer having the sequence of SEQ ID
NO.16.
18. The method of claim 1, wherein expression of IL8Rb is
determined using a probe having the sequence of SEQ ID NO.15.
19. The method of claim 1 for treating a patient having an
inflammatory condition of the bladder, comprising the steps: a)
providing a sample of urine from said patient; b) detecting the log
concentration of IL8Rb in said sample using a forward primer having
the sequence of SEQ ID NO.15, wherein if the level of IL8Rb in said
sample is greater than the levels of IL8Rb in a group of patients
not having an inflammatory condition of the bladder, said patient
is treated using an antiinflammatory agent.
20. The method of claim 1, where said human genotypic markers are
measured as mRNA or as cDNA.
Description
CLAIM OF PRIORITY
[0001] This application is a Continuation under 35 U.S.C. 111(a) of
U.S. patent application Ser. No. 16/159,359 filed 19 May 2016, now
entitled "Methods for Detecting Genetic and Phenotypic Biomarkers
of Urothelial Carcinoma and Treatment Thereof" (Now U.S. Pat. No.
10,131,955 issued 20 Nov. 2018), inventors David Darling, Jamese
Miller Suttie, Mark Dalphin, Laimonis Kavalieris, Paul O'Sullivan,
and Satish Kumar, which is a Continuation under 35 U.S.C. 111(a) of
International Patent Application No. PCT/US2014066678 filed 20 Nov.
2014, titled "Triaging Patients Having Asymptomatic Hematuria Using
Genotypic and Phenotypic Biomarkers," which claims priority to U.S.
Provisional Patent Application No. 61/907,013 filed 21 Nov. 2013;
Inventors David Darling, Satish Kumar, Mark Dalphin, and Paul
O'Sullivan. Each of these applications are herein incorporated
fully by reference.
FIELD OF THE INVENTION
[0002] This invention relates to the detection of patients not
having disease. Specifically, this invention relates to the use of
genetic markers and phenotypic markers for triaging patients that
present with hematuria without cancer. Particularly, this invention
relates to analysis of genetic markers and phenotypic markers in
triaging patients with either macroscopic or microscopic hematuria.
More particularly, this invention relates to use of genetic and
phenotypic markers in combination to triage patients with
asymptomatic macroscopic or microscopic hematuria and to predict
whether a patient's condition warrants further clinical
procedures.
BACKGROUND
[0003] Survival of cancer patients is greatly enhanced when the
cancer is treated early. In the case of bladder cancer, patients
diagnosed with disease that is confined to the primary site have a
5 year survival rate of 73%, compared to 6% for patients with
metastatic disease (Altekruse et al). Therefore, developments that
lead to early and accurate diagnosis of bladder cancer can lead to
an improved prognosis for the patients. To aid in early detection
of cancer a number of cancer specific markers have been identified.
However the use of these markers can result in false positive
results in patients having inflammatory bladder diseases, and not
bladder cancer.
[0004] Asymptomatic hematuria ("AH") is one of the most frequent
urological findings, with incidence rates of between 2% and 30%
depending on the population (Schwartz G: Proper evaluation of
asymptomatic microscopic hematuria in the era of evidence-based
medicine-progress is being made. Mayo Clin Proc. 2013, 88(2);
123-125, McDonald M, Swagerty D, Wetzel L: Assessment of
Microscopic hematuria in adults. AFP 2006 73:10, Grossfield G, Wolf
J, Litwan M, Hricak H, Shuler C, Agerter D, et al. Asymptomatic
microscopic hematuria in adults: summary of AUA best practice
policy recommendations. AFP 2001:63:1145-54).
[0005] AH is, however, indicative of broad range of pathologies
with urinary tract malignancy incidences in the AH population
ranging from 1.9-7%. Full diagnostic work up on all confirmed AH
patients puts a considerable burden on many healthcare systems. Use
of phenotypic indicators to segregate high and low risk patients
has been explored in a recent study by Loo et al. (Loo R, Lieberman
S, Slezak J, Landa H, Mariani A, Nicolaisen G, Aspera A and
Jaconsen S:
[0006] Stratifying risk of urinary tract malignant tumors in
patients with asymptomatic microscopic hematuria. Mayo Clin Proc.
2013, 88(2); 129-138).
[0007] The above-mentioned study of 4414 patients presenting with
confirmed AH showed that 73% of patients had no cause identified,
while 26% of patients warranted some form of urological work up to
identify the cause. Approximately 2.5% of patients presenting with
AH were diagnosed with urothelial malignancy, with other conditions
such as urinary tract infection (UTI) (2.3%), kidney stones
(16.2%), prostatic bleeding (4%), and contamination (0.4%) making
up the alterative diagnoses (Loo et al., Id.).
SUMMARY
[0008] We have identified a new problem in the field, namely how to
identify patients presenting with hematuria who do not have or are
at low risk for having bladder cancer. This solves the problem that
many patients with hematuria and without bladder cancer may undergo
expensive and invasive further workup when such workups are not
needed. Thus, this invention is useful to exclude individuals from
the hazards and costs associated with full work-up for bladder
cancer when a combination of genetic information and phenotypic
information provides identification of patients that do not have,
or are at low risk of having bladder cancer, and to effectively
triage patients having no cancer, from those having cancerous
conditions, including urothelial carcinomas, transitional cell
carcinoma (TCC) and non-cancerous conditions, including
inflammatory disease. This invention represents a new approach to a
new problem, in that it is unexpectedly useful, not to diagnose
cancer, but rather to diagnose non-cancers. The use of combinations
of genetic and phenotypic criteria provide unexpectedly better
discrimination than either genetic or phenotypic variables alone.
Data were obtained from 541 observations from validated under CLIA
standards and CURT+North Shore product trial using bootstrap
procedures for internal validation. Phenotypic variables included
presence of: (1) smoking history, (2) hematuria, (3) gender, and
(4) age. Genetic variables included analysis of expression of IGF,
HOXA13, MDK, CDC, and IL8R. The genetic+phenotypic model ("G+P")
performed unexpectedly better than either genetic or phenotypic
variables alone.
[0009] Macrohematuria, or finding of visually identified blood in
the urine is a common finding in patients with bladder cancer. For
those patients, it is often standard practice to perform additional
diagnostic procedures to diagnose bladder cancer. However, readily
identifying patients with microhematuria and understanding the
implications of microhematuria in urothelial carcinomas, remained a
problem.
[0010] We herein provide improved methods for determining whether a
patient presenting with either macrohematuria or microhematuria
could avoid invasive and expensive further clinical procedures to
detect urothelial carcinomas (UC) including bladder cancer, if such
patient is at a sufficiently low probability of having bladder
cancer to warrant not carrying out additional procedures.
[0011] Factors attributable to high probability of urothelial
carcinoma (UC) are described. Demographic factors such as gender,
race and age in addition to environmental factors such as smoking
history and occupational exposure to aromatic amines contribute
significantly to the risk of developing UC. Characterization of
patients in healthcare assessment based on these factors is used
routinely on an ad-hoc basis. For example, it is well accepted that
a 60 year old male with smoking history presenting with hematuria
has a higher probability of being positive for UC than a 35 year
old non-smoking female presenting with the same symptoms, however
these differences have not been quantitated to contribute to the
overall probability of the patient having UC. Attribution of
specific weights to various genotypic and phenotypic factors and
combining these with a diagnostic test output can add significantly
to the accuracy of the diagnostic power of non-invasive tests and
provide clinicians with greater certainty in segregating patients
on the basis of their probability of having UC as defined by the
clinical and biomarker test results.
[0012] Although there are methods available to detect the presence
of bladder cancer, there are no reliable and accurate methods to
determine whether a patient does not have, or is at low risk of
having bladder cancer. To address this need, we have developed new
analytical methods for distinguishing between cancerous conditions
from non-cancerous ones in patients presenting with hematuria,
either macrohematuria or microhematuria. In some aspects of this
invention, we combine quantified phenotypic variables and
quantified expression of genetic markers to form a combined
segregation index (the "G+P INDEX") in order to effectively triage
out AH patients with a low probability of having UC from those AH
patients that have high probability of UC. This segregation defines
those that don't need a complete urological workup from those that
do require a complete workup and thereby avoids unnecessary
work-ups on patients of low probability of UC.
[0013] Phenotypic Assays
[0014] Phenotypic variables evaluated in the G+P INDEX include
frequency of hematuria (HFREQ), age, gender, smoking history, and
red blood count (RBC). These terms are defined herein below.
Phenotypic variables are defined herein to include clinical
findings and observations.
[0015] Genotypic Assays
[0016] In general, preferred genotypic assays developed by Pacific
Edge Ltd. include quantification of expression of the genetic
markers CDC2, HOXA13, MDK and IGFBP5 (a "4-marker assay"). In
another preferred assay, the above 4 markers and a fifth marker,
IL8R, is quantified (a "5-marker" or Cxbladder.RTM. assay; a
trademark of Pacific Edge Ltd., Dunedin, New Zealand) (Holyoake A,
O'Sullivan P, Pollock R et al: Development of a multiplex RNA urine
test for the detection and stratification of transitional cell
carcinoma of the bladder. Clin Cancer Res 2008; 14: 742, and
O'Sullivan P, Sharples K, Dalphin M et al: A Multigene Urine Test
for the Detection and Stratification of Bladder Cancer in Patients
Presenting with Hematuria. J Urol 2012, Vol. 188 No 3; 746), and
International Patent Application No. PCT/NZ2011/000238, entitled
"Novel Markers for Detection of Bladder Cancer." Each of these
publications and patent application are herein incorporated fully
by reference as if separately so incorporated.
[0017] In preferred embodiments, a 4-marker assay can be performed
on unfractionated urine using PCR amplification to quantify four
mRNA markers (for CDC2, HOXA13, MDK and IGFBP5), which are
overexpressed in urothelial carcinoma. IL8R is highly overexpressed
in neutrophils and is consequently elevated in non-malignant
inflammatory conditions. Inclusion of this 5th mRNA marker
significantly reduced the risk of false positive detection of
transitional cell carcinoma (TCC). From the patient's perspective,
the test is non-invasive and very simple. A single sample of urine
often mid-stream urine but not exclusively, is taken, and this can
often be done at home without coming into the clinic.
[0018] The Cxbladder.RTM. assay has been shown to be considerably
more sensitive than cytology in patients presenting with
macroscopic hematuria. Most notably, the Cxbladder.RTM. assay
achieved a sensitivity of 100% (at a pre-specified specificity of
85%) for all urothelial carcinomas with a stage greater than Ta,
and 97% for all high-grade tumors. The Cxbladder.RTM. assay
attributes a single value score that combines the quantitative gene
expression of five genes represented in the patients urine. The
score segregates patients into three classes based on the
probability that the patient has a urothelial carcinoma.
[0019] For patients presenting with hematuria, (either
macrohematuria or microhematuria), this invention has been shown to
enhance these genotypic tools (either the 4-marker assay or
Cxbladder.RTM. assay) with the addition of phenotypic variables
collected from the patient over the same time period, and to
combine these into a new tool, an index that can be used to
segregate patients into three defined risk classes relative to the
patient's probability of having urothelial carcinoma ("UC").
Aspects
[0020] Aspects of this invention are illustrated below. It can be
understood that these are not the only aspects or embodiments of
this invention. Persons of ordinary skill can combine one or more
aspects together to produce additional aspects or embodiments.
[0021] One aspect includes a method for determining, in a patient
presenting with hematuria, or the level of risk for having
urothelial cancer, comprising:
[0022] providing a sample of urine from said patient;
[0023] quantifying a value, MI, comprising quantifying the levels
of expression of human MDK, CDC2, HOXA13, and IGFBP5 in said
sample;
[0024] assessing the phenotypic variables HFREQ, AgeGT, sex, SMK,
and RBC of said patient;
[0025] calculating G+P INDEX according to either:
G+P INDEX=(1*HFREQ+3*Gender+4*SMK)+(5*M1+2*IL-8), or formula
(i),
G+P
INDEX=(w1*HFREQ+w2*AgeGT50+w3*Gender+w4*SMK+w5*RBC)+(w6*M1+w7*IL-8),
or formula (ii),
G+P INDEX=-8.46+0.79 IGF-1.60 HOXA+2.10 MDK+0.95 CDC 0.38 IL8R+0.98
SNS+0.56 Hfreq+1.11 Gender+0.64 Age; and formula (iii),
[0026] determining whether the G+P INDEX is greater than a
threshold indicating the level of risk that the patient has
urothelial cancer.
[0027] Additional aspects include the method of the other aspect,
where said threshold is selected from the group of G+P INDEX values
of from 0 to 5, from 6 to 10, or from 11-15, where said value of
from 0 to 5 indicates Low Risk, 6 to 10 indicates Moderate Risk,
and 11-15 indicates High Risk.
[0028] Further aspects include the method of any other aspect,
where if said threshold is a G+P INDEX value of from 6-10, said
patient undergoes additional clinical or laboratory tests.
[0029] Yet further aspects includc the method of any prior aspect,
where if said threshold is a G+P INDEX value of from 11-15, said
patient undergoes additional clinical or laboratory tests.
[0030] Still further aspects include the method of any other
aspect, where if said threshold is a G+P INDEX value of from 0-5,
the patient is placed on a watch list for further clinical or
laboratory tests.
[0031] Additional aspects include the method of any other aspect,
where the threshold is established using a statistical method.
[0032] Still further aspects include the method of any other
aspect, wherein the statistical method is any one of Linear
Discriminant Analysis (LDA), Logistic Regression (Log Reg), Support
Vector Machine (SVM), K-nearest 5 neighbors (KN5N), and Partition
Tree Classifier (TREE).
[0033] Additional aspects include the method of any other aspect,
further comprising quantifying expression of one additional
genotypic marker selected from FIG. 6 or FIG. 7.
[0034] Other aspects include the method of any previous aspect,
where said step of quantifying genetic expression is carried out by
detecting the levels of mRNA.
[0035] Further aspects include the method of any other aspect,
wherein said step of quantifying genetic expression is carried out
by detecting the levels of cDNA.
[0036] Yet further aspects include the method of any of any other
aspect, where said step of quantifying genetic expression is
carried out using an oligonucleotide complementary to at least a
portion of said cDNA.
[0037] Additional aspects include the method of any other aspect,
where said step of quantifying genetic expression is carried out
using qRT-PCR method using a forward primer and a reverse
primer.
[0038] Yet additional aspects include the method of any other
aspect, where said step of quantifying genetic expression is
carried out by detecting the levels of a protein.
[0039] Still other aspects include the method of any other aspect,
where said step of quantifying genetic expression is carried out by
detecting the levels of a peptide.
[0040] Additional aspects include the method of any of any other
aspect, where said step of quantifying genetic expression is
carried out using an antibody directed against said marker.
[0041] Yet further additional aspects include the method of any of
claim 1 to 8 or 13-15, where said step of quantifying genetic
expression is carried out using a sandwich-type immunoassay method,
or using an antibody chip.
[0042] Still further aspects include the method of any other
aspect, where said quantifying genetic expression is carried out
using a monoclonal antibody.
[0043] Other aspects include the method any of any other aspect,
where said quantifying genetic expression is carried out using a
polyclonal antiserum.
G+P INDEX
[0044] Phenotypic and genotypic variables described above are
combined into a G+P INDEX according to the following
relationship:
G+P
INDEX=(w1*HFREQ+w2*AgeGT50+w3*Gender+w4*SMK+w5*RBC)+(w6*M1+w7*IL-8),
where HFREQ means the frequency of finding 3 or more red blood
cells per high power field in a 6-month period; if frequency is low
then HFREQ=0, and if higher than 3 red blood cells per high power
field, then 1. AgeGT50 refers to subject's age, if greater than 50
years then AgeGT50=1, and if less than 50 years, then 0. Gender is
assigned a value of 1 for male, and 0 for female. SMK means whether
the subject is a current or ex-smoker; if non-smoker then SMK=0 and
if a smoker, then 1. RBC means red blood cell count; if 25 or more
then RBC is set to 1, and if less than 25, then 0. MI is a
combination of expression of the genetic markers MDK, CDC, IGFBP5,
and HOXA13; if M1>4.5 then set it to 1, if less than 4.5, 0.
IL-8 refers to expression level of RNA for IL-8; if IL-8>2.5
then IL-8 is set to 1, if less than 2.5, 0. The symbols "*" means
the multiplication operator, and weighting factors, w1-w7 are
respectively the weights assigned to each of the variables listed
above in the G+P INDEX.
[0045] In other preferred embodiments, (AgeGT50 and RBC) may be
dropped from the model as shown below:
G+P INDEX=(1*HFREQ+3*Gender+4*SMK).+-.(5*M1+2*IL-8)
[0046] The G+P INDEX produces a value between 0 and 15. A patient
with G+P INDEX value of 11 to 15 is considered to be at "High Risk"
for bladder cancer, and indicates the need for additional work up
for bladder cancer. A patient with a G+P INDEX value of 6 to 10 is
considered to be at "Moderate Risk" for developing bladder cancer,
and additional work up is indicated. A patient with a G+P INDEX of
0 to 5 is considered to be at "Low Risk" for developing bladder
cancer. Patients in the "Low Risk" group are placed on a watchful
waiting list, and if additional symptoms appear, or if recurrent
episodes of microhematuria occur, they are reevaluated for possible
further work up.
[0047] As described more fully in Example 3 (FIGS. 18 and 19), we
found that the ROC curve for the quantified phenotypic variables
alone produced a modest level of diagnostic power. The ROC curve
for the quantified genotypic markers alone produced a significant
level of diagnostic power. We found an unexpectedly better
diagnostic power when both genotypic and phenotypic variables were
combined into a G+P INDEX.
[0048] Quantification of Genetic Expression
[0049] Proteins or nucleic acids that are secreted by or cleaved
from the cell, or lost by apoptotic mechanisms, either alone or in
combination with each other, have utility as serum or body fluid
markers for the diagnosis of disease, including inflammatory
disease in bladder and/or bladder cancer or as markers for
monitoring the progression of established disease. Detection of
protein and cell markers can be carried out using methods known in
the art, and include the use of RT-PCT, qRT-PCR, monoclonal
antibodies, polyclonal antisera and the like.
[0050] Specifically the present invention provides methods for
triaging patients presenting with hematuria, (either macrohematuria
or microhematuria), comprising: (i) providing a biological sample;
(ii) detecting one or more bladder tumor markers (BTMs) in said
sample. Bladder tumor markers of particular interest include MDK,
CDC2, HOXA13, and IGFBP5 (a "4-marker assay"). Optionally, one can
also detect the levels of human neutrophil marker interleukin 8
receptor B (IL8Rb) in the sample (Cxbladder.RTM. assay). The
presence of cancer can be established by comparing the levels of
the one or more BTMs with the levels in normal patients, patients
having early stage bladder cancer, and/or patients having an
inflammatory disease. For example, the presence of cancer can be
established by comparing the expression of BTMs against a threshold
of expression. The threshold may be in the order of expression that
is at least 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4,
5, 6, 7, 8, 9, or 10, 100, 1000, or up to 10,000 times the level of
expression in a group of patients not having cancer. In other
aspects, a high expression of IL8Rb without altered expression of a
bladder tumor marker can be indicative of an inflammatory disease
rather than cancer.
[0051] The methods of the present invention can be used in
conjunction with any suitable marker for detecting bladder cancer.
Examples of suitable markers for use in the invention are outlined
in FIG. 6 or 7. The present invention includes the use of any one
or more of the markers outlined in FIG. 6 or 7.
[0052] Optionally, in other preferred embodiments, the present
invention can include any combination of IL8Rb with one or more of
the markers MDK, CDC2, HOXA13, and IGFBP5, which can also be in
combination with one or more other marker suitable for detecting
bladder cancer, for example, any one of more of the markers
outlined in FIG. 6 or 7. More specifically, the present invention
includes quantification of expression of any one or more
combination of markers: IL8Rb/MDK, IL8Rb/CDC2, IL8Rb/HOXA13,
IL8Rb/IGFBP5, IL8Rb/MDK/CDC2, IL8Rb/MDK/HOXA13, IL8Rb/MDK/IGFBP5,
IL8Rb/CDC2/HOXA13, IL8Rb/CDC2/IGFBP5, IL8Rb/HOXA13/IGFBP5,
IL8Rb/MDK/CDC2/HOXA13, IL8Rb/MDK/CDC2/IGFBP5,
IL8Rb/CDC2/HOXA13/IGFBP5, and IL8Rb/MDK/CDC2/HOXA13/IGFBP5. These
combinations can optionally include one or more further markers
suitable for detecting bladder cancer, for example any one of more
of the markers outlined in FIG. 6 or 7.
[0053] The present invention also provides for a method for
detecting inflammatory conditions of the bladder, comprising: (i)
providing a biological sample from a patient; and (ii) detecting
the levels of human neutrophil marker interleukin 8 receptor B
(IL8Rb) in said sample. The presence of inflammatory conditions of
the bladder is established by comparing the levels of IL8Rb with
the levels in normal patients, patients having hematuria, and
patients having an inflammatory condition of the bladder. For
example, the presence of an inflammatory condition of the bladder
can be established by comparing the expression of the marker IL8Rb
against a threshold, The threshold may be in the order of
expression that is at least 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8,
1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 100, 1000, or up to 10,000 times
the level of expression in another group of patients.
[0054] Preferred genotypic methods of the present invention can be
carried out by detecting any suitable marker of gene expression,
for example determining the levels of mRNA, cDNA, a protein or
peptide utilizing any suitable method.
[0055] The establishment of a diagnosis can be established through
the use classifier system, for example Linear Discriminant Analysis
(LDA), Logistic Regression (Log Reg), Support Vector Machine (SVM),
K-nearest 5 neighbors (KN5N), and Partition Tree Classifier
(TREE).
BRIEF DESCRIPTION OF THE FIGURES
[0056] This invention is described with reference to specific
embodiments thereof and with reference to the figures (the same as
FIG., Fig., and Figure), in which:
[0057] FIGS. 1A, 1B, and 1C depict protein and cDNA sequences of
IL8Rb (also known as CXCR2).
[0058] FIGS. 2A-2F depict graphs of scatter plots showing the
effect of IL8Rb on the separation of TCC from non-malignant disease
(prostate disease, cystitis, urinary tract infection and
urolithiasis). IL8Rb has been substituted for different bladder
cancer RNA markers in FIGS. 2C and 2F.
[0059] FIG. 2A. MDK/IGFBP5;
[0060] FIG. 2B. MDK/HOXA13;
[0061] FIG. 2C. MDK/IL8Rb;
[0062] FIG. 2D. CDC2/IGFBP5;
[0063] FIG. 2E. CDC2/HOXA13;
[0064] FIG. 2F. CDC2/IL8Rb.
[0065] FIGS. 3A-3B depict ROC curve analysis (sensitivity vs
specificity) showing the effect of including IL8Rb in diagnostic
algorithms derived using linear discriminate analysis (LD) and
linear regression (LR). The ROC curves were derived from patients
with TCC and upper urinary tract cancers (n=61), and the
non-malignant diseases cystitis, urinary tract infection and
urolithiasis (n=61).
[0066] FIG. 3A. LD1 (solid) and LD2 (dashed).
[0067] FIG. 3B. LR1 (solid) and LR2 (dashed). IL8Rb is included in
LD2 and LR2.
[0068] FIGS. 4a and 4b depict extended ROC curve analysis showing
the effect of including IL8Rb in diagnostic algorithms derived
using linear discriminate analysis (LD) and linear regression (LR).
The ROC curves are derived from patients with TCC (n=56) and,
unlike FIG. 3, any non-malignant disease in the cohort (n=386).
[0069] FIG. 4a. LD1 (solid) and LD2 (dashed).
[0070] FIG. 4b. LR1 (dashed) and LR2 (solid). IL8Rb is included in
LD2 and LR2.
[0071] FIG. 5 depicts box plots showing the accumulation of IL8Rb
mRNA in the urine of patients with non-malignant urological
disease. The RNA has been quantified by qRT-PCR using the delta-Ct
method (Holyoake et al, 2008). With this method a lower Ct reflects
higher RNA levels. BPH: benign prostatic hyperplasia; UTI: urinary
tract infection; NS prostate: non-specific prostate diseases; Vasc.
Prostate: vascular prostate; warfarin: hematuria secondary to
warfarin use. The observations in patients with cystitis/UTI are
significantly different (p=0.001) to the other non-malignant
presentations shown.
[0072] FIGS. 6A-6YY depict markers known to be over expressed in
bladder cancer, and are suitable for use in the present
invention.
[0073] FIGS. 7A-7D depict markers known to be under expressed in
bladder cancer, and are suitable for use in the present
invention.
[0074] FIG. 8 depicts a flow chart for the patient recruitment
procedures and numbers for Example 2.
[0075] FIG. 9 depicts baseline clinical and demographic
characteristics of the patients by disease status at 3 months.
[0076] FIG. 10 depicts overall sensitivity and specificity of the
urine tests.
[0077] FIGS. 11A-11B depict various ROC curves;
[0078] FIG. 11A depicts ROC curves for NMP22 ELISA and uRNA-D (test
comprising the four markers MDK+CDC2+IGFBP5+HOXA13); and
[0079] FIG. 11B depicts ROC curve for the five markers MDK, CDC2,
HOXA13, IGFBP5 and IL8Rb.
[0080] FIG. 12 depicts the sensitivity of urine tests by stage,
grade, location of tumour, multiplicity of tumor, hematuria status,
creatinine of urine sample and sex. Tables show numbers and percent
with a positive urine test among those with TCC.
[0081] FIG. 13 depicts specificity of urine tests by diagnosis,
macrohematuria or, creatinine and sex. Tables show number and %
with a negative urine test result among those without TCC.
[0082] FIGS. 14A(I)-14O(V): depict ROC curves for the combinations
of markers:
[0083] FIGS. 14A(I)(-14A(V): MDK,
[0084] FIGS. 14B(I)-14B(V): CDC,
[0085] FIGS. 14C(I)-14C(V): IGFBP5,
[0086] FIGS. 14D(I)-14D(V): HOXA13,
[0087] FIGS. 14E(i)-14E(v): MDK+CDC2,
[0088] FIGS. 14F(i)-14F(v): MDK+IGFBP5,
[0089] FIGS. 14G(i)-14G(v): MDK+HOXA13,
[0090] FIGS. 14H(I)-14H(V): CDC2+IGFBP5,
[0091] FIGS. 14I(I)-14I(V): CDC+HOXA13,
[0092] FIGS. 14J(I)-16J)V): IGF+HOXA13,
[0093] FIGS. 14K(I)-14K(V): MDK+CDC2+IGFBP5,
[0094] FIGS. 14L(I)-14L(V): MDK+CDC2+HOXA13,
[0095] FIGS. 14M(I)-14M(V): MDK+IGFBP5+HOXA13,
[0096] FIGS. 14N(I)-14N(V): CDC2+IGFBP5+HOXA13,
[0097] FIGS. 14O(I)-14O(V): MDK+CDC2+IGFBP5+HOXA13, plus or minus
IL8Rb, using five different classifier models (i) Linear
Discriminant Analysis (LDA), (ii) Logistic Regression (Log Reg),
(iii) Support Vector Machine (SVM), (iv) K-nearest 5 neighbors
(KN5N), and (v) Partition Tree Classifier (TREE).
[0098] FIGS. 15A-15B depict results of sensitivity selectivity
studies.
[0099] FIG. 15A depicts "Area Under the Curve" (AUC) for up to 20%
false positive rate (at 80% specificity) of the ROC curves from
FIG. 14 and
[0100] FIG. 15B shows the difference the AUC resulting from the
inclusion of IL8Rb.
[0101] FIGS. 16a-16e depict graphs of the sensitivity of the
combinations of the four markers MDK, CDC2, IGFBP5, and HOXA13,
plus or minus IL8Rb, using five different classifier models (i)
Linear Discriminant Analysis (LDA), (ii) Logistic Regression (Log
Reg), (iii) Support Vector Machine (SVM), (iv) K-nearest 5
neighbors (KN5N), and (v) Partition Tree Classifier (TREE), at
different set specificities; (a) 80%, (b) 85%, (c) 90%, (d) 95%,
(e) 98%.
[0102] FIGS. 17a-17j depict the gains in sensitivity from adding
IL8Rb at different set specificities FIG. 17a 80%, FIG. 17b 85%,
FIG. 17c 90%, FIG. 17d 95%, FIG. 17e 98%, and the resulting gains
in specificity from adding IL8Rb at different set specificities
FIG. 17f 80%, FIG. 17g 85%, FIG. 17h 90%, FIG. 17i 95%, FIG. 17j
98%.
[0103] FIG. 18 depicts ROC curves for testing of patients having
hematuria studied using either genetic testing alone, phenotype
evaluation alone, and/or both genetic testing and phenotypic
evaluation.
[0104] FIG. 19 depicts a graph of odds ratios (horizontal axis) for
variables gender, smoking history and HFREQNEW of this
invention.
[0105] FIGS. 20A and 20B depict flow charts for standards of
reporting diagnostic accuracy.
[0106] FIG. 20A depicts a flow chart for patients with
macrohaematuria across all three cohorts in this study.
[0107] FIG. 20B depicts flow chart for reporting diagnostic
accuracy in patients with microhematuria included in this
study.
[0108] FIG. 21 depicts ROC curves representing the three
classification models. P INDEX (dotted line), G INDEX (dashed line)
and G+P INDEX (solid line).
[0109] FIG. 22 depicts NPV versus proportion of patients with
haematuria testing negative according each model. P INDEX (dotted
line), G INDEX (dashed line), and G+P INDEX (solid line).
[0110] FIG. 23 depicts a graph of the relationship between detect
results (horizontal axis) versus Triage result (vertical axis).
[0111] FIG. 24 depicts a graph of G2 INDEX (horizontal axis) versus
G1+P INDEX (vertical axis).
DETAILED DESCRIPTION
Definitions
[0112] Before describing the embodiments of the invention in
detail, it will be useful to provide some definitions of terms as
used herein.
[0113] The term "marker" refers to a molecule that is associated
quantitatively or qualitatively with the presence of a biological
phenomenon. Examples of "markers" include a polynucleotide, such as
a gene or gene fragment, whether coding or non-coding, DNA or DNA
fragment RNA or RNA fragment, whether coding or non-coding; or a
gene product, including a polypeptide such as a peptide,
oligopeptide, protein, or protein fragment; or any related
metabolites, by products, or any other identifying molecules, such
as antibodies or antibody fragments, whether related directly or
indirectly to a mechanism underlying the phenomenon. The markers of
the invention include the nucleotide sequences (e.g., GenBank
sequences) as disclosed herein, in particular, the full-length
sequences, any coding sequences, any fragments, any possible probes
(e.g., created across an exon-exon boundary), including those with
capture motifs, hairpins or fluorophores, or any complements
thereof, and any measurable marker thereof as defined above.
[0114] As used herein "antibodies" and like terms refer to
immunoglobulin molecules and immunologically active portions of
immunoglobulin (Ig) molecules, i.e., molecules that contain an
antigen binding site that specifically binds (immunoreacts with) an
antigen. These include, but are not limited to, polyclonal,
monoclonal, chimeric, single chain, Fc, Fab, Fab', and Fab.sub.2
fragments, and a Fab expression library. Antibody molecules relate
to any of the classes IgG, IgM, IgA, IgE, and IgD, which differ
from one another by the nature of heavy chain present in the
molecule. These include subclasses as well, such as IgG1, IgG2, and
others. The light chain may be a kappa chain or a lambda chain.
Reference herein to antibodies includes a reference to all classes,
subclasses, and types. Also included are chimeric antibodies, for
example, monoclonal antibodies or fragments thereof that are
specific to more than one source, e.g., a mouse or human sequence.
Further included are camelid antibodies, shark antibodies or
nanobodies.
[0115] The terms "cancer" and "cancerous" refer to or describe the
physiological condition in mammals that is typically characterized
by abnormal or unregulated cell growth. Cancer and cancer pathology
can be associated, for example, with metastasis, interference with
the normal functioning of neighboring cells, release of cytokines
or other secretory products at abnormal levels, suppression or
aggravation of inflammatory or immunological response, neoplasia,
pre-malignancy, malignancy, invasion of surrounding or distant
tissues or organs, such as lymph nodes, etc.
[0116] The term "tumor" refers to all neoplastic cell growth and
proliferation, whether malignant or benign, and all pre-cancerous
and cancerous cells and tissues.
[0117] The term "bladder cancer" refers to a tumor originating in
the bladder. These tumors are able to metastasize to any organ.
[0118] The term "BTM" or "bladder tumor marker" or "BTM family
member" means a tumor marker (TM) that is associated with
urothelial cancers, bladder cancer, transitional cell carcinoma of
the bladder (TCC), squamous cell carcinomas, and adenocarcinomas of
the bladder. The term BTM also includes combinations of individual
markers, whose combination improves the sensitivity and specificity
of detecting bladder cancer. It is to be understood that the term
BTM does not require that the marker be specific only for bladder
tumors. Rather, expression of BTM can be altered in other types of
cells, diseased cells, tumors, including malignant tumors.
[0119] The term "under expressing BTM" means a marker that shows
lower expression in bladder tumors than in non-malignant bladder
tissue.
[0120] The term "over expressing BTM" means a marker that shows
higher expression in bladder tumors than in non-malignant
tissue.
[0121] The terms "differentially expressed," "differential
expression," and like phrases, refer to a gene marker whose
expression is activated to a higher or lower level in a subject
(e.g., test sample) having a condition, specifically cancer, such
as melanoma, relative to its expression in a control subject (e.g.,
reference sample). The terms also include markers whose expression
is activated to a higher or lower level at different stages of the
same condition; in diseases with a good or poor prognosis; or in
cells with higher or lower levels of proliferation. A
differentially expressed marker may be either activated or
inhibited at the polynucleotide level or polypeptide level, or may
be subject to alternative splicing to result in a different
polypeptide product. Such differences may be evidenced by a change
in mRNA levels, surface expression, secretion or other partitioning
of a polypeptide, for example.
[0122] Differential expression may include a comparison of
expression between two or more markers (e.g., genes or their gene
products); or a comparison of the ratios of the expression between
two or more markers (e.g., genes or their gene products); or a
comparison of two differently processed products (e.g., transcripts
or polypeptides) of the same marker, which differ between normal
subjects and diseased subjects; or between various stages of the
same disease; or between diseases having a good or poor prognosis;
or between cells with higher and lower levels of proliferation; or
between normal tissue and diseased tissue, specifically cancer, or
melanoma. Differential expression includes both quantitative, as
well as qualitative, differences in the temporal or cellular
expression pattern in a gene or its expression products among, for
example, normal and diseased cells, or among cells which have
undergone different disease events or disease stages, or cells with
different levels of proliferation.
[0123] The term "expression" includes production of polynucleotides
and polypeptides, in particular, the production of RNA (e.g., mRNA)
from a gene or portion of a gene, and includes the production of a
polypeptide encoded by an RNA or gene or portion of a gene, and the
appearance of a detectable material associated with expression. For
example, the formation of a complex, for example, from a
polypeptide-polypeptide interaction, polypeptide-nucleotide
interaction, or the like, is included within the scope of the term
"expression". Another example is the binding of a binding ligand,
such as a hybridization probe or antibody, to a gene or other
polynucleotide or oligonucleotide, a polypeptide or a protein
fragment, and the visualization of the binding ligand. Thus, the
intensity of a spot on a microarray, on a hybridization blot such
as a Northern blot, or on an immunoblot such as a Western blot, or
on a bead array, or by PCR analysis, is included within the term
"expression" of the underlying biological molecule.
[0124] The terms "gene expression threshold," and "defined
expression threshold" are used interchangeably and refer to the
level of a marker in question, outside which the expression level
of the polynucleotide or polypeptide serves as a predictive marker
for a condition in the patient. For example, the expression of
IL8Rb above a certain threshold is diagnostic that the patient has
an inflammatory condition. A threshold can also be used when
testing a patient for suspected bladder cancer, using bladder
cancer makers. Expression levels above a threshold indicates that
the patient has an inflammatory bladder condition, likely to cause
a false positive test for cancer, whereas an expression level of
IL8Rb below a threshold is predictive that the patient does not
have an inflammatory bladder condition. By including the
measurement of IL8Rb any result from the expression of the bladder
tumor markers can be relied upon if the levels of IL8Rb is below
the threshold (i.e. a positive result is likely to be positive for
the patient having cancer rather than increased levels of the
bladder tumor markers actually resulting from exfoliation of
non-malignant cells from the mucosa from inflammation).
[0125] The term "diagnostic threshold" refers to a threshold in
which a patient can be said to have been diagnosed either with or
without a given condition, for example bladder cancer. A diagnostic
threshold is generally set to achieve a desired sensitivity and
specificity, depending on factors such as population, prevalence,
and likely clinical outcome. In general the diagnostic threshold
can be calculated and/or established using algorithms, and/or
computerized data analysis.
[0126] The exact threshold will be dependent on the population and
also any model being used to predict disease (predictive model). A
threshold is established experimentally from clinical studies such
as those described in the Examples below. Depending on the
prediction model used, the expression threshold may be set to
achieve maximum sensitivity, or for maximum specificity, or for
minimum error (maximum classification rate). For example a higher
threshold may be set to achieve minimum errors, but this may result
in a lower sensitivity. Therefore, for any given predictive model,
clinical studies will be used to set an expression threshold that
generally achieves the highest sensitivity while having a minimal
error rate.
[0127] The term "sensitivity" means the proportion of individuals
with the disease who test (by the model) positive. Thus, increased
sensitivity means fewer false negative test results.
[0128] The term "specificity" means the proportion of individuals
without the disease who test (by the model) negative. Thus,
increased specificity means fewer false positive test results.
[0129] The term "Receiver Operating Characteristic" ("ROC curve")
means a plot of the true positive rate (sensitivity) against the
false positive rate (specificity) for different cut off points for
a particular marker or test. Each point on the ROC curve represents
a specific sensitivity/specificity point that will correspond to a
given threshold. ROC curves can be important to establish a
threshold to give a desired outcome. The area under a ROC curve
represents (expressed as an Area Under the Curve (AUC) analysis,
can be a measure of how well a given marker or test consisting of a
number of markers, can distinguish between two or more diagnostic
outcomes. ROC curves can also be used to compare the accuracy of
two different tests.
[0130] The term "oligonucleotide" refers to a polynucleotide,
typically a probe or primer, including, without limitation,
single-stranded deoxyribonucleotides, single- or double-stranded
ribonucleotides, RNA: DNA hybrids, and double-stranded DNAs.
Oligonucleotides, such as single-stranded DNA probe
oligonucleotides, are often synthesized by chemical methods, for
example using automated oligonucleotide synthesizers that are
commercially available, or by a variety of other methods, including
in vitro expression systems, recombinant techniques, and expression
in cells and organisms.
[0131] The term "overexpression" or "overexpressed" refers to an
expression level of a gene or marker in a patient that is above
that seen in normal tissue. Expression may be considered to be
overexpressed if it is 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9,
2, 10, 100, 1000, or up to 10,000 times the expression in normal
tissue or in tissues from another group of patients.
[0132] The term "polynucleotide," when used in the singular or
plural, generally refers to any polyribonucleotide or
polydeoxribonucleotide, which may be unmodified RNA or DNA or
modified RNA or DNA. This includes, without limitation, single- and
double-stranded DNA, DNA including single- and double-stranded
regions, single- and double-stranded RNA, and RNA including single-
and double-stranded regions, hybrid molecules comprising DNA and
RNA that may be single-stranded or, more typically, double-stranded
or include single- and double-stranded regions. Also included are
triple-stranded regions comprising RNA or DNA or both RNA and DNA.
Specifically included are mRNAs, cDNAs, and genomic DNAs, and any
fragments thereof. The term includes DNAs and RNAs that contain one
or more modified bases, such as tritiated bases, or unusual bases,
such as inosine. The polynucleotides of the invention can encompass
coding or non-coding sequences, or sense or antisense sequences. It
will be understood that each reference to a "polynucleotide" or
like term, herein, will include the full-length sequences as well
as any fragments, derivatives, or variants thereof.
[0133] The term "phenotypic," means a trait that is observable in a
clinical setting, or in a clinical interview, or in a patient's
history. When used in a formula for calculating G+P index,
"phenotypic" or "P" means the patient's age, sex, incidence of
hematuria, and smoking history.
[0134] "Polypeptide," as used herein, refers to an oligopeptide,
peptide, or protein sequence, or fragment thereof, and to naturally
occurring, recombinant, synthetic, or semi-synthetic molecules.
Where "polypeptide" is recited herein to refer to an amino acid
sequence of a naturally occurring protein molecule, "polypeptide"
and like terms, are not meant to limit the amino acid sequence to
the complete, native amino acid sequence for the full-length
molecule. It will be understood that each reference to a
"polypeptide" or like term, herein, will include the full-length
sequence, as well as any fragments, derivatives, or variants
thereof.
[0135] The term "qPCR" or "QPCR" refers to quantitative polymerase
chain reaction as described, for example, in PCR Technique:
Quantitative PCR, J. W. Larrick, ed., Eaton Publishing, 1997, and
A-Z of Quantitative PCR, S. Bustin, ed., IUL Press, 2004.
[0136] The term "Reverse Transcription" means a process in which an
oligoribonucleotide, including a messenger RNA ("mRNA") is used as
a template for biochemical synthesis of a complementary
oligodeoxyribonucleotide ("cDNA") using an enzyme ("Reverse
Transcriptase"), which binds to the template RNA, and catalyzes a
series of addition reactions that sequentially attaches
deoxyribonucleotide bases to form an oligodeoxyribonucleotide
strand that is complementary to the RNA template.
[0137] The term "Hematuria" is defined as the presence of blood in
the urine. It may present as macroscopic hematura (visible traces
of blood cells) or microscopic hematuria (microscopic traces of
blood) within the urine. A confirmed indication of microhematuria
is defined as 3 or more red blood cells present per microscopic
high-powered field (HPF) on a minimum of 3 properly collected urine
samples. Microhematuria may also be detected by urine dipstick
(colorimetric comparison estimate) at clinic. Hematuria (either
microscopic or macroscopic) may be asymptomatic (no additional
symptoms associated with hematuria) or symptomatic. Additional
symptoms include dysuria (painful urination), a feeling of
incomplete emptying of the bladder or increased frequency or
urination.
[0138] "Stringency" of hybridization reactions is readily
determinable by one of ordinary skill in the art, and generally is
an empirical calculation dependent upon probe length, washing
temperature, and salt concentration. In general, longer probes
require higher temperatures for proper annealing, while shorter
probes need lower temperatures. Hybridization generally depends on
the ability of denatured DNA to re-anneal when complementary
strands are present in an environment below their melting
temperature. The higher the degree of desired homology between the
probe and hybridizable sequence, the higher the relative
temperature which can be used. As a result, it follows that higher
relative temperatures would tend to make the reaction conditions
more stringent, while lower temperatures less so. Additional
details and explanation of stringency of hybridization reactions,
are found e.g., in Ausubel et al., Current Protocols in Molecular
Biology, Wiley Interscience Publishers, (1995).
[0139] "Stringent conditions" or "high stringency conditions", as
defined herein, typically: (1) employ low ionic strength and high
temperature for washing. For example 0.015 M sodium chloride/0.0015
M sodium citrate/0.1% sodium dodecyl sulfate at 50.degree. C.; (2)
employ a denaturing agent during hybridization, such as formamide,
for example, 50% (v/v) formamide with 0.1% bovine serum
albumin/0.1% Fico11/0.1% polyvinylpyrrolidone/50 mM sodium
phosphate, buffer at pH 6.5 with 750 mM sodium chloride, 75 mM
sodium citrate at 42.degree. C.; or (3) employ 50% formamide,
5.times.SSC (0.75 M NaCl, 0.075 M sodium citrate), 50 mM sodium
phosphate (pH 6.8), 0.1% sodium pyrophosphate, 5.times., Denhardt's
solution, sonicated salmon sperm DNA (50 ug/ml), 0.1% SDS, and 10%
dextran sulfate at 42.degree. C., with washes at 42.degree. C. in
0.2.times.SSC (sodium chloride/sodium citrate) and 50% formamide at
55.degree. C., followed by a high-stringency wash comprising
0.1.times.SSC containing EDTA at 55.degree. C.
[0140] "Moderately stringent conditions" may be identified as
described by Sambrook et al., Molecular Cloning: A Laboratory
Manual, New York: Cold Spring Harbor Press, 1989, and include the
use of washing solution and hybridization conditions (e. g.,
temperature, ionic strength, and % SDS) less stringent that those
described above. An example of moderately stringent conditions is
overnight incubation at 37.degree. C. in a solution comprising: 20%
formamide, 5.times.SSC (150 mM NaCl, 15 mM trisodium citrate), 50
mM sodium phosphate (pH 7.6), 5.times.Denhardt's solution, 10%
dextran sulfate, and 20 mg/ml denatured sheared salmon sperm DNA,
followed by washing the filters in 1.times.SSC at about
37-50.degree. C. The skilled artisan will recognize how to adjust
the temperature, ionic strength, etc. as necessary to accommodate
factors such as probe length and the like.
[0141] The term "IL8Rb" means neutrophil marker interleukin 8
receptor B (also known as chemokine (C--X--C motif) receptor 2
[CXCR2]) (FIG. 1; SEQ ID NOs. 1 and 2), and includes the marker
IL8Rb. The term includes a polynucleotide, such as a gene or gene
fragment, RNA or RNA fragment; or a gene product, including a
polypeptide such as a peptide, oligopeptide, protein, or protein
fragment; or any related metabolites, by products, or any other
identifying molecules, such as antibodies or antibody
fragments.
[0142] The term "reliability" includes the low incidence of false
positives and/or false negatives. Thus, with higher reliability of
a marker, fewer false positives and/or false negatives are
associated with diagnoses made using that marker.
[0143] "Accuracy" is the proportion of true results (true positives
plus true negatives) divided by the number of total cases in the
population, according to the formula:
True Positives + True Negatives Total Number of Measurements
##EQU00001##
[0144] The term "triage" means to differentiate patients with
hematuria that have a low probability of having bladder cancer from
those patients with hematuria that have a reasonable probability of
having bladder cancer and requiring further clinical work up,
including cystoscopy or other clinical procedure.
Embodiments
[0145] Therefore, in certain preferred embodiments, a combination
of genetic markers and phenotypic markers are provided that permit
differentiating patients having a low probability of having bladder
cancer from those patients with a sufficient risk of having bladder
to warrant further clinical work up, possibly including cystoscopy
or other procedures. In other embodiments, markers are provided
that have reliability greater than about 70%; in other embodiments,
greater than about 73%, in still other embodiments, greater than
about 80%, in yet further embodiments, greater than about 90%, in
still others, greater than about 95%, in yet further embodiments
greater than about 98%, and in certain embodiments, about 100%
reliability.
[0146] For genetic analysis, the practice of the present invention
will employ, unless otherwise indicated, conventional techniques of
molecular biology (including recombinant techniques), microbiology,
cell biology, and biochemistry, which are within the skill of the
art. Such techniques are explained fully in the literature, such
as, Molecular Cloning: A Laboratory Manual, 2nd edition, Sambrook
et al., 1989; Oligonucleotide Synthesis, M J Gait, ed., 1984;
Animal Cell Culture, R. I. Freshney, ed., 1987; Methods in
Enzymology, Academic Press, Inc.; Handbook of Experimental
Immunology, 4th edition, D. M. Weir & CC. Blackwell, eds.,
Blackwell Science Inc., 1987; Gene Transfer Vectors for Mammalian
Cells, J. M. Miller & M. P. Calos, eds., 1987; Current
Protocols in Molecular Biology, F. M. Ausubel et al., eds., 1987;
and PCR: The Polymerase Chain Reaction, Mullis et al., eds.,
1994.
[0147] It is to be understood that the above terms may refer to
protein, DNA sequence and/or RNA sequence. It is also to be
understood that the above terms also refer to non-human proteins,
DNA and/or RNA having homologous sequences as depicted herein.
Embodiments of this Invention
[0148] Often patients referred to a urologist with hematuria are
booked to be seen in a dedicated hematuria clinic. Patients with
macroscopic hematuria are often prioritized over those with micro
hematuria. The information provided at the time of referral from
the primary care provider can be highly variable, making accurate
stratification of patients by probability of having a urothelial
cancer difficult. Often urine cytology is routinely requested
before the patient is seen and, if positive, is used to increase
the patients priority of receiving a full clinical work-up, however
urine cytology, whilst highly specific, has a very low sensitivity
and hence is of little practical value with its high rate of
false-negatives.sup.(6).
Phenotypic and Genotypic Analysis of Patients With Hematuria Not
Having Bladder Cancer
[0149] G+P INDEX
[0150] Phenotypic and genotypic variables described above are
combined into a G+P INDEX according to the following
relationship:
[0151] G+P
INDEX=(w1*HFREQ+w2*AgeGT50+w3*Gender+w4*SMK+w5*RBC)+(w6*M1+w7*I-
L-8), where HFREQ means the frequency of finding 3 or more red
blood cells per high power field in a 6-month period; if frequency
is low then HFREQ=0, and if higher than 3 red blood cells per high
power field, then 1. AgeGT50 refers to subject's age, if greater
than 50 years then AgeGT50=1, and if less than 50 years, then 0.
Gender is assigned a value of 1 for male, and 0 for female. SMK
means whether the subject is a current or ex-smoker; if non-smoker
then SMK=0 and if a smoker, then 1. RBC means red blood cell count;
if 25 or more then RBC is set to 1, and if less than 25, then 0. M1
is a combination of expression of the genetic markers MDK, CDC,
IGFBP5, and HOXA13; if M1>4.5 then set it to 1, if less than
4.5, 0. IL-8 refers to expression level of RNA for IL-8; if
IL-8>2.5 then IL-8 is set to 1, if ness than 2.5, 0. The symbols
"*" means the multiplication operator, and weighting factors, w1-w7
are respectively the weights assigned to each of the variables
listed above in the G+P INDEX.
[0152] In other preferred embodiments, (AgeGT50 and RBC) may be
dropped from the model as shown below:
G+P INDEX=(1*HFREQ+3*Gender+4*SMK).+-.(5*M1+2*IL-8)
[0153] The G+P INDEX produces a value between 0 and 15. A patient
with G+P INDEX value of 11 to 15 is considered to be at "High Risk"
for bladder cancer, and indicates the need for additional work up
for bladder cancer. A patient with a G+P INDEX value of 6 to 10 is
considered to be at "Moderate Risk" for developing bladder cancer,
and additional work up is indicated. A patient with a G+P INDEX of
0 to 5 is considered to be at "Low Risk" for developing bladder
cancer. Patients in the "Low Risk" group are placed on a watchful
waiting list, and if additional symptoms appear, or if recurrent
episodes of microhematuria occur, they are reevaluated for possible
further work up.
[0154] Genotypic variables useful for differentiating patients
without and with bladder cancer include expression of RNA markers
"M1" being a combination of MDK, CDC, IGFBP5 (IGBP5), and HOXA13.
Another genotypic variable is expression of RNA for IL8R.
Coefficients for these genotypic variables are shown in Table 7
below. A threshold of 4.5 and 2.5 was used for M1 and IL8R,
respectively, and a coefficient of 5 and 2, respectively, were
assigned.
[0155] The G+P INDEX produces a value between 0 and 15. A G+P INDEX
value of 11 to 15 is considered "High Risk" for bladder cancer, and
indicates the need for additional work up for bladder cancer. A G+P
INDEX value of 6 to 10 is considered "Moderate Risk" for developing
bladder cancer, and additional work up is indicated. A G+P INDEX of
0 to 5 is considered "Low Risk" for developing bladder cancer, and
these patients are placed on a waiting list, and if additional
symptoms appear, or if recurrent episodes of microhematuria occur,
are reevaluated for possible fuller work up.
[0156] As described more fully in Example 3 (FIGS. 18 and 19), we
found that the ROC curve for phenotypic data alone produced a
modest level of diagnostic power. The ROC curve for genotypic data
alone produced a significant level of diagnostic power. We found an
unexpectedly better diagnostic power when both genotypic and
phenotypic data was combined into a G+P INDEX.
Genetic Analysis of Patients not Having Bladder Cancer
[0157] In some preferred embodiments, this invention combines use
of a 4-marker assay or a Cxbladder.RTM. assay (genotypic variables)
and one or more of five key risk factors (phenotypic variables) to
produce a selection index that can be used to triage patients with
microscopic or macroscopic hematuria in terms of their potential
risk of having urothelial cancer. While not precluding the need for
a flexible cystoscopy, patients deemed at high risk of urothelial
cancer based on phenotypic variables and genotypic variables may be
seen earlier, potentially improving overall patient outcome.
[0158] Genotypic markers can be used as tools to detect cancer-free
patients or to select patient groups that are at low, medium or
high risk of having a disease. The markers can, for example, be
differentially expressed between disease tissue and corresponding
non-disease tissue. In this situation, the detection of
differential expression is associated with the presence of the
disease. Alternatively, the marker can be associated directly with
changes occurring in the disease tissues, or changes resulting from
the disease Inflammatory diseases are associated with an increase
in neutrophils. It has been found that the neutrophil marker
interleukin 8 receptor B (IL8Rb; FIG. 1; SEQ ID NOs 1 and 2), can
provide a good marker for the presence of neutrophils in a sample,
and therefore can be used as a diagnostic marker for the detection
of inflammatory disease in a sample, and in particular, in the
detection of inflammatory disease of the bladder.
[0159] As shown in FIG. 5, accumulation of IL8Rb in urine is
indicative of the presence of inflammatory disease of the bladder.
Specifically, FIG. 5 shows the accumulation of IL8Rb in the urine
of patients having the conditions; benign prostatic hyperplasia,
urinary tract infection, non-specific prostate diseases, vascular
prostate and secondary warfarin use. It will be appreciated
however, that the use of IL8Rb is not be limited to the detection
of these diseases only, but that these examples show that IL8Rb
does increase in samples from patients having an inflammatory
disease of the bladder. That is, IL8Rb can be used as a marker of
inflammation associated with bladder disease and therefore is
suitable for use in detecting any condition associated with
inflammation. Therefore, the detection of the amount of IL8Rb can
be used as a marker for inflammatory disease of the bladder. More
particularly, IL8Rb can be used to detect inflammatory disease of
the bladder associated with the accumulation of neutrophils.
[0160] Urine tests for TCC rely largely on the presence of markers
in the urine derived from exfoliated tumor cells. The ability to
detect these cells can be masked by the presence of large numbers
of contaminating cells, such as, blood and inflammatory cells.
Moreover, inflammation of the bladder lining can result in the
increased exfoliation of non-malignant cells from the mucosa. As a
result, urine tests that use markers derived from bladder
transitional cells have a higher likelihood of giving a false
positive result from urine samples taken from patients with
cystitis, urinary tract infection or other conditions resulting in
urinary tract inflammation or transitional cell exfoliation, such
as, urolithiasis (Sanchez-Carbayo et al).
[0161] One way to try and avoid such false positive results has
been to select markers with low relative expression in blood or
inflammatory cells. The use of such markers results in fewer false
positives in TCC patients presenting with non-malignant,
inflammatory conditions. However, low expression of the markers in
hematologically-derived cells fails to compensate for the enhanced
rate of exfoliation of non-malignant transitional cells.
[0162] It has been discovered that the negative impact of
exfoliated transitional cells from inflamed tissue has on the
accuracy of bladder cancer urine tests can be minimized by
improving the identification of patients with inflammatory
conditions of the urinary tract. Here it has been surprisingly
found that using the marker IL8Rb in combination with one or more
bladder tumor markers (BTM's) provides for a more accurate
detection of bladder cancer. In particular, a marker based test for
bladder cancer that includes the marker IL8Rb is less susceptible
to false positive results, which can result in patients suffering
from an inflammatory non-cancer condition.
[0163] In general, the presence or absence of an inflammatory
condition is established by having a threshold of gene expression,
above which expression of IL8Rb is indicative of an inflammatory
condition. For example, the expression of IL8Rb above a certain
threshold is diagnostic that the patient has an inflammatory
condition (see thresholds described above)
[0164] When IL8Rb is used in conjunction with one or more markers
predictive for the presence of bladder cancer, the presence of
elevated expression of the bladder tumor marker(s), and expression
of IL8Rb, above a certain threshold, is predictive of the patient
having an inflammatory condition and not cancer. Furthermore, if
the test is preformed on urine from the patient, then this result
is predictive of the patient having an inflammatory bladder
condition. The high levels of the bladder tumor markers are most
likely the result of non-malignant cells coming from the mucosa as
a result of the inflammation. That is, the patient, although having
high levels of the bladder tumor marker(s) does not actually have
bladder cancer--a false positive.
[0165] Alternatively, if the patient has abnormally high levels or
diagnostic levels of one or more bladder tumor markers, but the
level of IL8Rb is below a threshold, then the patient is likely to
have cancer, and in particular bladder or urothelial cancer. This
is especially so, if the test is preformed on urine from the
patient. This result is of significant benefit to the health
provider because they can be sure that the patient does have
cancer, and can start treatment immediately, and not be concerned
that the result is actually caused an inflammatory condition giving
a false positive result.
[0166] It has been surprisingly shown that the quantification of
RNA from the gene encoding the neutrophil marker interleukin 8
receptor B (IL8Rb) improves the overall performance of detecting
patients with TCC, using known TCC or BTM markers. The reference
sequences for IL8Rb are shown in FIG. 1 and SEQ ID NOs 1 and 2). In
addition to its role in TCC detection, it has been explored whether
IL8Rb could be used as a urine marker to aid in the diagnosis of
inflammatory disease (FIG. 5).
[0167] The use of IL8Rb marker can be used in isolation for the
detection of inflammatory conditions of the bladder utilizing known
methods for detecting gene expression levels. Examples of methods
for detecting gene expression are outlined below.
[0168] Alternatively, IL8Rb can be combined with one of more BTMs
to detect bladder cancer. It has been shown that by utilizing the
inflammatory disease marker IL8Rb as part of the test for bladder
cancer, the influence of inflamed tissue on creating a false
positive result is minimized. The marker IL8Rb can be used in
association with any bladder cancer markers, or alternatively can
be used with two or more markers, as part of a signature, for
detecting bladder cancer.
[0169] Reducing the number of false positive results means that
fewer patients not having bladder cancer are subjected to
potentially unnecessary procedures, including cystoscopy, which
carries its own risks. Reducing the number of false negative
results means that it is more likely that a patient with bladder
cancer is detected, and can therefore be further evaluated for
cancer.
[0170] The action of IL8Rb to improve the detection of bladder
cancer results from the ability to separate non-malignant
conditions from patients having bladder cancer. This is achieved
because an increase of IL8Rb is indicative of an increase in the
presence of neutrophils in a sample. Therefore, the ability of
IL8Rb is not dependent on the bladder tumor marker used. As shown
in FIGS. 2, and 12 to 15, when combined with a variety of bladder
tumor markers and combinations of bladder tumor markers, IL8Rb had
the general effect of increasing the specificity of the ability of
the marker(s) to detect cancer in the subjects.
[0171] One example of a signature according to the present
invention is the use of IL8Rb in combination with MDK, CDC2, IGFBP5
and HOXA13, which may also be in combination with one or more other
marker suitable for detecting bladder cancer, for example any one
of more of the markers outlined in FIG. 6 or 7. As shown in FIGS.
14 and 15, IL8Rb can be used in any combination of the markers,
specifically the combinations IL8Rb/MDK, IL8Rb/CDC2, IL8Rb/HOXA13,
IL8Rb/IGFBP5, IL8Rb/MDK/CDC2, IL8Rb/MDK/HOXA13, IL8Rb/MDK/IGFBP5,
IL8Rb/CDC2/HOXA13, IL8Rb/CDC2/IGFBP5, IL8Rb/HOXA13/IGFBP5,
IL8Rb/MDK/CDC2/HOXA13, IL8Rb/MDK/CDC2/IGFBP5,
IL8Rb/CDC2/HOXA13/IGFBP5, and IL8Rb/MDK/CDC2/HOXA13/IGFBP5. As
shown in FIGS. 14 and 15, the inclusion of IL8Rb increased the
ability of the marker, or the combination of markers to accurately
diagnose bladder cancer in a subject. The present invention is not
to be limited to these specific combinations but can optionally
include one or more further markers suitable for detecting bladder
cancer, for example any one of more of the markers outlined in FIG.
6 or 7. Table 1 below shows the identifiers for the specific
markers MDK, CDC2, IGFBP5 and HOXA13 and IL8Rb.
TABLE-US-00001 TABLE 1 Identifiers for Bladder Tumor Markers HGNC
NCBI Gene Entrez PE Gene Name Gene Name (Official) NCBI RefSeq ID
HGNC URL MDK MDK NM_002391 4192
http://www.genenames.org/data/hgnc_data.php?hgnc_id=6972 CDC CDK1
NM_001170406 983
http://www.genenames.org/data/hgnc_data.php?hgnc_id=1722 IGF IGFBP5
NM_000599 3488
http://www.genenames.org/data/hgnc_data.php?hgnc_id=5474 HOXA
HOXA13 NM_000522 3209
http://www.genenames.org/data/hgnc_data.php?hgnc_id=5102 IL8Rb
CXCR2 NM_001168298 3579
http://www.genenames.org/data/hgnc_data.php?hgnc_id=6027
[0172] FIGS. 2 to 4 and 12 to 17 show the effect of using IL8Rb in
combination with four known, representative, markers of bladder
cancer; MDK, CDC2, IGFBP5 and HOXA13. The results show that by
incorporating the use if IL8Rb individually with each marker (FIGS.
2, 14 and 15 to 17), but also when used with all possible
combinations of the four BTM markers as a signature, there is an
improvement in the ability to separate the samples of patients with
TCC and those with non-malignant conditions.
[0173] As shown in FIGS. 10 to 13, the inclusion of IL8Rb with the
four markers MDK, CDC2, IGFBP5 and HOXA13 (uRNA-D) not only
increased the overall performance of the test compared to the four
markers alone, the test also compared extremely favorably with
other known tests, NMP22.RTM. "a registered trademark of Matritech,
Inc., of Massachusetts, United States" Elisa, NMP22
BladderChek.RTM. (a registered trademark of Matritech, Inc., of
Massachusetts, United States), and cytology. FIGS. 14 through 17
also show the effect of IL8Rb in the various combinations of the
four markers MDK, CDC2, IGFBP5 and HOXA13.
[0174] FIG. 14 shows the ROC curves for all the combinations of the
four markers MDK, CDC2, IGFBP5 and HOXA13, with and without IL8Rb,
calculated using five different classifier models (i) Linear
Discriminant Analysis (LDA), (ii) Logistic Regression (Log Reg),
(iii) Support Vector Machines (SVM), (iv) K-nearest 5 neighbors
(KN5N), and (v) Partition Tree Classifier (TREE). FIG. 15 tabulates
the Area Under the Curve (AUC) for all 5 classifiers and all 15
combinations of the 4 biomarkers, with and without IL8Rb. This AUC
calculation is restricted to the area from a false positive rate of
0 to a false positive rate of 20%, covering the useful ranges of
specificity (80-100%). The AUC quantifies the visible differences
on the ROC curves of FIG. 14. FIG. 16 shows the sensitivity of all
combinations of the four markers measured with and without IL8Rb at
specificities of FIG. 16(a) 80%, FIG. 16(b) 85%, FIG. 16(c) 90%,
FIG. 16(d) 95%, and FIG. 16(e) 98%. FIG. 17 tabulates the changes
in either sensitivity (vertical direction on the ROC curves; better
is "up") or specificity (horizontal direction on the ROC curve;
better is to the left) at the fixed specificities of FIGS. 17(a, f)
80%, FIGS. 17(b, g) 85%, FIGS. 17(c, h) 90%, FIGS. 17(d, I) 95%,
and FIGS. 17(e, j) 98%, respectively.
[0175] These results show that IL8Rb, in general, improves the
ability of the biomarkers (MDK, CDC, IGFBP5, and HOXA13), singly or
in combination, to classify tumor from normal samples.
[0176] These results generally show that the IL8Rb was able to
increase the accuracy at which the test could detect bladder or
urothelial cancer. The biggest gains were seen with either markers
that did not perform as well without the inclusion of IL8Rb or with
classifiers that did not perform as well. Smaller gains were seen
for markers and/or classifiers that performed well prior to adding
IL8Rb and therefore there was less room for improvement. It is
important to note that the results show a population based analysis
and the benefit of incorporating IL8Rb could be greater when
diagnosis individual patients, especially those whose diagnosis on
the expression of the BTM markers maybe unclear.
[0177] These results show that not only can IL8Rb be used to detect
inflammatory disease of the bladder, but also when used in
combination with markers for bladder cancer, results in an improved
detection of bladder cancer, arising from a reduction in "false
positive" results.
[0178] These results also show the utility of IL8Rb in that it
affects the overall performance of the various markers
combinations, and confirms the ability of IL8Rb to improve the
performance of one or more bladder cancer markers to accurately
detect cancer in a patient. Further, FIG. 14 and FIG. 15 show that
the same results can be achieved using a range of classifier
models, and shows that the result is not dependent on a classifier
model or algorithm, but rather the combination of markers used.
These results confirm that any suitable classifier model or
algorithm could be used in the present invention. In particular,
FIG. 14 and FIG. 15 show that IL8Rb has a greater effect at the
higher specificities, and in particular in the most clinically
applicable ranges.
[0179] Therefore, using the G+P Index of this invention, we are
able to accurately triage patients with hematuria based on
phenotypic and genetic variables into groups being at "High Risk"
for having urothelial cancer and warrant immediate further work-up,
"At Risk" and warrant immediate work-up, and those with "Low Risk"
who may be placed on a watch list for later evaluation.
Detection of Genetic Markers in Body Samples
[0180] In several preferred embodiments, assays for cancer can be
desirably carried out on samples obtained from blood, plasma,
serum, peritoneal fluid obtained for example using peritoneal
washes, or other body fluids, such as urine, lymph, cerebrospinal
fluid, gastric fluid or stool samples. For the detection of
inflammatory conditions of the bladder or bladder cancer the test
is ideally preformed on a urine sample.
[0181] Specifically, present methods for detecting inflammatory
bladder disease or bladder cancer can be conducted on any suitable
sample from the body that would be indicative of the urine, but
ideally the level of IL8Rb, and any further cancer marker is
established directly from a urine sample.
[0182] A test can either be performed directly on a urine sample,
or the sample may be stabilized by the addition of any suitable
compounds or buffers known in the art to stabilize and prevent the
breakdown of RNA and/or protein in the sample so that it can be
analyzed at a later date, or even to ensure that the RNA and/or
protein is stabilized during the analysis.
[0183] The determination of either the protein and/or RNA level in
the subject's urine can be performed directly on the urine, or the
urine can be treated to further purify and/or concentrate the RNA
and/or protein. Many methods for extracting and/or concentrating
proteins and/or RNA are well known in the art and could be used in
the present invention.
[0184] It can be appreciated that many methods are well known in
the art for establishing the expression level of a particular gene,
either at the RNA and/or protein level, and any suitable method can
be used in the present invention. Some common methods are outlined
below, however, the invention is not restricted to these methods
and any method for quantifying protein and/or RNA levels is
suitable for use in the present invention.
General Approaches to Disease and Cancer Detection Using Genotypic
Markers
[0185] General methodologies for determining expression levels are
outlined below, although it will be appreciated that any method for
determining expression levels would be suitable.
Quantitative PCR (qPCR)
[0186] Quantitative PCR (qPCR) can be carried out on tumor samples,
on serum and plasma using specific primers and probes. In
controlled reactions, the amount of product formed in a PCR
reaction (Sambrook, J., E Fritsch, E. and T Maniatis, Molecular
Cloning: A Laboratory Manual 3.sup.rd. Cold Spring Harbor
Laboratory Press: Cold Spring Harbor (2001)) correlates with the
amount of starting template. Quantification of the PCR product can
be carried out by stopping the PCR reaction when it is in log
phase, before reagents become limiting. The PCR products are then
electrophoresed in agarose or polyacrylamide gels, stained with
ethidium bromide or a comparable DNA stain, and the intensity of
staining measured by densitometry. Alternatively, the progression
of a PCR reaction can be measured using PCR machines such as the
Applied Biosystems' Prism7000.TM. (a trademark of Applera
Corporation, Connecticut, United States) or the Roche
LightCycler.TM. (a trademark of Roche Molecular Systems, Inc.,
California, United States), which measure product accumulation in
real-time. Real-time PCR measures either the fluorescence of DNA
intercalating dyes such as Sybr Green into the synthesized PCR
product, or the fluorescence released by a reporter molecule when
cleaved from a quencher molecule; the reporter and quencher
molecules are incorporated into an oligonucleotide probe which
hybridizes to the target DNA molecule following DNA strand
extension from the primer oligonucleotides. The oligonucleotide
probe is displaced and degraded by the enzymatic action of the Taq
polymerase in the next PCR cycle, releasing the reporter from the
quencher molecule. In one variation, known as Scorpion, the probe
is covalently linked to the primer.
Reverse Transcription PCR (RT-PCR)
[0187] RT-PCR can be used to compare RNA levels in different sample
populations, in normal and tumor tissues, with or without drug
treatment, to characterize patterns of expression, to discriminate
between closely related RNAs, and to analyze RNA structure.
[0188] For RT-PCR, the first step is the isolation of RNA from a
target sample. The starting material is typically total RNA
isolated from human tumors or tumor cell lines, and corresponding
normal tissues or cell lines, respectively. RNA can be isolated
from a variety of samples, such as tumor samples from breast, lung,
colon (e.g., large bowel or small bowel), colorectal, gastric,
esophageal, anal, rectal, prostate, brain, liver, kidney, pancreas,
spleen, thymus, testis, ovary, uterus, bladder etc., tissues, from
primary tumors, or tumor cell lines, and from pooled samples from
healthy donors. If the source of RNA is a tumor, RNA can be
extracted, for example, from frozen or archived paraffin-embedded
and fixed (e.g., formalin-fixed) tissue samples.
[0189] The first step in gene expression profiling by RT-PCR is the
reverse transcription of the RNA template into cDNA, followed by
its exponential amplification in a PCR reaction. The two most
commonly used reverse transcriptases are avian myeloblastosis virus
reverse transcriptase (AMV-RT) and Moloney murine leukemia virus
reverse transcriptase (MMLV-RT). The reverse transcription step is
typically primed using specific primers, random hexamers, or
oligo-dT primers, depending on the circumstances and the goal of
expression profiling. For example, extracted RNA can be
reverse-transcribed using a GeneAmp.RTM. RNA PCR kit (Perkin Elmer,
Calif., USA), following the manufacturer's instructions. The
derived cDNA can then be used as a template in the subsequent PCR
reaction.
[0190] Although the PCR step can use a variety of thermostable
DNA-dependent DNA polymerases, it typically employs the Taq DNA
polymerase, which has a 5'-3' nuclease activity but lacks a 3'-5'
proofreading endonuclease activity. Thus, TaqMan.RTM. qPCR (a
registered trademark of Roche Molecular Systems, Inc., California,
United States) typically utilizes the 5' nuclease activity of Taq
or Tth polymerase to hydrolyze a hybridization probe bound to its
target amplicon, but any enzyme with equivalent 5' nuclease
activity can be used.
[0191] Two oligonucleotide primers are used to generate an amplicon
typical of a PCR reaction. A third oligonucleotide, or probe, is
designed to detect nucleotide sequence located between the two PCR
primers. The probe is non-extendible by Taq DNA polymerase enzyme,
and is labeled with a reporter fluorescent dye and a quencher
fluorescent dye. Any laser-induced emission from the reporter dye
is quenched by the quenching dye when the two dyes are located
close together as they are on the probe. During the amplification
reaction, the Taq DNA polymerase enzyme cleaves the probe in a
template-dependent manner. The resultant probe fragments
disassociate in solution, and signal from the released reporter dye
is free from the quenching effect of the second fluorophore. One
molecule of reporter dye is liberated for each new molecule
synthesized, and detection of the unquenched reporter dye provides
the basis for quantitative interpretation of the data.
[0192] TaqMan.RTM. RT-PCR (a registered trademark of Roche
Molecular Systems, Inc., California, United States) can be
performed using commercially available equipment, such as, for
example, ABI PRISM7700.TM. Sequence Detection System (a trademark
of Applera Corporation, Connecticut, United States)
(Perkin-Elmer-Applied Biosystems, Foster City, Calif., USA), or
Lightcycler.TM. (a registered trademark of Roche Molecular Systems,
Inc., California, United States((Roche Molecular Biochemicals,
Mannheim, Germany). In a preferred embodiment, the 5' nuclease
procedure is run on a real-time quantitative PCR device such as the
ABI PRISM 7700.TM. Sequence Detection System. The system consists
of a thermocycler, laser, charge-coupled device (CCD), camera, and
computer. The system amplifies samples in a 96-well format on a
thermocycler. During amplification, laser-induced fluorescent
signal is collected in real-time through fiberoptic cables for all
96 wells, and detected at the CCD. The system includes software for
running the instrument and for analyzing the data.
[0193] 5' nuclease assay data are initially expressed as Cp, or the
threshold cycle. As discussed above, fluorescence values are
recorded during every cycle and represent the amount of product
amplified to that point in the amplification reaction. The point
when the fluorescent signal is first recorded as statistically
significant is the threshold cycle, Cp.
Real-Time Quantitative PCR (qRT-PCR)
[0194] A more recent variation of the RT-PCR technique is the real
time quantitative PCR, which measures PCR product accumulation
through a dual-labeled fluorigenic probe (i.e., TaqMan.RTM. probe.
Real time PCR is compatible both with quantitative competitive PCR
and with quantitative comparative PCR. The former uses an internal
competitor for each target sequence for normalization, while the
latter uses a normalization gene contained within the sample, or a
housekeeping gene for RT-PCR. Further details are provided, e.g.,
by Held et al., Genome Research 6: 986-994 (1996).
[0195] Expression levels can be determined using fixed,
paraffin-embedded tissues as the RNA source. According to one
aspect of the present invention, PCR primers are designed to flank
intron sequences present in the gene to be amplified. In this
embodiment, the first step in the primer/probe design is the
delineation of intron sequences within the genes. This can be done
by publicly available software, such as the DNA BLAT software
developed by Kent, W. J., Genome Res. 12 (4): 656-64 (2002), or by
the BLAST software including its variations. Subsequent steps
follow well established methods of PCR primer and probe design.
[0196] In order to avoid non-specific signals, it is useful to mask
repetitive sequences within the introns when designing the primers
and probes. This can be easily accomplished by using the Repeat
Masker program available on-line through the Baylor College of
Medicine, which screens DNA sequences against a library of
repetitive elements and returns a query sequence in which the
repetitive elements are masked. The masked sequences can then be
used to design primer and probe sequences using any commercially or
otherwise publicly available primer/probe design packages, such as
Primer Express (Applied Biosystems); MGB assay-by-design (Applied
Biosystems); Primer3 (Steve Rozen and Helen J. Skaletsky (2000)
Primer3 on the VIMNV for general users and for biologist
programmers in: Krawetz S, Misener S (eds) Bioinformatics Methods
and Protocols: Methods in Molecular Biology. Humana Press, Totowa,
N.J., pp 365-386).
[0197] The most important factors considered in PCR primer design
include primer length, melting temperature (Tm), and G/C content,
specificity, complementary primer sequences, and 3' end sequence.
In general, optimal PCR primers are generally 1730 bases in length,
and contain about 20-80%, such as, for example, about 50-60% G+C
bases. Melting temperatures between 50 and 80.degree. C., e.g.,
about 50 to 70.degree. C., are typically preferred. For further
guidelines for PCR primer and probe design see, e.g., Dieffenbach,
C. W. et al., General Concepts for PCR Primer Design in: PCR
Primer, A Laboratory Manual, Cold Spring Harbor Laboratory Press,
New York, 1995, pp. 133-155; Innis and Gelfand, Optimization of
PCRs in: PCR Protocols, A Guide to Methods and Applications, CRC
Press, London, 1994, pp. 5-11; and Plasterer, T. N. Primerselect:
Primer and probe design. Methods Mol. Biol. 70: 520-527 (1997), the
entire disclosures of which are hereby expressly incorporated by
reference.
Microarray Analysis
[0198] Differential expression can also be identified, or confirmed
using the microarray technique. Thus, the expression profile of
disease specific markers can be measured in either fresh or
paraffin-embedded tumor tissue, using microarray technology. In
this method, polynucleotide sequences of interest (including cDNAs
and oligonucleotides) are plated, or arrayed, on a microchip
substrate. The arrayed sequences (i.e., capture probes) are then
hybridized with specific polynucleotides from cells or tissues of
interest (i.e., targets). Just as in the RT-PCR method, the source
of RNA typically is total RNA isolated from human tumors or tumor
cell lines, and corresponding normal tissues or cell lines. Thus
RNA can be isolated from a variety of primary tumors or tumor cell
lines. If the source of RNA is a primary tumor, RNA can be
extracted, for example, from frozen or archived formalin fixed
paraffin-embedded (FFPE) tissue samples and fixed (e.g.,
formalin-fixed) tissue samples, which are routinely prepared and
preserved in everyday clinical practice.
[0199] In a specific embodiment of the microarray technique, PCR
amplified inserts of cDNA clones are applied to a substrate. The
substrate can include up to 1, 2, 5, 10, 15, 20, 25, 30, 35, 40,
45, 50, or 75 nucleotide sequences. In other aspects, the substrate
can include at least 10,000 nucleotide sequences. The microarrayed
sequences, immobilized on the microchip, are suitable for
hybridization under stringent conditions. As other embodiments, the
targets for the microarrays can be at least 50, 100, 200, 400, 500,
1000, or 2000 bases in length; or 50-100, 100-200, 100-500,
100-1000, 100-2000, or 500-5000 bases in length. As further
embodiments, the capture probes for the microarrays can be at least
10, 15, 20, 25, 50, 75, 80, or 100 bases in length; or 10-15,
10-20, 10-25, 10-50, 10-75, 10-80, or 20-80 bases in length.
[0200] Fluorescently labeled cDNA probes may be generated through
incorporation of fluorescent nucleotides by reverse transcription
of RNA extracted from tissues of interest. Labeled cDNA probes
applied to the chip hybridize with specificity to each spot of DNA
on the array. After stringent washing to remove non-specifically
bound probes, the chip is scanned by confocal laser microscopy or
by another detection method, such as a CCD camera. Quantitation of
hybridization of each arrayed element allows for assessment of
corresponding mRNA abundance. With dual colour fluorescence,
separately labeled cDNA probes generated from two sources of RNA
are hybridized pairwise to the array. The relative abundance of the
transcripts from the two sources corresponding to each specified
gene is thus determined simultaneously.
[0201] The miniaturized scale of the hybridization affords a
convenient and rapid evaluation of the expression pattern for large
numbers of genes. Such methods have been shown to have the
sensitivity required to detect rare transcripts, which are
expressed at a few copies per cell, and to reproducibly detect at
least approximately two-fold differences in the expression levels
(Schena et al., Proc. Natl. Acad. Sci. USA 93 (2): 106-149 (1996)).
Microarray analysis can be performed by commercially available
equipment, following manufacturer's protocols, such as by using the
Affymetrix GenChip.RTM. technology, Illumina microarray technology
or Incyte's microarray technology. The development of microarray
methods for large-scale analysis of gene expression makes it
possible to search systematically for molecular markers of cancer
classification and outcome prediction in a variety of tumor
types.
RNA Isolation, Purification, and Amplification
[0202] General methods for mRNA extraction are well known in the
art and are disclosed in standard textbooks of molecular biology,
including Ausubel et al., Current Protocols of Molecular Biology,
John Wiley and Sons (1997). Methods for RNA extraction from
paraffin embedded tissues are disclosed, for example, in Rupp and
Locker, Lab Invest. 56: A67 (1987), and De Sandres et al.,
BioTechniques 18: 42044 (1995). In particular, RNA isolation can be
performed using purification kit, buffer set, and protease from
commercial manufacturers, such as Qiagen, according to the
manufacturer's instructions. For example, total RNA from cells in
culture can be isolated using Qiagen RNeasy.RTM. "a registered
trademark of Qiagen GmbH, Hilden, Germany" mini-columns. Other
commercially available RNA isolation kits include MasterPure.TM.
Complete DNA and RNA Purification Kit (EPICENTRE (D, Madison,
Wis.), and Paraffin Block RNA Isolation Kit (Ambion, Inc.). Total
RNA from tissue samples can be isolated using RNA Stat-60
(Tel-Test). RNA prepared from tumour can be isolated, for example,
by cesium chloride density gradient centrifugation.
[0203] The steps of a representative protocol for profiling gene
expression using fixed, paraffin-embedded tissues as the RNA
source, including mRNA isolation, purification, primer extension
and amplification are given in various published journal articles
(for example: T. E. Godfrey et al. J. Molec. Diagnostics 2: 84-91
(2000); K. Specht et al., Am. J. Pathol. 158: 419-29 (2001)).
Briefly, a representative process starts with cutting about 10
micron thick sections of paraffin-embedded tumor tissue samples.
The RNA is then extracted, and protein and DNA are removed. After
analysis of the RNA concentration, RNA repair and/or amplification
steps may be included, if necessary, and RNA is reverse transcribed
using gene specific promoters followed by RT-PCR. Finally, the data
are analyzed to identify the best treatment option(s) available to
the patient on the basis of the characteristic gene expression
pattern identified in the tumor sample examined.
Immunohistochemistry and Proteomics
[0204] Immunohistochemistry methods are also suitable for detecting
the expression levels of the proliferation markers of the present
invention. Thus, antibodies or antisera, preferably polyclonal
antisera, and most preferably monoclonal antibodies specific for
each marker, are used to detect expression. The antibodies can be
detected by direct labeling of the antibodies themselves, for
example, with radioactive labels, fluorescent labels, hapten labels
such as, biotin, or an enzyme such as horseradish peroxidase or
alkaline phosphatase. Alternatively, unlabeled primary antibody is
used in conjunction with a labeled secondary antibody, comprising
antisera, polyclonal antisera or a monoclonal antibody specific for
the primary antibody. Immunohistochemistry protocols and kits are
well known in the art and are commercially available.
[0205] Proteomics can be used to analyze the polypeptides present
in a sample (e.g., tissue, organism, or cell culture) at a certain
point of time. In particular, proteomic techniques can be used to
assess the global changes of polypeptide expression in a sample
(also referred to as expression proteomics). Proteomic analysis
typically includes: (1) separation of individual polypeptides in a
sample by 2-D polyacrylamide gel electrophoresis (2-D PAGE); (2)
identification of the individual polypeptides recovered from the
gel, e.g., by mass spectrometry or N-terminal sequencing, and (3)
analysis of the data using bioinformatics. Proteomics methods are
valuable supplements to other methods of gene expression profiling,
and can be used, alone or in combination with other methods, to
detect the products of the proliferation markers of the present
invention.
Hybridization Methods Using Nucleic Acid Probes Selective for a
Marker
[0206] These methods involve binding the nucleic acid probe to a
support, and hybridizing under appropriate conditions with RNA or
cDNA derived from the test sample (Sambrook, J., E Fritsch, E. and
T Maniatis, Molecular Cloning: A Laboratory Manual 3.sup.rd. Cold
Spring Harbor Laboratory Press: Cold Spring Harbor (2001)). These
methods can be applied to markers derived from a tumour tissue or
fluid sample. The RNA or cDNA preparations are typically labeled
with a fluorescent or radioactive molecule to enable detection and
quantification. In some applications, the hybridizing DNA can be
tagged with a branched, fluorescently labeled structure to enhance
signal intensity (Nolte, F. S., Branched DNA signal amplification
for direct quantitation of nucleic acid sequences in clinical
specimens. Adv. Clin. Chem. 33, 201-35 (1998)). Unhybridized label
is removed by extensive washing in low salt solutions such as
0.1.times.SSC, 0.5% SDS before quantifying the amount of
hybridization by fluorescence detection or densitometry of gel
images. The supports can be solid, such as nylon or nitrocellulose
membranes, or consist of microspheres or beads that are hybridized
when in liquid suspension. To allow washing and purification, the
beads may be magnetic (Haukanes, B-1 and Kvam, C., Application of
magnetic beads in bioassays. Bio/Technology 11, 60-63 (1993)) or
fluorescently-labeled to enable flow cytometry (see for example:
Spiro, A., Lowe, M. and Brown, D., A Bead-Based Method for
Multiplexed Identification and Quantitation of DNA Sequences Using
Flow Cytometry. Appl. Env. Micro. 66, 4258-4265 (2000)).
[0207] A variation of hybridization technology is the QuantiGene
Plex.RTM. assay (a registered trademark of Panomics, of California,
United States) (Genospectra, Fremont) which combines a fluorescent
bead support with branched DNA signal amplification. Still another
variation on hybridization technology is the Quantikine.RTM. mRNA
assay (R&D Systems, Minneapolis). Methodology is as described
in the manufacturer's instructions. Briefly the assay uses
oligonucleotide hybridization probes conjugated to Digoxigenin.
Hybridization is detected using anti-Digoxigenin antibodies coupled
to alkaline phosphatase in colorometric assays.
[0208] Additional methods are well known in the art and need not be
described further herein.
Enzyme-Linked Immunological Assays (ELISA)
[0209] Briefly, in sandwich ELISA assays, a polyclonal or
monoclonal antibody against the marker is bound to a solid support
(Crowther, J. R. The ELISA guidebook. Humana Press: New Jersey
(2000); Harlow, E. and Lane, D., Using antibodies: a laboratory
manual. Cold Spring Harbor Laboratory Press: Cold Spring Harbor
(1999)) or suspension beads. Other methods are known in the art and
need not be described herein further. Monoclonal antibodies can be
hybridoma-derived or selected from phage antibody libraries (Hust
M. and Dubel S., Phage display vectors for the in vitro generation
of human antibody fragments. Methods Mol Biol. 295:71-96 (2005)).
Nonspecific binding sites are blocked with non-target protein
preparations and detergents. The capture antibody is then incubated
with a preparation of sample or tissue from the patient containing
the antigen. The mixture is washed before the antibody/antigen
complex is incubated with a second antibody that detects the target
marker. The second antibody is typically conjugated to a
fluorescent molecule or other reporter molecule that can either be
detected in an enzymatic reaction or with a third antibody
conjugated to a reporter (Crowther, Id.). Alternatively, in direct
ELISAs, the preparation containing the marker can be bound to the
support or bead and the target antigen detected directly with an
antibody-reporter conjugate (Crowther, Id.).
[0210] Methods for producing monoclonal antibodies and polyclonal
antisera are well known in the art and need not be described herein
further.
Immunodetection
[0211] The methods can also be used for immunodetection of marker
family members in sera or plasma from bladder cancer patients taken
before and after surgery to remove the tumour, immunodetection of
marker family members in patients with other cancers, including but
not limited to, colorectal, pancreatic, ovarian, melanoma, liver,
oesophageal, stomach, endometrial, and brain and immunodetection of
marker family members in urine and stool from bladder cancer
patients.
[0212] Disease markers can also be detected in tissues or samples
using other standard immunodetection techniques such as
immunoblotting or immunoprecipitation (Harlow, E. and Lane, D.,
Using antibodies: a laboratory manual. Cold Spring Harbor
Laboratory Press: Cold Spring Harbor (1999)). In immunoblotting,
protein preparations from tissue or fluid containing the marker are
electrophoresed through polyacrylamide gels under denaturing or
non-denaturing conditions. The proteins are then transferred to a
membrane support such as nylon. The marker is then reacted directly
or indirectly with monoclonal or polyclonal antibodies as described
for immunohistochemistry. Alternatively, in some preparations, the
proteins can be spotted directly onto membranes without prior
electrophoretic separation. Signal can be quantified by
densitometry.
[0213] In immunoprecipitation, a soluble preparation containing the
marker is incubated with a monoclonal or polyclonal antibody
against the marker. The reaction is then incubated with inert beads
made of agarose or polyacrylamide with covalently attached protein
A or protein G. The protein A or G beads specifically interact with
the antibodies forming an immobilized complex of
antibody-marker-antigen bound to the bead. Following washing the
bound marker can be detected and quantified by immunoblotting or
ELISA.
Establishing a Diagnosis Based on Genotypic Analysis
[0214] Once the level of expression of IL8Rb, and optionally one or
more further cancer markers, has been obtained then a diagnosis for
that subject can be established. If the expression of IL8Rb is
above the expression seen in subjects that do not have an
inflammatory bladder disease, and/or is consistent with the level
of expression in subjects known to have an inflammatory bladder
disease, then the subject will be diagnosed as having an
inflammatory bladder disease. Alternatively, if the expression is
not above the expression seen in subjects that do not have an
inflammatory bladder disease, and/or is below the levels of
expression in subjects known to have an inflammatory bladder
disease, then the subject will be diagnosed as not an inflammatory
bladder disease.
[0215] In the situation where IL8Rb is used in conjunction with one
or more markers for Bladder cancer, then the expression level of
IL8Rb will be compared with the level of expression of subjects
without an inflammatory bladder disease, and/or subjects known to
have an inflammatory bladder disease. The one or more cancer
markers are compared to the expression level in subjects without
bladder cancer and/or subjects known to have bladder cancer. If the
expression level of the IL8Rb is consistent with a subject that
does not have an inflammatory bladder disease (less than a subject
having an inflammatory bladder disease) and the expression level of
the one or more bladder cancer markers are consistent with a
subject having bladder cancer (differential to a subject that does
not have bladder cancer), then the subject is diagnosed as having
bladder cancer. If the expression level of the IL8Rb is greater
than a subject that does not have an inflammatory bladder disease
(consistent with a subject having an inflammatory bladder disease)
and the expression level of the one or more bladder cancer markers
are consistent with a subject having bladder cancer (differential
to a subject that does not have bladder cancer), then the subject
is diagnosed as having an inflammatory bladder disease. If the
expression level of the IL8Rb is consistent with a subject that
does not have an inflammatory bladder disease (less than a subject
having an inflammatory bladder disease) and the expression level of
the one or more bladder cancer markers are consistent with a
subject that does not have bladder cancer (differential to a
subject that does have bladder cancer), then the subject is
diagnosed as having neither bladder cancer or an inflammatory
bladder disease.
[0216] Because there is often an overlap in expression levels
between the normal and disease expression of a diagnostic marker,
in order to establish a diagnosis for a subject it is typical to
establish a classifying threshold. A classifying threshold is a
value or threshold which distinguishes subjects into disease or non
disease categories. A threshold is commonly evaluated with the use
of a Receiver Operating Characteristic (ROC) curve, which plots the
sensitivity against specificity for all possible thresholds.
Determination of Diagnostic Thresholds
[0217] For tests using disease markers, diagnostic thresholds can
be derived that enable a sample to be called either positive or
negative for the disease, e.g., bladder cancer. These diagnostic
thresholds are determined by the analysis of cohorts of patients
that are investigated for the presence of bladder cancer or
inflammatory bladder disease. Diagnostic thresholds may vary for
different test applications; for example, diagnostic thresholds for
use of the test in population screening are determined using
cohorts of patients who are largely free of urological symptoms,
and these diagnostic thresholds may be different to those used in
tests for patients who are under surveillance for bladder cancer
recurrence. A diagnostic threshold can be selected to provide a
practical level of test specificity in the required clinical
setting; that is, a specificity that allows reasonable sensitivity
without excessive numbers of patients receiving false positive
results. This specificity may be within the range of 80-100%.
[0218] A diagnostic threshold is determined by applying an
algorithm that combines the genotypic expression levels of each
marker to each sample from a prospective clinical trial.
[0219] Samples used are from patients with bladder cancer and a
range of non-malignant urological disorders. A diagnostic threshold
is selected by determining the score of the algorithm that resulted
in the desired specificity. For example, in some applications a
specificity of 85% is desired. The A diagnostic threshold is then
set by selecting an algorithm score that results in 85% of patients
without bladder cancer being correctly classed as negative for
cancer. In other applications (such as population screening),
higher specificity, such as 90%, is favoured. To set a threshold
for this application, an algorithm score that results in 90% of
patients without bladder cancer being correctly classed as negative
for cancer is selected. Examples of the use of an algorithm is
outlined in the Examples.
[0220] As an alternative to single thresholds, the test may use
test intervals which provide different degrees of likelihood of
presence of disease and which have different clinical consequences
associated with them. For example, a test may have three intervals;
one associated with a high (e.g. 90%) risk of the presence of
bladder cancer, a second associated with a low risk of bladder
cancer and a third regarded as being suspicious of disease. The
"suspicious" interval could be associated with a recommendation for
a repeat test in a defined period of time.
[0221] Data Analysis
[0222] Once the method to test for the amount of RNA and/or protein
has been completed, the data then has to be analyzed in order to
determine the distribution of biomarker values associated with
tumor and non-tumor samples. This typically involves normalizing
the raw data, i.e., removing background "noise" etc and averaging
any duplicates (or more), comparison with standards and
establishing cut-offs or thresholds to optimally separate the two
classes of samples. Many methods are known to do this, and the
exact method will depend on specific method for determining the
amount of RNA and/or protein used.
[0223] Below is an example of how the data analysis could be
performed when using qRT-PCR. However, it will be appreciated the
general process could be adapted to be used for other methods of
establishing the RNA and/or protein content, or other methods could
be established by someone skilled in the art to achieve the same
result.
[0224] Data
[0225] Measurements of fluorescence are taken at wavelengths
.omega..sub.i i=1,2 at each cycle of the PCR. Thus for each well we
observe a pair of fluorescence curves, denoted by
f.sub.t(.omega..sup.i), where t=1, . . . , k denotes cycle number
and i=1,2 indexes the wavelengths.
[0226] Fluorescence curves have a sigmoidal shape beginning with a
near horizontal baseline and increasing smoothly to an upper
asymptote. The location of a point C.sub.p where the fluorescence
curve departs from the linear baseline will be used to characterize
the concentration of the target gene. A precise definition of
C.sub.p follows later. The following is an example of a scheme to
process these data. [0227] Compensate for fluorescence overlap
between frequency bands, [0228] Estimate a smooth model for each
fluorescence curve in order to estimate C.sub.p [0229] Combine data
from replicated wells. [0230] Estimate standard curves [0231]
Compute a concentration relative to the standard. Each biological
sample yields relative concentrations of 5 genes, which are the
inputs to the discriminant function.
[0232] Color Compensation
[0233] Denote the level of fluorescence of dye j at cycle t and
frequency .omega. by W.sub.tj(.omega.). In a multiplexed assay the
measured response at any frequency .omega. is the sum of
contributions from all dyes at that frequency, so for each
cycle.
f.sub.t(.omega.)=W.sub.t1(.omega.)+W.sub.t2(.omega.)+ . . .
The purpose of color compensation is to extract the individual
contributions W.sub.tj(.omega.), from the observed mixtures
f.sub.t(.omega.).
[0234] In the ideal situation, fluorescence
W.sub.tj(.omega..sub.o), due to dye j at a frequency .omega. is
proportional to its fluorescence W.sub.tj(.omega..sub.o) at
reference frequency .omega..sub.o, regardless of the level of
W.sub.tj(.omega..sub.o). This suggests the linear relationship
[ f t ( .omega. 1 ) f t ( .omega. 2 ) ] = [ W t 1 ( .omega. 1 ) + W
t 2 ( .omega. 1 ) W t 1 ( .omega. 2 ) + W t 2 ( .omega. 2 ) ] = [ 1
A 12 A 21 1 ] [ W t 1 ( .omega. 1 ) W t 2 ( .omega. 2 ) ]
##EQU00002##
for some proportionality constants A.sub.12 and A.sub.21 that are
to be determined.
[0235] In reality, there are additional effects, which are
effectively modeled by introducing linear terms in this system,
so
[ f t ( .omega. 1 ) f t ( .omega. 2 ) ] = [ 1 A 12 A 21 1 ] [ W t 1
( .omega. 1 ) W t 2 ( .omega. 2 ) ] + [ a 1 + b 1 t a 2 + b 2 t ]
##EQU00003##
After estimating the "color compensation" parameters A.sub.12 and
A.sub.21 we can recover W.sub.t1(.omega..sub.1) and
W.sub.t2(.omega..sub.2), albeit distorted by a linear baseline, by
matrix multiplication:
[ W t 1 ( .omega. 1 ) W t 2 ( .omega. 2 ) ] = [ 1 A 12 A 21 1 ] - 1
[ f t ( .omega. 1 ) f t ( .omega. 2 ) ] + [ a 1 * + b 1 * t a 2 * +
b 2 * t ] ##EQU00004##
[0236] W.sub.t1(.omega..sub.1) and W.sub.t2(.omega..sub.2) are
called "color compensated" data. The linear distortions
a.sub.i*+b.sub.i*t in the last term of this expression will be
accommodated in the baseline estimate when estimating a model for
the colour compensated data below 2. It has no influence on the
estimate of C.sub.p.
[0237] Estimation of the color compensation coefficients requires a
separate assay using single (as opposed to duplex) probes. Then
W.sub.t2(.omega..sub.2)=0 giving:
[ f t ( .omega. 1 ) f t ( .omega. 2 ) ] = [ 1 A 12 A 21 1 ] [ W t 1
( .omega. 1 ) 0 ] + [ a 1 + b 1 t a 2 + b 2 t ] ##EQU00005##
Thus,
f.sub.t(.omega..sub.2)=A.sub.21f.sub.t(.omega..sub.1)+a*+b*t
[0238] The coefficient A.sub.21 can be estimated by ordinary linear
regression of f.sub.t(.omega..sub.2) on f.sub.t(.omega..sub.1) and
PCR cycle t for t=1, . . . , k.
Model Estimation
[0239] In this section, let y.sub.t t=1, . . . , k denote a color
compensated fluorescence curve.
[0240] Amplification
[0241] Models are only estimated for fluorescence curves that show
non-trivial amplification. We define the term "amplification" as a
non-trivial departure from the linear baseline of the color
compensated fluorescence curve. Use signal to noise ratio (SNR) to
quantify amplification. Here SNR is defined as the ratio of signal
variance to noise variance. Noise variance is set as part of
calibration of the assay procedure and remains unchanged: for this
purpose, use the residual variance from a linear model for the
baseline from wells that can have no amplification, i.e., wells
without RNA. For each fluorescence curve, estimate the signal
variance as the residual variance from the best fitting straight
line ("best" is meant in the least squares sense.) [0242] If SNR is
less than a specified threshold, the fluorescence curve is close to
linear and no amplification is present. Then there is no point of
departure from the baseline and the concentration in the sample may
be declared as zero. [0243] If the SNR is above the threshold,
amplification is present and a concentration can be estimated.
Thresholds for the (dimensionless) SNR are selected to provide
clear discrimination between "amplified" and "non amplified"
curves. For example, the following ranges for thresholds are
effective for the markers.
TABLE-US-00002 [0243] Fluor Gene Range JOE MDK 40-120 JOE CDC 35-70
JOE IL8R 30-60 FAM IGF 50-80 FAM HOXA 50-150 FAM XENO 50-80
[0244] Model
[0245] Estimate a sigmoidal model for each fluorescence curve. Any
suitable parametric form of model can be used, but it must be able
to model the following features: [0246] linear baseline that may
have a non-zero slope, [0247] asymmetries about the mid point.
[0248] asymptotes at lower and upper levels [0249] smooth increase
from baseline to upper asymptote An example of a model that
achieves these requirements is
[0249] g t ( .theta. ) = A + A s t + D ( 1 + ( t B ) E ) F
##EQU00006##
We call this the "6PL model". The parameter vector
.theta.=[A,As,D,B,E,F] is subject to the following constraints to
ensure that g.sub.t (.theta.) is an increasing function of t and
has the empirical properties of a fluorescence curve.
D>0,B>0,E<0,F<0
[0250] The other two parameters determine the base line A+A.sub.st,
and these parameters do not need explicit constraints though A is
always positive and the slope parameter A.sub.s is always small.
The parameter D determines the level of amplification above the
baseline. The remaining parameters B.E.F have no intrinsic
interpretation in themselves but control the shape of the curve.
These parameters are also the only parameters that influence the
estimate of C.sub.p. When A.sub.s=0 this is known as the
five-parameter logistic function (5PL) and if, in addition, F=1
this model reduces four-parameter logistic model (4PL), Gottschalk
and Dunn (2005), Spiess et al. (2008).
[0251] Initialization
[0252] Initial values for non-linear estimation are set as [0253]
A.sub.s=0,F=1 [0254] A=mean(y, . . . , y.sub.5) [0255]
D=range(y.sub.1, . . . , yk) [0256] B=cycle corresponding to half
height [0257] E is initialized by converting g.sub.t (.theta.) into
a linear form having set the values of the remaining parameters to
their initial values defined above. Linearization obtains
[0257] E log ( 1 + t B ) = log ( D y t - A ) ##EQU00007##
[0258] Now estimate E by regression of log
( D y t - A ) ##EQU00008##
on log
( 1 + t B ) ##EQU00009##
for t selected so that
A + D 10 < y t < A + 9 D 10 ##EQU00010##
[0259] An alternative form of this model that leads to an almost
identical analysis (with its own initialization) is:
A + A s t + D ( 1 + exp ( - t - B E ) F ##EQU00011##
When A.sub.s=0 this is sometimes known as the Richards equation,
Richards (1959).
[0260] Estimation Criterion
[0261] Estimate parameters to minimize a penalized sum of squares
criterion:
t ( y t - g t ( .theta. ) ) 2 + .lamda. ( .theta. )
##EQU00012##
[0262] Here .lamda.(.theta.) is a non-negative function that
penalizes large values of some (or all) of the parameters in
.theta.. This method is known as regularization or ridge regression
(Hoerl, 1962) and may be derived from a Bayesian viewpoint by
setting a suitable prior distribution for the parameter vector
.theta.. A satisfactory choice for the penalty is:
.lamda.(.theta.)=.lamda.(B.sup.2+D.sup.2+E.sup.2+F.sup.2)
[0263] Large values of .lamda. bias the parameter estimates towards
zero and reduce the variance of the parameter estimates.
Conversely, small (or zero) .lamda. leads to unstable parameter
estimates and convergence difficulties in minimization algorithms.
The choice of .lamda. is a compromise between bias and variance or
stability. Empirical evidence shows that a satisfactory compromise
between bias and variance may be achieved if .lamda. is chosen in
the range:
0.01>.lamda.>0.0001.
This choice also ensures convergence of the optimization
algorithm.
[0264] Algorithm Choice
[0265] For any choice of .lamda. in the above range, the
description in the previous paragraph completely defines the
parameter estimates. A non-linear least squares procedure based on
the classical Gauss-Newton procedure (such as the
Levenberg-Marquardt algorithm as implemented in More, 1978) has
been successfully used and is a suitable approach. General purpose
optimizing algorithms such as Nelder and Mead, 1965, or
Broyden-Fletcher algorithm as implemented by Byrd, et al., 1995)
have also been successfully trialed in this context.
[0266] C.sub.p Estimate
[0267] C.sub.p is the point at time t that maximizes the second
derivative of g.sub.t (.theta.). Each fluorescence curve yields a
C.sub.p that characterizes the concentration of the target gene.
The average of the estimated C.sub.ps for each set of technical
replicates is computed and used in the subsequent analysis.
[0268] Standard Curves
[0269] Absolute or relative concentrations are derived from a
comparison with standard curves on the same PCR plate. Model a
dilution series using the linear model:
C.sub.p=R+S log.sub.10 Conc
where Conc is an absolute or relative concentration of the
standard. The intercept and slope parameters are plate specific.
Model the between-plate variability in the intercept and slope
parameters by setting population models
R.about.N(.mu..sub.R,.sigma..sub.R.sup.2)
S.about.N(.mu..sub.s,.sigma..sub.S.sup.2)
where the parameters .mu..sub.R, .sigma..sub.R.sup.2, .mu..sub.S,
.sigma..sub.S.sup.2 as are set on the basis of prior data as
described below. Then for a given plate R and S can be interpreted
as observations from these populations.
[0270] For replicate i of standard at concentration Conc.sub.j the
following model can be used:
C.sub.p(i,j)=R+S log.sub.10 Conc.sub.j+.di-elect cons..sub.ij
where .di-elect cons..sub.ij.about.N(0, .sigma..sub.j.sup.2). Note
that the variance of the residuals depends on C.sub.p. Empirical
estimates of Var(.di-elect cons..sub.ij) are given in Table 2.
Estimate the parameters R and S using by maximizing the likelihood
function. Interpret the slope parameter in terms of the efficiency
of the PCR process through the expression:
S = - 1 log 10 Efficiency ##EQU00013##
[0271] This model has a Bayesian interpretation: Give vague
(non-informative) prior distributions to the parameters .mu..sub.R,
.sigma..sub.R.sup.2, .mu..sub.S, .sigma..sub.S.sup.2. Then the
population models for R and S and for C.sub.p(i,j) fully determine
a probability model for the prior data. A Markov chain Monte Carlo
(MCMC) algorithm (Lunn et al., 2009) allows estimation of
.mu..sub.R, .sigma..sub.R.sup.2, .mu..sub.S, .sigma..sub.S.sup.2.
If the prior distribution is omitted, a traditional frequentist
interpretation results. Following this estimation procedure it is
possible to obtain the gene-dependent population parameter
estimates in Table 3.
TABLE-US-00003 TABLE 2 Variance of Residuals C.sub.p .sigma..sup.2
12 0.0100 13 0.0108 14 0.0119 15 0.0134 16 0.0155 17 0.0184 18
0.0224 19 0.0279 20 0.0356 21 0.0466 22 0.0625 23 0.0860 24 0.1212
25 0.1750 26 0.2591 27 0.3931 28 0.6112 29 0.9741
TABLE-US-00004 TABLE 3 Population Parameters for Slopes and
Intercepts of Standard Curves .mu..sub.R .sigma..sub.R.sup.2
.mu..sub.S .sigma..sub.S.sup.2 MDK 19.49 0.5112 -3.426 0.0481 CDC
18.91 0.2343 -3.414 0.0198 IL8R 31.43 0.0919 -3.192 0.0017 IGF
20.63 0.3835 -3.275 0.0247 HOXA 22.51 0.1544 -3.270 0.0037
[0272] The estimates of intercept and slope of the standard curve
are denoted by {circumflex over (R)} and S.
[0273] Relative Concentrations .DELTA.C.sub.p
[0274] Use the standard curve to compute C.sub.p(REF) at the
concentration Conc.sub.REF from the expression:
C.sub.p(REF)={circumflex over (R)}+S log.sub.10 Conc.sub.REF. The
relative concentration of a sample is given by the expression:
.DELTA. C p = C p - C p ( REF ) S ^ = log 10 Conc SAMPLE Conc REF
##EQU00014##
[0275] Alternatively S may be approximated at a fixed level
corresponding to a PCR efficiency of 2. Then
S=-1/log.sub.10(2)=-3.32. Use the same notation .DELTA.C.sub.P for
either choice. The resulting .DELTA.C.sub.P estimates, one for each
gene, are inputs to the discriminant function in the next step.
[0276] Discriminant Function
[0277] The .DELTA.C.sub.P values correspond to a relative biomarker
value with plate-to-plate variation removed. Examination of the 5
.DELTA.C.sub.P values in comparison with each other (for example,
see FIG. 2), shows how tumor samples typically have different
biomarker values than non-tumor samples. Furthermore, while there
is overlap in the areas for tumor and normal, a large number of
samples are effectively well separated. Under these circumstances,
many different statistical classifiers could be used to separate
the normal from the tumor samples. We show here that a sample of
several classifiers do work to separate these samples. We used 5
different classification methods: 1) Linear Discriminant Analysis
(LDA), 2) Logistic Regression (Log Reg); 3) Support Vector Machines
(SVM); 4) K-nearest-neighbor (KNN) based on 5 neighbors (KNSN); and
5) Recursive partitioning trees (TREE) (Cite: Venables & Ripley
and Dalgaard).
[0278] Creation of a classifier requires a dataset containing the
biomarker values for a large number of samples which should
represent the ultimate population to be tested by the classifier.
For example, if a classifier is to be used for screening an at-risk
population (eg age 50 and older, smokers), then the set of data
required for creating the classifier (called the "training set")
should mirror that population and contain only samples from people
older than 50 who smoke. Typically to obtain measurement precision
of smaller than 10% error for parameters like sensitivity and
specificity, the training set needs to be larger than 300
samples.
[0279] Estimation of the effectiveness of a classifier can be made
using cross-validation. In cross-validation (Wikipedia:
Cross-validation), the dataset is divided into a small number of
equally sized partitions (typically 3 to 10). One section is left
out and the remaining sections used to build a classifier; then the
left out section is tested by the new classifier and its
predictions noted. This is done for each section in turn and all
the predictions combined and analysed to compute the
characteristics of the classifier: Sensitivity, Specificity, etc.
If the cross-validation is performed by partitioning the data into
10 parts, it is called 10-fold cross-validation; similarly, 3 parts
would be 3-fold cross-validation. If the data are partitioned into
as many classes are there are samples, this is called "leave one
out cross-validation". By testing on data not used to build the
classifier, this method provides an estimate of the classifier
performance in the absence of additional samples.
[0280] We have built classifiers using all 15 combinations of 4
biomarkers, MDK, IGFBP5, CDC2, and HOXA13, all with and without the
IL8Rb biomarker, using the clinical trial dataset described
elsewhere in this document (Example 1) and tested those 30
classifiers using 10-fold cross-validation. This was done for each
of the 5 classifier types listed above and the ROC curves computed.
All work was performed using the R Statistical Programming
Environment (CITE). These results (FIG. 14) show that in most
cases, the classifier with IL8Rb is more sensitive for values of
specificity which are useful diagnostically (False Positive Rate of
0 to 20%; Specificity from 100 to 80%). The Area Under the Curve
(AUC) for the region with diagnostic utility of specificities is
used to quantify how well classifiers perform with larger values
indicating better classifier performance. FIG. 15a tabulates the
AUC for each classifier and biomarker combination, while FIG. 15b
shows the amount of increase in AUC for each condition when IL8Rb
is added. In most cases, the addition of IL8Rb improves the ability
to make accurate diagnoses. Specific sensitivity values for
diagnostically useful specificity values are tabulated for all the
classifiers in FIGS. 16a-16e. In addition, FIGS. 17a-17j tabulate
the amount of gain in sensitivity or specificity which the addition
of IL8Rb provides.
[0281] The utility of the classifier is created when, having
created it and tested it, it is used to test a new sample. To
simplify the interpretation of results, a cut-off score or
threshold is established; samples on one side of the cut-off are
considered positive and on the other side, negative for tumors.
Additional cut-offs may be established for example to indicate
increasing levels of certainty of results. In this case, we have
established a cut-off which gives a false positive rate of 15% in
our training set. Using our cross-validated ROC curves, we can then
estimate our sensitivity. Typically, we also establish a cut-off at
a positive predictive value of 75%. To use these cut-offs we
establish a "negative" result for scores less than the cut-off
established by the 85% specificity. Scores greater than the 75% PPV
are called "positive" and score between the two are called
"indeterminate" " or "suspicious".
Antibodies to IL8Rb
[0282] In additional aspects, this invention includes manufacture
of antibodies against IL8Rb. The marker IL8Rb can be produced in
sufficient amount to be suitable for eliciting an immunological
response. In some cases, a full-length IL8Rb can be used, and in
others, a peptide fragment of a IL8Rb may be sufficient as an
immunogen. The immunogen can be injected into a suitable host
(e.g., mouse, rabbit, etc) and if desired, an adjuvant, such as
Freund's complete adjuvant or Freund's incomplete adjuvant can be
injected to increase the immune response. It can be appreciated
that making antibodies is routine in the immunological arts and
need not be described herein further. As a result, one can produce
antibodies, including monoclonal or phage-display antibodies,
against IL8Rb.
[0283] In yet further embodiments, antibodies can be made against
the protein or the protein core of the tumor markers identified
herein or against an oligonucleotide sequence unique to a IL8Rb.
Although certain proteins can be glycosylated, variations in the
pattern of glycosylation can, in certain circumstances, lead to
mis-detection of forms of IL8Rb that lack usual glycosylation
patterns. Thus, in certain aspects of this invention, IL8Rb
immunogens can include deglycosylated IL8Rb or deglycosylated IL8Rb
fragments. Deglycosylation can be accomplished using one or more
glycosidases known in the art. Alternatively, IL8Rb cDNA can be
expressed in glycosylation-deficient cell lines, such as
prokaryotic cell lines, including E. coli and the like.
[0284] Expression vectors can be made having IL8Rb-encoding
oligonucleotides therein. Many such vectors can be based on
standard vectors known in the art. Vectors can be used to transfect
a variety of cell lines to produce IL8Rb-producing cell lines,
which can be used to produce desired quantities of IL8Rb for
development of specific antibodies or other reagents for detection
of IL8Rb or for standardizing developed assays for IL8Rb.
Kits
[0285] Based on the discoveries of this invention, several types of
test kits can be envisioned and produced. First, kits can be made
that have a detection device pre-loaded with a detection molecule
(or "capture reagent"). In embodiments for detection of IL8Rb mRNA,
such devices can comprise a substrate (e.g., glass, silicon,
quartz, metal, etc) on which oligonucleotides as capture reagents
that hybridize with the mRNA to be detected is bound. In some
embodiments, direct detection of mRNA can be accomplished by
hybridizing mRNA (labeled with cy3, cy5, radiolabel or other label)
to the oligonucleotides on the substrate. In other embodiments,
detection of mRNA can be accomplished by first making complementary
DNA (cDNA) to the desired mRNA. Then, labeled cDNA can be
hybridized to the oligonucleotides on the substrate and
detected.
[0286] Antibodies can also be used in kits as capture reagents. In
some embodiments, a substrate (e.g., a multi-well plate) can have a
specific IL8Rb and BTM capture reagents attached thereto. In some
embodiments, a kit can have a blocking reagent included. Blocking
reagents can be used to reduce non-specific binding. For example,
non-specific oligonucleotide binding can be reduced using excess
DNA from any convenient source that does not contain IL8Rb and BTM
oligonucleotides, such as salmon sperm DNA. Non-specific antibody
binding can be reduced using an excess of a blocking protein such
as serum albumin. It can be appreciated that numerous methods for
detecting oligonucleotides and proteins are known in the art, and
any strategy that can specifically detect marker associated
molecules can be used and be considered within the scope of this
invention.
[0287] Antibodies can also be used when bound to a solid support,
for example using an antibody chip, which would allow for the
detection of multiple markers with a single chip.
[0288] In addition to a substrate, a test kit can comprise capture
reagents (such as probes), washing solutions (e.g., SSC, other
salts, buffers, detergents and the like), as well as detection
moieties (e.g., cy3, cy5, radiolabels, and the like). Kits can also
include instructions for use and a package.
[0289] Detection of IL8Rb and BTMs in a sample can be performed
using any suitable technique, and can include, but are not limited
to, oligonucleotide probes, qPCR or antibodies raised against
cancer markers.
[0290] It will be appreciated that the sample to be tested is not
restricted to a sample of the tissue suspected of being an
inflammatory disease or tumor. The marker may be secreted into the
serum or other body fluid. Therefore, a sample can include any
bodily sample, and includes biopsies, blood, serum, peritoneal
washes, cerebrospinal fluid, urine and stool samples.
[0291] It will also be appreciate that the present invention is not
restricted to the detection of cancer in humans, but is suitable
for the detection of cancer in any animal, including, but not
limited to dogs, cats, horses, cattle, sheep, deer, pigs and any
other animal known to get cancer.
General Tests for Inflammatory Disease or Cancer Markers in Body
Fluids
[0292] In general, methods for assaying for oligonucleotides,
proteins and peptides in these fluids are known in the art.
Detection of oligonucleotides can be carried out using
hybridization methods such as Northern blots, Southern blots or
microarray methods, or qPCR. Methods for detecting proteins include
such as enzyme linked immunosorbent assays (ELISA), protein chips
having antibodies, suspension beads radioimmunoassay (RIA), Western
blotting and lectin binding. However, for purposes of illustration,
fluid levels of a disease markers can be quantified using a
sandwich-type enzyme-linked immunosorbent assay (ELISA). For plasma
assays, a 5 uL aliquot of a properly diluted sample or serially
diluted standard marker and 75 uL of peroxidase-conjugated
anti-human marker antibody are added to wells of a microtiter
plate. After a 30 minute incubation period at 30.degree. C., the
wells are washed with 0.05% Tween 20 in phosphate-buffered saline
(PBS) to remove unbound antibody. Bound complexes of marker and
anti-marker antibody are then incubated with o-phenylendiamine
containing H.sub.20.sub.2 for 15 minutes at 30.degree. C. The
reaction is stopped by adding 1 M H.sub.2SO.sub.4, and the
absorbance at 492 nm is measured with a microtiter plate
reader.
[0293] It can be appreciated that anti-IL8Rb antibodies can be
monoclonal antibodies or polyclonal antisera. It can also be
appreciated that any other body fluid can be suitably studied.
[0294] It is not necessary for a marker to be secreted, in a
physiological sense, to be useful. Rather, any mechanism by which a
marker protein or gene enters the serum can be effective in
producing a detectable, quantifiable level of the marker. Thus,
normal secretion of soluble proteins from cells, sloughing of
membrane proteins from plasma membranes, secretion of alternatively
spliced forms of mRNA or proteins expressed therefrom, cell death
(either apoptotic) can produce sufficient levels of the marker to
be useful.
[0295] There is increasing support for the use of serum markers as
tools to diagnose and/or evaluate efficacy of therapy for a variety
of cancer types.
EXAMPLES
[0296] The examples described herein are for purposes of
illustrating embodiments of the invention. Other embodiments,
methods and types of analyses are within the scope of persons of
ordinary skill in the molecular diagnostic arts and need not be
described in detail hereon. Other embodiments within the scope of
the art are considered to be part of this invention.
Example 1: Genotypic Analysis of Bladder Cancer
[0297] Methods
[0298] Patients:
[0299] Between April 2008 and September 2009, 485 patients
presenting with macroscopic hematuria, but no prior history of
urinary tract malignancy, were recruited at eleven urology clinics
in New Zealand and Australia. Each patient provided a urine sample
immediately prior to undergoing cystoscopy and any additional
diagnostic procedures. A diagnosis was made by three months
following enrollment in the study. Of these 485 patients, gene
expression data on all five study genes was successfully obtained
for 442 patients using the methods described below. The
characteristics of these patients are shown in Table 4.
TABLE-US-00005 TABLE 4 Characteristics of the Study Population I
Diagnosis Number Benign prostatic hyperplasia 18 Cystitis 18
Exercise-induced hematuria 3 Non-specific kidney disease 3
Non-specific neoplasia 3 Non-specific prostate disease 63 Vascular
prostate 49 Other urological cancer (non-TCC) 5 Superficial vessels
3 Urethral stricture 6 Urinary tract infection 18 Urolithiasis 25
Warfarin use 10 Unknown etiology 155 Miscellaneous 7 TCC 56 Total
442
[0300] Table 4 shows the number of patients in each of the main
diagnostic categories at three months after the patient's initial
presentation with gross hematuria.
[0301] Urine Analysis:
[0302] Urine samples were analyzed by central review cytology
(Southern Community Laboratories, Dunedin, New Zealand). The
diagnostic tests NMP22 BladderChek.RTM. (Matritech) and NMP22 ELISA
(Matritech) were carried out according to the manufacturer's
instructions at the clinical site (BladderChek.RTM.) or by Southern
Community Laboratories (NMP22 ELISA).
[0303] RNA Quantification:
[0304] 2 mls or urine from each patient was mixed with RNA
extraction buffer containing 5.64M guanidine thiocyanate, 0.5%
sarkosyl and 50 mM NaoAc pH6.5. Total RNA was then extracted by
Trizol extraction (Invitrogen) and the RNeasy procedure (Qiagen),
as previously described1. RNA was eluted from the columns in 35 ul
water and 3 ul was used in each subsequent monoplex or duplex
quantitative reverse transcription polymerase chain reaction
(qRT-PCR) assay. Each 16 ul qRT-PCR reaction contained 0.3 U
RNAse-OUT (Invitrogen), 0.225 uM each Taqman probe, 1.25 U
Superscript III (Invitrogen), 0.275 uM each primer, 1.5 U Fast
Start Taq polymerase (Roche), 10 mM DTT, 0.375 mM dNTPs, 4.5 mM
MgSO4, 1.6 ul 10.times.Fast Start PCR buffer (Roche) and 2.6 ul GC
Rich solution (Roche). Primers and fluorescently dual-labeled
probes were obtained from Integrated DNA Technologies (Coralville
USA) for each of the five study genes: MDK, CDC2, HOXA13, IGFBP5
and IL8Rb. Primer/probe sequences are shown in Table 2. Reactions
were set up in 96 well plates and cycled as follows on a Roche
Light Cycler.RTM. 480: 50.degree. C., 15 mins; 95.degree. C. 8
mins; 10 cycles of 95.degree. C. 15 sec, 60.degree. C. 2 mins and
30 cycles of 95.degree. C. 15 secs, 60.degree. C. 1 min. Standard
curves of 1/16 serial dilutions of a reference RNA (derived from
pooled cell line RNAs) were included on each plate to generate
range of 0.3 pg/.mu.l to 20 ng/.mu.l. Data was collected at the
extension phase of the final 30 thermocycles and exported as a raw
text file. Table 5 below depicts primers and probe sequences used
for qRT-PCR quantification of the five RNA markers.
TABLE-US-00006 TABLE 5 Marker Forward Seq Reverse Seq Probe MDK TGC
ACC CCC TGA TTA AAG CTA ACG AGC CCT TCC CTT TCT AAG ACC AAA AGA CAG
AA TGG CTT TGG CCT TT (SEQ ID NO. 3) (SEQ ID NO. 4) (SEQ ID NO. 5)
IGFBP5 CGT TGT ACC GGG ACG CAT CAC TCA ACG AAG AGA AAG CAG TGC CCA
ATT TT TGC AAA CCT TCC GTG A (SEQ ID NO. 7) CGT (SEQ ID NO. 8) (SEQ
ID NO. 6) CDC2 GCC GCC GCG TGT CTA CCC TTA TAC ACA AGC CGG GAT CTA
GAA TAA T ACT CCA TAG G CCA TAC CCA TTG (SEQ ID NO. 9) (SEQ ID NO.
10) ACT AAC T (SEQ ID NO. 11) HOXA13 TGG AAC GGC TGG CGT ATT CCC
GTT CAA ACT CTG CCC GAC CAA ATG TAC TG GT GTG GTC TCC CA (SEQ ID
NO. 12) (SEQ ID NO. 13) (SEQ ID NO. 14) IL8Rb CCT TGA GGC CCT GTA
GGA CAC CTC CAG TGG CCA CTC CAA ACA GTG AAG AAG AG TAA CAG CAG GTC
ACA TC (SEQ ID NO. 16) ACA (SEQ ID NO. 15) (SEQ ID NO. 17)
[0305] qRT-PCR Data Analysis
[0306] Raw fluorescence data was exported from the Roche
LightCycler.RTM. 480 as a tab-delimited file containing cycle
number versus two channels of fluorescence data for all wells on
the plate. The data were processed using an R program that applied
color compensation (Bernard 1999) to the data to correct for bleed
over from one fluorescent channel into another. It then fitted a
5-point logistic model to estimate the C.sub.P using the second
derivative maximum (Spiess 2008).
[0307] All samples and controls were applied in duplicate to the
PCR plates. The C.sub.P values from the duplicate wells were
averaged before use. If the difference between the two C.sub.P
values exceeded 3 units, that sample was repeated. To provide
standardization across PCR plates, C.sub.p's were expressed as
.DELTA.C.sub.P's relative to a reference RNA (derived from pooled
cell line RNAs) at 20 ng/.mu.l: .DELTA.C.sub.P=C.sub.P
(sample)-C.sub.P (reference RNA)
[0308] Statistical Analysis
[0309] qRT-PCR .DELTA.C.sub.P values from MDK, CDC2, HOXA13, IGFBP5
and IL8Rb were used to generate classifiers to separate samples
containing TCCs from samples containing no TCCs, based on Linear
Discriminant Analysis or Logistic Regression (Venables 2002). In
both cases, interactions between genes were permitted in the
classifier models. The generation of the LDA followed standard
procedures, as described, for example in "Modern Applied Statistics
with S, 4th edition" by W. N. Venables and B. D. Ripley (2002),
Springer. The dataset from the study was cleaned of any incomplete
data then the R Statistical Environment (R Development Core Team
(2009) and the function "lda" from the package MASS (Venables and
Ripley (2002)) were used to generate and test the linear
discriminant on the clinical trial data.
[0310] The generation of the Logistic Regression classifier was
performed in a similar manner to the generation of the LDA. Again,
the study data was cleaned of incomplete data. A logistic
regression classifier was created using R; no additional packages
were required. Logistic regression was performed as described by
Dalgaard (2008). Comparison among classifiers was made using ROC
curves, using the R package, ROCR (Sing et al. 2009). Confidence
intervals for ROC curves were generated using the methods of
Macskassy et al (Macskassy 2005). The following algorithms were
generated:
[0311] Linear Discriminant Classifier
[0312] The first classifier, a linear discriminant, (called LDA-3),
Is based on five gene values (normalized to a Reference value by
subtracting the reference value) allowing for multiway interactions
between the genes. The classifier was built in R using the `lda( )`
function from the package called "MASS". (R version 2.9.1; MASS
version 7.2-49). The classifier was built using the following
equation:
lda3<-lda(TCC.YN.about.MDK*IGF*CDC*HOXA*IL8R,data=uRNA.Trial)
Where lda3 is the created model; TCC.YN is the true value for
"presence of TCC in urine" (Yes or No) as determined by cystoscopy;
MDK, IGF, CDC, HOXA and IL8R are the normalized gene Cp value; and
uRNA Trial is a data file containing the Cp values for the each of
the 5 genes and TCC.YN (yes or no) from the clinical trial. Use of
the ASTERISK, `*`, in the formula signifies multiplication.
Evaluation of the classifier score takes as input a new data frame
containing the five gene values as well as the classifier, lda3, to
output a classifier score:
score<-c(predict(lda3,new.data)$x),
where "score" is the output used from the classifier to predict the
presence of TCCs; "1da3" is the classifier created above and
"new.data" is a data FILE containing the measured values of the
five genes called by the same names as used in classifier creation.
The syntax, `$x` and "c( . . . )" is present to extract the score
specifically from the large amount of information returned by the
predict function. Setting the score cut off to 0.112 and above,
sets our specificity to 85% for presence of TCCs in the urine
sample. The coefficients for LDA-3 are shown in Table 6.
TABLE-US-00007 TABLE 6 MDK.d.R100 5.333639e+00 IGF.d.R100
3.905978e+00 CDC.d.R100 6.877143e-01 HOXA.d.R100 6.073742e+00
IL8R.d.R100 -1.229466e+00 MDK.d.R100:IGF.d.R100 -7.420480e-01
MDK.d.R100:CDC.d.R100 -2.611158e-01 IGF.d.R100:CDC.d.R100
-1.965410e-01 MDK.d.R100:HOXA.d.R100 -8.491556e-01
IGF.d.R100:HOXA.d.R100 -4.037102e-01 CDC.d.R100:HOXA.d.R100
-3.429627e-01 MDK.d.R100:IL8R.d.R100 1.903118e-01
IGF.d.R100:IL8R.d.R100 2.684005e-01 CDC.d.R100:IL8R.d.R100
-1.229809e-01 HOXA.d.R100:IL8R.d.R100 2.909062e-01
MDK.d.R100:IGF.d.R100:CDC.d.R100 4.108895e-02
MDK.d.R100:IGF.d.R100:HOXA.d.R100 7.664999e-02
MDK.d.R100:CDC.d.R100:HOXA.d.R100 4.832034e-02
IGF.d.R100:CDC.d.R100:HOXA.d.R100 2.116340e-02
MDK.d.R100:IGF.d.R100:IL8R.d.R100 -3.750854e-02
MDK.d.R100:CDC.d.R100:IL8R.d.R100 1.664612e-02
IGF.d.R100:CDC.d.R100:IL8R.d.R100 2.089442e-03
MDK.d.R100:HOXA.d.R100:IL8R.d.R100 -1.539486e-02
IGF.d.R100:HOXA.d.R100:IL8R.d.R100 -3.894153e-02
CDC.d.R100:HOXA.d.R100:IL8R.d.R100 6.295032e-03
MDK.d.R100:IGF.d.R100:CDC.d.R100:HOXA.d.R100 -4.359738e-03
MDK.d.R100:IGF.d.R100:CDC.d.R100:IL8R.d.R100 -2.019317e-04
MDK.d.R100:IGF.d.R100:HOXA.d.R100:IL8R.d.R100 3.746882e-03
MDK.d.R100:CDC.d.R100:HOXA.d.R100:IL8R.d.R100 -2.902150e-03
IGF.d.R100:CDC.d.R100:HOXA.d.R100:IL8R.d.R100 4.799489e-04
MDK.d.R100:IGF.d.R100:CDC.d.R100:HOXA.d.R100:IL8R.d.R100
7.512308e-05
[0313] Logistic Regression Classifier
[0314] A second classifier based on Logistic Regression was derived
from the same cleaned dataset as LDA-3. Instead of using the lda( )
function, however, we used the glm( ) function from the package
stats (included with a base install of R) as shown below:
lr1<-glm(TCC.YN.about.CDC*IGF*HOXA*IL8R*MDK,
family=binomial("log it"),data=uRNA.Trial),
where "lr1" is the classifier created and the other parameters are
as described for the linear discriminant. Once again, full
interaction is specified using the operator. Classification is
performed in a manner very similar to that for LDA-3:
score<-predict(lr1,new.data,type=`response`),
where "score" is the value used to classify urine samples based on
the measurement of the five genes in "new.data", as above. The cut
off for lr1 is set to 0.102 to achieve a specificity of 85%; values
about the cut off are considered to be positive to TCCs. The
coefficients for the classifier are: [0315]
-103.0818143+3.9043769*CDC.d.R100+13.1120675*IGF.d.R100+17.4771819*HOXA.d-
.R100+-10.7711519*IL8R.d.R100+21.1027595*MDK.d.R100+-0.5938881*CDC.d.R100*-
IGF.d.R100+-1.0736184*CDC.d.R100*HOXA.d.R100+-1.3340189*IGF.d.R100*HOXA.d.-
R100+0.3126461*CDC.d.R100*IL8R.d.R100+1.4597355*IGF.d.R100*IL8R.d.R100+1.8-
739459*HOXA.d.R100*IL8R.d.R100+-1.035054*CDC.d.R100*MDK.d.R100+-2.5885156*-
IGF.d.R100*MDK.d.R100+-2.7013483*HOXA.d.R100*MDK.d.R100+1.4546134*IL8R.d.R-
100*MDK.d.R100+0.0767503*CDC.d.R100*IGF.d.R100*HOXA.d.R100+-0.0663361*CDC.-
d.R100*IGF.d.R100*IL8R.d.R100+-0.1015552*CDC.d.R100*HOXA.d.R100*IL8R.d.R10-
0+-0.2110656*IGF.d.R100*HOXA.d.R100*IL8R.d.R100+0.1361215*CDC.d.R100*IGF.d-
.R100*MDK.d.R100+0.1601118*CDC.d.R100*HOXA.d.R100*MDK.d.R100+0.259745*IGF.-
d.R100*HOXA.d.R100*MDK.d.R100+-0.0106468*CDC.d.R100*IL8R.d.R100*MDK.d.R100-
+-0.1947899*IGF.d.R100*IL8R.d.R100*MDK.d.R100+-0.185286*HOXA.d.R100*IL8R.d-
.R100*MDK.d.R100+0.0136603*CDC.d.R100*IGF.d.R100*HOXA.d.R100*IL8R.d.R100+--
0.0151368*CDC.d.R100*IGF.d.R100*HOXA.d.R100*MDK.d.R100+0.0056651*CDC.d.R10-
0*IGF.d.R100*IL8R.d.R100*MDK.d.R100+0.0030538*CDC.d.R100*HOXA.d.R100*IL8R.-
d.R100*MDK.d.R100+0.0232556*IGF.d.R100*HOXA.d.R100*IL8R.d.R100*MDK.d.R100+-
-0.000867*CDC.d.R100*IGF.d.R100*HOXA.d.R100*IL8R.d.R100*MDK.d.R100
Results
[0316] qRT-PCR Analysis of Urine Samples
[0317] To obtain an overview of the effect of IL8Rb on TCC
detection, two dimensional scatter plots were constructed using
qRT-PCR data obtained from the urine of patients with either TCC
(n=56) or the non-malignant conditions urolithiasis (n=25), urinary
tract infection (n=18) or cystitis (n=18). The scatter plots were
constructed using pairs of genes from a four gene signature (MDK,
CDC2, HOXA13, IGFBP5). IL8Rb was then substituted for one gene of
each pair and the data re-plotted. These plots are shown in FIGS.
2a-2f. Substitution of IL8Rb for IGFBP5 and HOXA13 in plots with
MDK (FIGS. 2a-2c) showed improved separation between samples from
patients with TCC and those with non-malignant conditions. The same
trend was observed in plots with CDC2 in which IL8Rb was
substituted for IGFBP5 and HOXA13 (FIGS. 2d-2f).
[0318] The contribution of IL8Rb to the correct diagnosis of TCC in
patients presenting with gross hematuria was then quantified by ROC
curve analysis. qRT-PCR data for each gene in the signature (MDK,
CDC2, IGFBP5 and HOXA13) and IL8Rb was used to develop linear
discriminate algorithms that maximized the discrimination between
the patients with TCC and those without. Two linear discriminate
algorithms were developed using the entire cohort of 442 samples:
LD1, which used the qRT-PCR data from MDK, CDC2, HOXA13 and IGFBP5
and LD2, which used MDK, CDC2, HOXA13, IGFBP5 and IL8Rb. LD1 and
LD2 were then used to generate ROC curves showing the sensitivity
and specificity of TCC detection in the group of patients with
confirmed TCC (n=56) or the non-malignant conditions urolithiasis
(n=25), urinary tract infection (n=18) or cystitis (n=18). FIG. 3a
shows the ROC curves for LD1 and LD2. The area under the ROC curve
for LD1 was 78% compared to 84% for LD2.
[0319] As an alternative to linear discriminate analysis, logistic
regression was used as an independent method to develop an
algorithm for the discrimination between patients with TCC and
those with non-malignant disease. As for the linear discriminate
analysis, the logistic regression algorithms were developed using
the entire cohort of 442 samples. The ROC curves obtained using
logistic regression and the 56 TCC and 61 non-malignant samples
described above are shown in FIG. 3b. The area under the ROC curve
for LR1 (obtained using qRT-PCR data from MDK, CDC2, HOXA13 and
IGFBP5) was 80% compared to 86% for LR2 (obtained using qRT-PCR
data from MDK, CDC2, HOXA13, IGFBP5 and IL8Rb). This data clearly
illustrates that inclusion of IL8Rb in methods for the detection of
TCC using urine samples can lead to improved discrimination between
patients with TCC and non-malignant diseases such as cystitis,
urinary tract infection and urolithiasis.
[0320] To confirm the improved accuracy afforded by IL8Rb for the
discrimination between patients with TCC and urolithiasis, urinary
tract infection or cystitis was maintained in an unselected cohort
of patients comprising a larger number and diversity of
non-malignant patients, the ROC curve analyses were repeated with
the entire cohort of 442 samples described in Table 1. In this
analysis, the area under the curve for LD1 and LD2 was 86 and 89%,
respectively (FIG. 4a). Similarly, the area under the curve for LR1
was 87% and for LR2 91% (FIG. 4b). This result confirms that IL8Rb
leads to improved accuracy in the detection of TCC using urine
samples.
[0321] This improvement in cancer detection due to the inclusion of
IL8Rb was further illustrated by applying LD1/LD2 and LR1/LR2 to
the 442 patient cohort and then determining the sensitivity of
detection of stage Ta TCC alone. Stage Ta tumors are smaller, more
differentiated tumors that are typically more difficult to detect
than higher stage tumors. LD1 detected 18/31 (58%) of the Ta tumors
compared to 19/31 (61%) for LD2 at a specificity of 85%. LR1
detected 21/31 (68%) compared to 24/31 (77%) for LR2 (specificity
of 85%). This data shows that the inclusion of IL8Rb into the LD
and LR algorithms increased the sensitivity of detection of stage
Ta tumors by up to 9%. In comparison to these RNA tests, the three
other bladder cancer tests in this study showed markedly lower
accuracy for the detection of Ta tumors: urine cytology (39%
sensitivity, 94% specificity), NMP22 ELISA (35% sensitivity, 88%
specificity) and NMP22 (BladderChek.RTM. "a registered trademark of
Matritech, Inc. of Massachusetts, United States") (39% sensitivity,
96% specificity).
IL8Rb as an Aid in the Diagnosis of Inflammation of the Urinary
Tract
[0322] To determine the ability of IL8Rb to be used in the
diagnosis of patients with inflammation of the urinary tract due to
causes such as cystitis or urinary tract infections, the urine
levels of IL8Rb mRNA in hematuria patients diagnosed with benign
prostate hyperplasia, non-specific prostate disease, vascular
prostate, hematuria secondary to warfarin use, and cystitis/urinary
tract infection were determined by qRT-PCR. The mean IL8Rb
.DELTA.Ct levels for each of these conditions were -3.12, -3.10,
-2.84, -1.98 and -5.27, respectively. The difference between the
mean of the IL8Rb level in patients with cystitis/urinary tract
infection and the other non-malignant states combined was
determined to be significant (p=0.001) using the Wilcoxon rank sum
test. Box plots portraying this data are shown in FIG. 5. This data
shows an elevation of IL8Rb levels in the majority of patients
diagnosed with either cystitis or urinary tract infection compared
to the other non-malignant conditions examined. Overlap between
plots is likely to be explained by a combination of three factors:
(i) the inability of standard clinical practice to correctly
diagnose each condition, (ii) co-morbidity (e.g., infection and
benign prostate hyperplasia), and (iii) the normal association of
high urine neutrophil counts in a subset of patients with benign
prostate hyperplasia, non-specific prostate disease, vascular
prostate or hematuria secondary to warfarin use. Regardless, given
the strict association between inflammation and neutrophil numbers,
the quantification of IL8Rb in urine provides an accurate method of
detecting inflammation of the urinary tract, be it as a consequence
of infection or in association with other non-malignant
conditions.
Example 2
[0323] Methods
[0324] Study Population
[0325] A consecutive series of patients without a prior history of
TCC were recruited prospectively from nine urology clinics in New
Zealand and two in Australia between 28 Apr. 2008 and 11 Aug. 2009.
The patient set included the patients used in example 1, but
included an additional 46 patients, whose data was not available
for the first analysis. The further studying also includes further
analysis of the results obtained. The samples were collected and
RNA collected and tested as described in Example 1.
[0326] RNA Test Development
[0327] uRNA.RTM. consists of four mRNA markers, CDC2, HOXA13, MDK
and IGFBP5. These markers were selected on the basis of their low
expression in blood and inflammatory cells and over-expression in
TCC..sup.2 In this cohort study, we prospectively specified a
linear discriminate algorithm (uRNA-D) that combined the four
markers into a single score. uRNA-D was independent, being
developed on an earlier dataset. It was not however, derived using
a strictly characterized patient group representing the intended
target population for the test. As a consequence, the study
protocol also defined the development of a new algorithm
(Classifier-D) for the use of the five markers CDC2, HOXA13, MDK,
IGFBP5 and IL8Rb using data obtained from the patients recruited to
the current cohort study.
[0328] In addition to Classifier-D, a second algorithm
(Classifier-S) was derived using the cohort study data to enable
identification of tumors that were either of advancing stage
(.gtoreq.stage 1) or high grade (WHO/ISUP 1998 classification).
Algorithm-S comprised all five markers, including CDC2 and HOXA13
which had previously been shown to be differentially expressed
between Stage Ta tumors and those .gtoreq.stage 1.
[0329] Classifier Development
[0330] Development of two classifiers for the use of the five
markers CDC2, HOXA13, MDK, IGFBP5 and IL8Rb (Classifier-D and
Classifier-S) were based on data obtained in this study, in
accordance with the methods outlined in this specification.
Briefly, logistic regression models were made using the statistical
programming environment, R (R Development Core Team (2011). R: A
language and environment for statistical computing. R Foundation
for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL
http://www.R-project.org/). Models made using .DELTA.C.sub.P values
for each of the five markers and their two way interactions (e.g.,
MDK.times.CDC2, MDK.times.IGFBP5, etc) were evaluated for their
ability to classify; those with the lowest AIC values were
evaluated in a leave-one-out cross validation procedure for their
sensitivity when the specificity was set to 85%. Several models
demonstrated comparable performance for each of Classifier-D and
Classifier-S, with the model with the fewest numbers of parameters
being selected.
[0331] Statistical Methods
[0332] Where a diagnostic test was specified in the protocol,
proportions and 95% confidence intervals were calculated for
sensitivity and specificity. Receiver operating characteristic
(ROC) curves were plotted and compared using the Stata roctab and
roccomp commands (Statacorp and Delong). For Classifier-D
confidence intervals are not appropriate, but Fishers exact or Chi
squared tests (where sample sizes allow) were used to test for an
association between TCC or patient characteristics and chances of
true positive or false positive results. Logistic regression models
were used to explore factors associated with false positive and
false negative results. All analyses were carried out in Stata
version 11.2.
Results
[0333] A total of 517 patients were initially recruited to the
study; 4% of patients were excluded because they were found to be
ineligible (n=10), did not undergo cystoscopy (n=9), TCC status was
not stated (n=2) or they did not provide an acceptable urine sample
(n=2) (FIG. 8). A further 10 patients were excluded from the
analysis because they did not have results for one or more of the
urine tests. The baseline demographic and clinical characteristics
of the 485 remaining patients are shown in FIG. 9.
[0334] The prevalence of TCC in the cohort was 13.6%. Two were
missing a review stage (both were Ta by local review) and two were
not given a review grade (one was grade 1 by the local pathologist,
the other low). Of the 66 tumors, 55 were superficial (stage Ta, T1
or Tis) and 11 were muscle invasive (T2). No patients had
detectable metastases or involvement of regional lymph nodes. Using
the 1973 grading system, 24 were classified as grade 3, 38 grade 2,
three grade 1 and one unknown. With the WHO98 system, 29 were
classified as high grade, four mixed, 32 low grade and one unknown.
In addition to the TCCs, two patients were diagnosed with a
papilloma, and seven with other neoplasms (five of these
urological).
[0335] The cutoff for the uRNA-D test was determined on the study
cohort, with specificity set at 85%. With this cutoff, uRNA-D
detected 41 of the 66 TCC cases (sensitivity of 62%), compared with
NMP22.TM. ELISA (50%), Bladderchek.RTM. (38%) and cytology (56%).
The RNA test developed on the cohort data Classifier-D detected 54
of the TCC cases (82%) at a specificity of 85% and 48 (73%) at a
specificity of 90%. uRNA-D and NMP22.TM. ELISA values can be
directly compared as both tests were fully specified prior to the
study. FIG. 21 shows the ROC curves; the area under the curves
(AUCs) are 0.81 and 0.73 respectively (p=0.03). The ROC curve for
Classifier-D was 0.87 (FIG. 21), and the improvement in performance
relative to uRNA-D appears to be mostly in the range of clinically
relevant specificities (above 80%).
[0336] Overall, Classifier-D detected 97% of the high/grade 3
tumors, compared to uRNA-D (83%), cytology (83%), NMP22 ELISA (69%)
and Bladderchek.RTM. (38%). Classifier-D was also more sensitive
for the detection of low-grade tumors (69%), with the other tests
ranging from 28-41% (FIG. 12) Classifier-D was positive for all the
TCC cases of Stage .gtoreq.1 plus both Tis, but the sensitivity was
68% for stage Ta (p=0.016, FIG. 12). This was still substantially
higher than the other tests, with uRNA-D being the next highest at
41%. TCC patients with macrohematuria or microhematuria evident in
their urine sample were more likely to have their TCC detected by
including IL8Rb than those without macrohematuria or microhematuria
(p<0.0005), though this is likely to be at least partially a
result of the higher proportion of high stage and grade TCCs among
those with macrohematuria or microhematuria. Numbers were
insufficient to explore this further in regression analyses.
[0337] Of the 12 cases missed by Classifier-D, all were stage Ta
and all except one were low grade (WHO ISUP 1998). Only two of the
twelve (both low grade, stage Ta TCC) were picked up by another
test (one by both NMP22.TM. ELISA and BladderChek.RTM. and one by
uRNA-D). Of the 12 cases missed by Classifier-D, all were stage Ta
and all except one were low grade (WHO ISUP 1998). Only two of the
twelve (both low grade, stage Ta TCC) were picked up by another
test (one by both NMP22.TM. ELISA and BladderChek.RTM. and one by
uRNA-D). Cytology did not pick up any TCCs that Classifier-D
missed.
[0338] Patient A: High Grade renal pelvic T2 tumour, no concurrent
Tis, no size given.
[0339] Patient B: High grade Bladder T3a no concurrent Tis,
2.times.3 cm
[0340] Patient C: a high grade tumour measuring 4.8.times.5.6 cm
with extensive stromal and muscularis propria invasion, extending
to the perivescical fat with no evidence of metastasis.
[0341] The specificity of the urine tests among those with
alternative diagnoses and according to urine sample characteristics
are shown in FIG. 13. Control patients with macrohematuria or
microhematuria were more likely to have false positive tests than
those without macrohematuria or microhematuria (p=0.002), and there
was a trend that patients with calculi may as well, although the
differences in specificity by diagnosis were not statistically
significant overall (p=0.12). There were five patients with other
urological cancers; only one of these gave a positive Classifier-D
test result. Results from fitting logistic regression models were
similar. In a logistic regression model with diagnosis and
macrohematuria or microhematuria, the association with
macrohematuria or microhematuria status remained significant
(p=0.006) and, when compared directly to no diagnosis those with
calculi had a 2.7 fold increased odds of a false positive test (95%
CI (1.1 to 6.4), p=0.03). Age did not affect the specificity of the
test.
[0342] Macrohematuria or microhematuria detected in the urine
sample was the only factor clearly associated with test
sensitivity. The predictive value of a positive test in this cohort
was 63% for those with macrohematuria or microhematuria and 24% for
those without, largely reflecting the greater prevalence of TCC in
the patients with macrohematuria or microhematuria (39% vs 6%).
[0343] There were 54 patients with TCC in whom the Classifier-D
test was positive. These patients were classified into severe and
less severe TCC using Classifier-S. Severe TCC was defined as stage
.gtoreq.1 or grade 3 at any stage. At a specificity of 90%,
Classifier-S correctly classified 32/35 (91%) of the severe TCC
cases.
Example 3: Combined Genotype and Phenotype Analysis of Patients
with Hematuria I
[0344] This study focuses on patients presenting with confirmed
asymptomatic microscopic hematuria (AH) who are undergoing a full
clinical work-up for the investigation of possible urothelial
cancer (UC). Approximately 500 patients are enrolled to participate
in the study.
[0345] As used herein, terms are defined below in additional
examples.
Objectives
[0346] Objectives are to determine the: (1) efficacy of a genotypic
and phenotypic algorithm in patients presenting with micro
hematuria who are scheduled for a full urological clinical work-up,
(2) performance characteristics (sensitivity, specificity, area
under the ROC curve, positive and negative predictive values) of
the genotypic and phenotypic algorithm G+P INDEX for the detection
of primary UC in patients presenting with confirmed microscopic
hematuria, and (3) number of patients correctly diagnosed as
negative for UC by a genotypic and phenotypic tool, the G+P INDEX
and therefore do not require investigative cystoscopy.
Study Population
[0347] The study population consists of patients presenting with
confirmed micro-hematuria who fulfill study requirements. Patients
are recruited from general practices that refer patients to the
urological clinics.
Informed Consent
[0348] Patients scheduled for investigative cystoscopy are
contacted to discuss possible participation. Patients are informed
of the nature of the study and consent is obtained. Patients
provide demographic, occupational, and smoking history information,
and ensure that they fully understand the patient information and
consent forms prior to provision of their urine samples. Study
coordinators complete a CRF page detailing the relevant inputs to
the genotypic and phenotypic index and transfer the data for
analysis.
Inclusion Criteria
[0349] Patients undergoing cystoscopic investigation for Urothelial
Carcinoma following a confirmed clinical finding of microscopic
hematuria (Minimum of 3 RBC per high power field (HPF)) on 2 or 3
properly collected urine specimens.sup.(3). [0350] Patients are
willing to comply with study requirements. [0351] Patients are over
18 years of age.
Exclusion Criteria
[0351] [0352] Prior history of urothelial caarcinoma (UC). [0353]
Current presentation of macroscopic hematuria [0354] Prior history
(past 12 months of an episode of Macroscopic Hematuria with
confirmed diagnosis (malignant or otherwise).
G+P Index
[0355] We developed a novel index, the "G+P Index," which comprises
of a combination of both genotypic and phenotypic data. The
Genotypic ("G") component utilizes RNA biomarker expression
information in conjunction with five clinical factors collected
from the patient in the same time window (Phenotypic data "P") to
determine the risk of UC in AH patients.
[0356] All patients receive a standard clinical work-up to
determine true clinical outcome and the outputs from this study are
simulated outputs based on the clinical data collected and the
genotypic data collected from the patients' urine samples. As such,
patient care is not altered as a result of the study output.
Patients provide urine samples, which are sent for genetic
analysis.
Patient Triage
[0357] No change to overall standard of care is made for patients
participating in the study. All patients scheduled for a full
urological work-up undertake the appropriate investigations
according to the current standard of care.
Study Data
[0358] Demographic and risk factor information are inputs to the
genotypic and phenotypic index. Final disease state (as determined
by flexible cystoscopy and follow up) are collated with (.PHI.)
results and demographic information and subjected to statistical
analysis.
Determination of the G+P Index
[0359] Using datasets obtained from samples collected from a large
number of sites in New Zealand and Australia from approximately 500
patients, a training model to predict the probability of `TCC=Yes`
was developed.
Data Collected for Variables Used in the Training and Validation
Populations
[0360] Phenotypic Variables [0361] Clinical findings are: Gender,
Age, Smoking history, and HFREQNEW (<=1 denoted as Low; >1
denoted as High).
[0362] Genotypic Variables
[0363] Genotypic Variables include expression of RNA markers: M1
(=MDK+CDC+IGBP5-HOXA13) and IL8R. Table 7 below shows estimates of
the coefficients of each of the factors in the validation G+P
INDEX:
TABLE-US-00008 TABLE 7 Analysis of Maximum Likelihood Estimates
Wald Standard Chi- Parameter DF Estimate Error Square Pr > ChiSq
Exp(Est) Intercept 1 -4.8445 0.6532 55.0092 <.0001 0.008 gender
2 1 -1.8544 0.6750 7.5484 0.0060 0.157 smoke 2 1 -0.9049 0.4537
3.9775 0.0461 0.405 HFREQNEW Low 1 -0.7015 0.3478 4.0681 0.0437
0.496 M1 1 1.2221 0.1553 61.9328 <.0001 3.394 IL8R 1 -0.2408
0.1789 1.8112 0.1784 0.786
[0364] FIG. 18 depicts ROC curves for the G+P Index. FIG. 18 is a
graph of sensitivity (vertical axis) versus 1-specificity
(horizontal axis) for results according to an embodiment of this
invention. For comparison, a diagonal line depicts the model. The
outcomes based solely upon Phenotypic information is shown as the
dash-dotted line, the outcome based solely upon Genotypic
information is shown as the dashed line, and the outcomes based on
the G+P Index are shown as the solid line. These results indicate
that the combination of Genotypic and Phenotypic information
provides an unexpected, substantial improvement in prediction of
outcome.
[0365] Exploratory models considered seven phenotypic variables,
but AgeGT50 showed insignificant effect while there was
insufficient data for RBC, so both of these variables were dropped
from the final model. Based on the significance level of the
remaining five phenotypic variables and the two RNA markers, an
index was constructed. Using relationship between M1, IL8R and
TCC(=yes) in the training dataset, a threshold of 4.5 and 2.5 was
used for M1 and IL8R respectively. A score of 5, 4, 3, 2 and 1 was
assigned to M1, Smokers, Male, IL8R, and HFREQ--which result in an
index score ranging from 0 to 15. The integrated algorithm based on
co-efficient is given below as the combined G+P index:
G+P INDEX=(1*HFREQ+3*Gender+4*SMK)+(5*M1+2*IL-8)
[0366] Odds ratios for different clinical factors that were
retained in the final model are shown in FIG. 19. An odds ratio can
be interpreted as having a harmful or protective effect upon the
subject depending on how far it deviates from 1 (i.e., no effect).
Odds ratios whose confidence limits exclude 1 are statistically
significant. Generally, the factors with higher odd ratio (e.g.
SMK, Gender) are assigned larger weights compared to factors with
small odds ratio (e.g. HFREQ).
[0367] The classification table for the full model is presented
below in Table 8.
TABLE-US-00009 TABLE 8 Classification Table Correct Incorrect
Percentages Prob Non- Non- False False Level Event Event Event
Event Correct Sensitivity Specificaty POS NEG 0.000 66 0 415 0 13.7
100.0 0.0 86.3 -- 0.050 61 245 170 5 63.6 92.4 59.0 73.6 2.0 0.100
54 319 96 12 77.5 81.8 76.9 64.0 3.6 0.150 47 358 57 19 84.2 71.2
86.3 54.8 5.0 0.200 45 376 39 21 87.5 68.2 90.6 46.4 5.3 0.250 40
381 34 26 87.5 60.6 91.8 45.9 6.4 0.030 37 391 24 29 89.0 56.1 94.2
39.3 6.9 0.350 36 398 17 30 90.2 54.5 95.9 32.1 7.0 0.400 35 402 13
31 90.9 53.0 96.9 27.1 7.2 0.450 35 403 12 31 91.1 53.0 97.1 25.5
7.1 0.500 31 407 8 35 91.1 47.0 98.1 20.5 7.9 0.550 29 408 7 37
90.0 43.9 98.3 19.4 8.3 0.600 29 408 7 37 90.9 42.4 98.6 17.6 8.5
0.650 28 409 6 38 90.9 42.4 98.6 17.6 8.5 0.700 24 409 6 42 90.0
36.4 98.6 20.0 9.3 0.750 20 409 6 46 89.2 30.3 98.6 23.1 10.1 0.800
12 410 5 54 87.7 18.2 98.9 29.4 11.6 0.850 11 411 4 55 87.7 16.7
99.0 26.7 11.8 0.900 6 412 3 60 86.9 9.1 99.3 33.3 12.7 0.950 1 413
2 65 86.1 1.5 99.5 66.7 13.6 1.000 0 415 0 66 86.3 0.0 100.0 --
13.7
[0368] Preliminary Validation Study of G+P Index
[0369] To further test the use of the G+P Index, we carried out
another study. Based on the statistical significance of various
clinical and RNA markers, an index was constructed. There were 98
subjects whose TCC status (yes or no) as well as G+P INDEX
variables were available.
[0370] A score of 5, 4, 3, 2 and 1 was assigned to M1 (genetic
tests), Smokers, Male, IL8R, and HFREQ--which results in an index
score ranging from 0 to 15. The number of true positives and true
negatives were 6 and 84 respectively. Similarly, the number of
false positives and false negatives were 5 and 3 respectively.
Thus, the overall accuracy of the proposed index was 0.92.
[0371] Implications and Follow Up Based on the G+P Index
[0372] If the G+P INDEX result indicates a "High Risk" of UC
defined as a score of 11-15 or above, the patient is prioritized
for a flexible cystoscopy and abdominal ultrasound as clinically
indicated.
[0373] If the G+P INDEX result indicates an "Moderate Risk" of UC,
defined as a score of from 6-10, the patient is reviewed and
followed up as per clinical practice. Consideration may be given to
the use of cytology, uretoscopy and/or a CT scan.
[0374] If the G+P INDEX result indicates a "Low Risk" of UC,
defined as a score of from 0-5, the patient will receive the normal
standard of care and be placed on the appropriate waiting list.
REFERENCES
[0375] The following references relate to the disclosure above.
[0376] Agresti, A. and Coull, B. A. (1998) Approximate is better
than "ex-act" for interval estimation of binomial proportions. The
American Statistician, 52, 119-126. [0377] Altekruse S F, Kosary C
L, Krapcho M, Neyman N, Aminou R, Waldron W, Ruhl J, Howlader N,
Tatalovich Z, Cho H, Mariotto A, Eisner M P, Lewis D R, Cronin K,
Chen H S, Feuer E J, Stinchcomb D G, Edwards B K (eds). SEER Cancer
Statistics Review, 1975-2007, National Cancer Institute. Bethesda,
Md., http://seer.cancer.gov/csr/1975_2007/, based on November 2009
SEER data submission, posted to the SEER web site, 2010. [0378]
Brown, L. D. Cal, T. T. and DasGupta, A. (2001) Interval estimation
for a binomial proportion. Statistical Science, 16, 101-133. [0379]
Buckhaults et al., (Cancer Research 61:6996-7001 (2002) describes
certain secreted and cell surface genes expressed in colorectal
tumours. [0380] Byrd, R. H., Lu, P., Nocedal, J. and Zhu, C. (1995)
A limited memory algorithm for bound constrained optimization. SIAM
J. Scientific Computing, 16, 1190-1208. [0381] Dalgaard (2008).
"Introductory Statistics with R, 2nd edition"; Peter Dalgaard,
Chapter 13 (2008), Springer, ISBN 978-0-387-79053-4. [0382] DeLong,
E. R., D. M. DeLong, and D. L. Clarke-Pearson. 1988. Comparing the
areas under two or more correlated receiver operating
characteristic curves: A nonparametric approach. Biometrics 44:
837-845. [0383] Gottschalk, P. G. and Dunn, J. R. (2005) The
five-parameter logistic: A characterization and comparison with the
four-parameter logistic. Analytic Biochemistry, 343, 54-65.
Doi:10.1016/j.ab0.2005.04.035 [0384] Grossfield G, Wolf J, Litwan
M, Hricak H, Shuler C, Agerter D, et al. Asymptomatic microscopic
hematuria in adults: summary of AUA best practice policy
recommendations. AFP 2001:63:1145-54. [0385] Hall et al
(Laryngoscope 113(1):77-81 (2003) (PMID: 12679418) (Abstract)
described predictive value of serum thyroglobulin in thyroid
cancer. [0386] Hoerl, A. E. (1962) Application of ridge analysis to
regression problems. Chemical Engineering Progress, 58, 54-59.
[0387] Holyoake A, O'Sullivan P, Pollock R, Best T, Watanabe J,
Kajita Y, et al. Development of a multiplex RNA urine test for the
detection and stratification of transitional cell carcinoma of the
bladder. Clin Cancer Res. 2008 Feb. 1; 14(3):742-9. [0388] Hotte et
al., (AJ. American Cancer Society 95(3):507-512 (2002) describes
plasma osteopontin as a protein detectable in human body fluids and
is associated with certain malignancies. [0389] Kim et al., (JAMA
287(13):1671-1679 (2002) describes osteopontin as a potential
diagnostic biomarker for ovarian cancer. [0390] Koopman et al.,
(Cancer Epidemiol. Biomarkers Pre.sup.y 13(3):487-491 (2004)
(Abstract) describes osteopontin as a biomarker for pancreatic
adenocarcinoma. [0391] Kuo et al (Clin. Chim. Acta.
294(1-2):157-168 (2000) (Abstract) describes serum matrix
metalloproteinase-2 and -9 in HCF- and HBV-infected patients.
[0392] Leman et al., (Urology, 69(4) 714-20 (2007) (Abstract)
describes EPCA-2 as a serum marker for prostate cancer. [0393] Loo
R, Lieberman S, Slezak J, Landa H, Mariani A, Nicolaisen G, Aspera
A and Jaconsen S: Stratifying risk of urinary tract malignant
tumors in patients with asymptomatic microscopic hematuria. Mayo
Clin Proc. 2013, 88(2); 129-138. [0394] Lunn, D., Spiegelhalter,
D., Thomas, A. and Best, N. (2009) The BUGS project: Evolution,
critique and future directions (with discussion), Statistics in
Medicine 28: 3049-3082. [0395] Marchi et al., (Cancer 112,
1313-1324 (2008) (Abstract) describes ProApolipoprotein Al as a
serum marker of brain metastases in lung cancer patients. [0396]
Martin et al., (Prostate Cancer Prostatic Dis. Mar. 9, 2004 (PMID:
15007379) (Abstract) described use of human kallikrein 2,
prostate-specific antigen (PSA) and free PSA as markers for
detection of prostate cancer. [0397] Mazzaferri et al., (J. Clin.
Endocrinol. Metab. 88(4):1433-1441 (2003) (Abstract) describes
thyroglobulin as a potential monitoring method for patients with
thyroid carcinoma. [0398] McDonald M, Swagerty D, Wetzel L:
Assessment of Microscopic hematuria in adults. AFP 2006 73:10.
[0399] Melle et al., (Clin. Chem. 53(4), 629-635 (2007) (Abstract)
describes HSP27 as a serum marker for pancreatic adenocarcinoma.
[0400] More, J. J. (1978) The Levenberg-Marquardt algorithm:
implementation and theory, in Lecture Notes in Mathematics 630:
Numerical Analysis, G. A. Watson (Ed.), Springer-Verlag: Berlin,
pp. 105-116. [0401] Nelder, J. A. and Mead, R. (1965) A simplex
algorithm for function minimization. Computer Journal 7, 308-31.
[0402] O'Sullivan P, Sharples K, Dalphin M et al: A Multigene Urine
Test for the Detection and Stratification of Bladder Cancer in
Patients Presenting with Hematuria. J Urol 2012, Vol 188 No 3; 746
[0403] Pellegrini et al., (Cancer Immunol. Immunother.
49(7):388-394 (2000) (Abstract) describes measurement of soluble
carcinoembryonic antigen and TIMP 1 as markers for pre-invasive
colorectal cancer.
[0404] (R Development Core Team (2009). R: A language and
environment for statistical computing. R Foundation for Statistical
Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://WWW.
R-project.org/) [0405] Richards, F. J. (1959) A flexible growth
function for empirical use. Journal of Experimental Botany. 10,
290-300. [0406] Rudland et al., (Cancer Rese arch 62: 3417-3427
(2002) describes osteopontin as a metastasis associated protein in
human breast cancer. [0407] Sing et al (2009). (Tobias Sing, Oliver
Sander, Niko Beerenwinkel and Thomas Lengauer (2009). ROCR:
Visualizing the performance of scoring classifiers. R package
version 1.0-4. http://CRAN.R-project.org/package=ROCR). [0408]
Speiss, A.-N., Feig, C. and Ritz, C. (2008) Highly accurate
sigmoidal fitting of real-time RCR data by introducing a parameter
for asymmetry. BMC Bioinformatics, 9, 221.
Doi:10.1186/1471-2105-9-211. [0409] Sprenger H, Lloyd A R, Lautens
L L, Bonner T I, Kelvin D J. Structure, genomic organization, and
expression of the human interleukin-8 receptor B gene. J Biol Chem.
1994 Apr. 15; 269(15):11065-72. [0410] StataCorp. 2009. Stata
Statistical Software: Release 11. College Station, Tex.: StataCorp
LP. [0411] Schwartz G: Proper evaluation of asymptomatic
microscopic hematuria in the era of evidence-based
medicine-progress is being made. Mayo Clin Proc. 2013, 88(2);
123-125. [0412] Trivedi C and Messing E M: Commentary: the role of
cytologic analysis of voided urine in the work-up of asymptomatic
microhematuria. BMC Urology 2009, 9:13 [0413] Tsigkou et al., (I
Clin Endocrinol Metab, 92(7) 2526-31 (2007) (Abstract) describes
total inhibin as a potential serum marker for ovarian cancer.
[0414] Tsigkou et al., (I Clin Endocrinol Metab, 92(7) 2526-31
(2007) (Abstract) describes total inhibin as a potential serum
marker for ovarian cancer. [0415] Venables, W. N. & Ripley, B.
D. (2002). Modern Applied Statistics with S. Fourth Edition.
Springer, New York. ISBN 0-387-95457-0). [0416] Whitley et al, (Dim
Lab. Med. 24(1):29-47 (2004) (Abstract) describes thyroglobulin as
a serum marker for thyroid carcinoma. [0417] Yoshikawa et al.,
(Cancer Letters, 151: 81-86 (2000) describes tissue inhibitor of
matrix metalloproteinase-1 in plasma of patients with gastric
cancer.
Example 4: Triage of Patients Presenting with Hematuria Using G+P
Index II
[0418] The G+P INDEX indicates a Positive when it takes values in
the range of 11 to 15.
Definitions
[0419] As used herein, the following definitions are used in this
and the following examples.
[0420] "AMH" means asymptomatic microhematuria;
[0421] "AUA" means American Urological Association;
[0422] "AUC" means area under the curve;
[0423] "CI" means confidence interval;
[0424] "CT" means computed tomography;
[0425] "ELISA" means enzyme-linked immunosorbent assay;
[0426] "FISH" means fluorescence in situ hybridization;
[0427] "HPF" means high-powered field;
[0428] "log OR" means log odds ratio;
[0429] "Hfreq" means average daily frequency of hematuria during
the most recent hematuria episode;
[0430] "ISUP" means International Society of Urological
Pathology;
[0431] "MRI" means magnetic resonance imaging;
[0432] "NPV" means negative predictive value;
[0433] "OR" means odds ratio;
[0434] "QC" means quality control;
[0435] "QoL" means quality of life;
[0436] "Phenotypic" is used to define clinical prognostic
characteristics and to distinguish them from gene expression-based
biomarkers that have been broadly defined as `genotypic`
variables.
[0437] "RBC" means red blood cell;
[0438] "ROC" means receiver operating curve;
[0439] "RT-qPCR" means quantitative reverse transcription
polymerase chain reaction;
[0440] "STARD" means Standards for Reporting of Diagnostic
Accuracy;
[0441] "UC" means urothelial carcinoma;
[0442] "WHO" means World Health Organization.
[0443] Introduction
[0444] Hematuria, which is most often associated with causes such
as benign prostatic enlargement, infection or urinary calculi, but
is also symptomatic of urothelial carcinoma (UC), is estimated to
occur in between 1 and 22% of patients in a general population
[1,2]. Macroscopic (macro-) hematuria is characterized by a visible
colour change in the urine of patients, while microscopic (micro-)
hematuria is defined more precisely as the presence of .gtoreq.3
red blood cells per high-powered field (RBCs/HPF) in three
concurrently collected urine samples [2]. The overall prevalence of
UC in patients with microhematuria has been reported to be
approximately 4%, whereas several studies have consistently shown
that the prevalence of UC is much higher in patients with
macrohematuria, ranging from approximately 12-23% [2-6], yet up to
four times as many patients with micro-versus macrohematuria
present for urological evaluation [7]. Notably, given that recent
changes to the American Urological Association (AUA) guidelines [2]
have seen the threshold for asymptomatic microhematuria (AMH)
lowered to .gtoreq.3 RBCs/HPF in a single sample, and even lower
thresholds (.gtoreq.1 RBC/HPF) have been proposed [8], a
consequential increase in the number of patients with hematuria who
will undergo a urological work-up to investigate potential UC and a
corresponding increase in the overall clinical and financial burden
of these patients on healthcare systems is expected.
[0445] Such hematuria-related referrals place a significant
clinical burden on urologists, as all patients must undergo a full
work-up to provide an often inconclusive diagnosis. Furthermore,
the existing diagnostic tests--many of which are invasive or have
high radiation loadings--can have a detrimental effect on patient
quality of life (QoL), especially if the patient receives repeated
cystoscopies as mandated in the current guidelines [2]. It has been
reported that for cystoscopies performed without prophylactic
antibiotics, 22% of patients had asymptomatic bacteriuria and 1.9%
of patients developed a febrile urinary tract infection (UTI)
within 30 days [9]. Other studies have also reported a high
prevalence of macrohematuria, pain on voiding and transient
erectile dysfunction in men following cystoscopy [10,11].
[0446] Healthcare systems also incur a significant financial burden
as a result of patients with hematuria undergoing a full urological
work up [12,13] and it has been concluded that urine cytology adds
costs without offering any significant diagnostic benefit [14-16].
Consequently, integrating an accurate, non-invasive test into the
primary clinical work-up of patients presenting with hematuria
allows physicians to effectively triage patients with hematuria,
thereby reducing the number of patients undergoing a full
urological work-up and investigative cystoscopy for UC, and offers
significant benefits to both patients and healthcare systems
[15-19].
[0447] Several clinical prognostic characteristics, including age,
gender, smoking history and degree of hematuria, are
well-established as risk factors for UC in patients with hematuria
[3,20-22]. Recently, several groups have attempted to develop
models based on clinical prognostic characteristics to predict the
risk of UC in patients with hematuria [20-22], but critically,
these models offer limited accuracy and have largely been focused
on detecting patients with UC rather than ruling out patients who
do not have disease. These detection-focused models have therefore
been insufficient to reliably identify patients with disease during
a primary evaluation, even if used in combination with urine
cytology [20-22].
[0448] Despite the higher incidence of UC in patients presenting
with macrohematuria, a number of studies show there is no
significant difference in the distribution of UC by grade and stage
in patients presenting with micro-compared with those presenting
with macrohematuria [5,23-25]. Therefore, the AUA recommends that
all patients with macrohematuria or AMH be referred to a urologist
for a full urological work-up, as severity of hematuria is not
sufficiently predictive for the presence of UC [2]. However, as
patients with hematuria may only undergo limited urinalysis in a
primary evaluation, consisting of cytology and in some cases
imaging studies, such as ultrasound, a full urological work-up is
often necessary to conclusively detect or rule out UC. While urine
cytology is specified in current guidelines and routinely used in
patients with suspected UC, cytology results are often inconclusive
with atypical or suspicious findings and also suffer from a low
diagnostic yield driven by a relatively high risk of false negative
results for patients with UC-related hematuria [2,26,27].
Consequently, it can be difficult to rule out benign causes of
hematuria, whether macrohematuria or AMH, during a primary
evaluation, especially if UC-related hematuria is intermittent and
appears to resolve following treatment for a benign cause [12].
[0449] A number of gene-based studies have set out to profile
urinary biomarkers in patients with UC, and these biomarkers may be
useful in their own right for detecting disease [28,29]. An
opportunity also exists to triage out patients on the basis of
their clinical characteristics and gene expression profile.
Combining NMP22 enzyme-linked immunosorbent assay (ELISA) tests or
a panel of gene markers with clinical characteristics has been
shown to improve diagnostic accuracy compared with clinical
characteristics alone, but these combined models have not yet
delivered significant advances in overall diagnostic accuracy,
especially when attempting to identify low-risk patients [30,31].
Nevertheless it is considered that incorporating clinical factors
and specific gene expression into a combined algorithm is likely to
provide the best guidance for diagnosing and managing patients with
hematuria or UC [32].
[0450] Cxbladder.TM. Detect (Pacific Edge Ltd., Dunedin, New
Zealand), a multigene test performed on unfractionated urine has
previously been shown to be more sensitive than urine cytology and
NMP22 for detecting UC in patients with macrohematuria [33] and
more accurate than urine cytology, NMP22 and fluorescence in situ
hybridization (FISH) in a comparative analysis (Kasabov, Darling,
Breen, et al., unpublished observations). Cxbladder Detect uses
quantitative reverse transcription polymerase chain reaction
(RT-qPCR) technology to quantify five mRNA markers, four markers
that are overexpressed in UC alongside a fifth marker that is
elevated in non-malignant inflammatory conditions, and offers a
high level of specificity and sensitivity when used to detect UC in
patients presenting with hematuria [33]. It was hypothesized that
an integrated model combining high-performance genetic biomarkers
with phenotypic variables collected from the same patients will
provide superior clinical resolution using high sensitivity (i.e. a
low probability of a patient with UC receiving a false negative
result), high negative predictive value (i.e. a high proportion of
all negative results being true) and a high test-negative rate to
enable the accurate triage of patients who have a low probability
of UC. These genotypic and phenotypic variables when combined into
a novel segregation model enable patients with hematuria who have a
low probability of UC to be identified and triaged, as opposed to
undergoing a full urological work-up.
[0451] Methods
[0452] Patient Selection
[0453] A prospective sample of 695 patients have been analysed,
where true clinical outcome was determined using a conventional
clinical evaluation. The study sample consists of an initial cohort
of patients with hematuria was consented and sampled as previously
described [33], where a consecutive series of 517 patients with a
recent history of macrohematuria, aged .gtoreq.45 years and without
a prior history of UC, were recruited prospectively from nine
urology clinics in Australia and New Zealand. These patients were
followed for three months for determination of UC status or
alternative diagnosis following multigene analysis of urine
samples, with a positive UC diagnosis being based on cystoscopical
appearance and histopathological examination. The stage of disease
was classified according to the TNM staging criteria determined by
pathology and diagnostic imaging investigations and tumor grade was
classified according to local pathology practice, using the 1998
World Health Organization (WHO)/International Society of Urological
Pathology (ISUP) consensus classification [34].
[0454] Additional cohorts of 94 and 84 patients undergoing
urological investigations following a macrohematuria event were
subsequently recruited from two centers in New Zealand between
March 2012 and April 2013 and included in the development of
models. Centers were selected to participate on the basis of their
previous experience participating in the initial study and their
willingness to evaluate the Cxbladder Detect product within
individual clinical settings.
[0455] An additional representative test set of 45 patients
presenting with microhematuria were prospectively collected and
used for further validation of the G+P INDEX, as set out below.
[0456] Eligibility criteria were similar to those of [33], except
that patients aged .gtoreq.18 years and those who had previously
undergone a cystoscopy to investigate UC that proved to be negative
were eligible for enrolment. Furthermore, as in [33], patients
exhibiting symptoms indicative of a UTI, or bladder or renal
calculi, were excluded.
[0457] Ethical approval for this study was granted by all
participating centers and informed consent obtained from all
patients providing samples.
[0458] Urine Sample Collection and Assessment
[0459] To provide gene expression data, a single mid-stream urine
sample was collected from participants using the Urine Sampling
System from Pacific Edge. Multigene analysis of samples from all
studies was carried out in accordance with the standard operating
procedure, as is used for the commercially available Cxbladder
Detect multigenic test. All urine samples (4.5 mL) from the initial
cohort were collected at a clinic prior to cystoscopy and
transferred to a stabilization liquid via vacuum driven aspiration
and sent to Pacific Edge within 48 hours. The samples were then
stored at -80.degree. C. until required for batch analysis. Samples
from the subsequent cohorts were collected in the same manner, but
shipped to Pacific Edge at ambient temperature and processed within
7 days of sample collection in accordance with revised quality
control (QC) limits and tolerance testing performed at the Pacific
Edge diagnostic laboratory.
[0460] Statistical Analysis
[0461] Univariate logistic regression was used to estimate the
unadjusted (raw) log odds ratio (log OR) co-efficients for four
binary phenotypic variables associated with UC: age, gender,
smoking history and average daily frequency of hematuria during the
patient's most recent hematuria episode (Hfreq; see Table 9).
TABLE-US-00010 TABLE 9 Definitions of binary phenotypic variables
associated with UC and their corresponding scores Score Phenotypic
parameter 0 1 Gender Female Male Age <60 years .gtoreq.60 years
Smoking history Never smoked Current or past smoker Hfreq .ltoreq.1
episode/day >1 episode/day
[0462] Multivariate logistic regression on all four phenotypic
variables was used to generate adjusted log OR co-efficients in the
phenotypic model (P INDEX).
[0463] G INDEX was developed using logistic regression to determine
the association between UC and mRNA concentrations for the five
Cxbladder.RTM. Detect genes (IGFBP5, HOXA13, MDK, CDK1 and CXCR2)
in urine samples. A multivariate genotypic-phenotypic model (G+P
INDEX) was generated using a combination of all nine variables from
the G INDEX and P INDEX. These linear models determined the log OR
from which the probability of a patient having UC was derived.
[0464] The relative performance of each of model was illustrated in
receiver operating curves (ROCs) plotting the false positive rate
versus the true positive rate when testing for UC, as determined by
each model. Area under the curve (AUC) was used to compare the
relative efficiency of each model with an AUC approaching 1 deemed
to be optimal.
[0465] To reduce potential bias when model estimation and
prediction are performed on the same data set, a bias-corrected AUC
was calculated for each of the three logistic regression models
using bootstrap resampling [35]. The difference between the nominal
AUC from the original sample and the average AUC from the bootstrap
samples is an estimate of the sample bias and the nominal AUCs were
adjusted accordingly. Bootstrap estimates of bias-corrected
confidence intervals (CIs) were also obtained [36].
[0466] Furthermore, it was a design criteria for this clinical test
that the performance characteristics of each model must exceed a
threshold NPV of 0.97, with as high a sensitivity as possible with
the further caveat of having a high test-negative rate. The test
negative rate is selected to provide a high clinical resolution
when triaging out patients presenting with hematuria who have a low
probability of having UC. Comparisons were made between the G
INDEX, P INDEX and G+P INDEX and the performance of each model was
determined in terms of sensitivity and NPV with a sufficiently high
test-negative rate to provide an effective tool for triaging out
patients with haematuria who have a low probability of UC.
[0467] Results
[0468] Sample Demographics
[0469] Of the 695 patients with macrohematuria registered across
the three cohorts, 23 were deemed to be ineligible and samples from
a further 85 patients were excluded after enrolment due to the
absence of sufficient data or samples failing to meet QC standards
(see FIG. 20A). In total, samples from 587 patients were available
for modelling comprising 72 UC-positive and 515 UC-negative
samples.
[0470] Of the 45 samples from patients with microhematuria
provided, 40 were suitable for analysis with 5 patients deemed
ineligible and excluded from the analysis (see FIG. 20B). All 45
patients had received a full urological evaluation and clinical
truth was confirmed as UC-negative. Full demographic data from both
sample populations is presented in Table 10.
TABLE-US-00011 TABLE 10 Sample population demographics for patients
with macro- and microhematuria with complete data Patients with
Patients with macrohematuria microhematuria Parameter (N = 587), n
(%) (N = 40), n (%) Age, years 0-49 65 (11.1) 21 (52.5) 50-59 111
(18.9) 60-69 145 (24.7) 19 (47.5) 70-79 175 (29.8) 80-100 91 (15.5)
Gender Female 113 (19.3) 25 (62.5) Male 474 (80.7) 15 (37.5)
Smoking history Never smoked 246 (41.9) 25 (62.5) Current or past
341 (58.1) 15 (37.5) smoker Hfreq .ltoreq.1 332 (56.6) 40 (100)
(episodes/day) >1 255 (43.4) -- Tumor stage Normal 515 (87.7) 40
(100) T1 16 (2.7) -- T2 11 (1.9) -- T3 2 (0.3) -- Ta 40 (6.8) --
Tis 3 (0.5) --
[0471] Relationship Between Phenotypic Variables and Risk of UC in
Patients with Macrohematuria
[0472] Unadjusted univariate logistic regression analyses of each
of the four binary phenotypic variables indicated that age
.gtoreq.60 years, male gender, a history of smoking and a high
frequency of macrohematuria were all associated with an increased
risk of UC (Table 11).
TABLE-US-00012 TABLE 11 Unadjusted and adjusted ORs for UC by
phenotypic and genotypic factors for patients with hematuria
Unadjusted Adjusted P Adjusted OR variable OR G + P variable
Phenotypic variables Control UC (95% CI) (95% CI) OR (95% CI) Age,
years <60 151 11 2.30 2.24 1.89 .gtoreq.60 364 61 (1.22-4.73)
(1.18-4.65) (0.85-4.64) Gender Female 105 8 2.05 1.58 3.03 Male 410
64 (1.01-4.75) (0.76-3.72) (1.12-9.36) Smoking Never 227 19 2.20
2.19 2.67 history smoked (1.29-3.91) (1.27-3.92) (1.34-5.64)
Current 288 53 or past smoker Hfreq (average .ltoreq.1 300 32 1.74
1.80 1.76 episodes/day) >1 215 40 (1.06-2.88) (1.08-3.00)
(0.93-3.35) Unadjusted Adjusted G Adjusted OR variable OR G + P
variable Genotypic variables (95% CI) (95% CI) OR (95% CI) IGFBP5
7.34 2.15 2.21 (4.59-12.33) (1.03-4.58) (1.03-4.83) HOXA13 6.27
0.33 0.20 (3.92-10.34) (0.13-0.83) (0.07-0.56) MDK 7.10 4.76 8.14
(4.73-11.10) (1.74-13.62) (2.64-26.60) CDK1 7.80 3.47 2.59
(5.11-12.39) (1.39-9.13) (0.98-7.18) CXCR2 1.69 0.65 0.69
(1.36-2.10) (0.45-0.92) (0.47-0.98)
Adjusted P INDEX, G INDEX and G+P INDEX variable ORs are the
exponentiated co-efficients in the P INDEX, G INDEX and G+P INDEX,
respectively. Adjusted log OR co-efficients were calculated in the
multivariate logistic regression model.
[0473] P
INDEX=-3.78+0.81.times.Age+0.46.times.Gender+0.78.times.Smoking
history+0.59.times.Hfreq, where each phenotypic variable is
assigned a binary score of 0 or 1, as designated in Table 9, and
the confidence intervals for the co-efficients are presented in
Table 11. The bias-corrected estimate for AUC for the P INDEX is
0.66 (95% CI: 0.55-0.67; FIG. 21).
[0474] Relationship Between Genotypic Variables and Risk of UC in
Patients with Macrohematuria
[0475] The G INDEX was estimated by logistic regression using the
log mRNA concentrations of the five genes IGFBP5, HOXA13, MDK, CDK1
and CXCR2 in urine samples to predict UC occurrence.
G
INDEX=-6.22+0.77.times.IGFBP5-1.11.times.HOXA13+1.56.times.MDK+1.24.ti-
mes.CDK1-0.43.times.CXCR2
[0476] The G INDEX gives a bias-corrected AUC of 0.83 (95% CI:
0.74-0.89; FIG. 21).
[0477] Relationship Between Genotypic and Phenotypic Variables and
Risk of UC in Patients with Macrohematuria
[0478] The five continuous genotypic variables were then combined
with the four binary phenotypic variables to estimate the G+P INDEX
using mulitvariate logistic regression.
G+P
INDEX=-8.46+(0.79.times.IGF-1.60.times.HOXA+2.10.times.MDK+0.95.time-
s.CDC-0.38.times.IL8R)+(0.64.times.Age+1.11.times.Gender+0.98.times.Smokin-
g history+0.56.times.Hfreq)
[0479] The G+P INDEX gives a bias-corrected AUC of 0.86 (95% CI:
0.80-0.91).
[0480] Comparison between G INDEX and G+P INDEX
[0481] There is overlap between the confidence intervals for the G
INDEX and G+P INDEX, so a bootstrap version of a paired test was
constructed by determining the difference in AUC for the G INDEX
and G+P INDEX for each bootstrap sample. Ten thousand bootstrap
samples with a sample size of n=587 were generated by random
sampling with replacement from the original 587 samples available
for analysis. The resulting 95% CI for the difference between
models was 0.01-0.08. Thus the probability that the true difference
between the two AUCs is less than 0.01 is <0.025, indicating
that there is a high likelihood of the AUC for the G+P INDEX being
significantly greater than the AUC for the G INDEX.
[0482] NPV and Sensitivity of Models
[0483] The G+P INDEX generated an NPV>0.97 over the range of
test-negative rates from 0.2 to 0.7 and was almost always higher
than the NPV for the G INDEX model (FIG. 22). The G+P INDEX offered
performance characteristics of sensitivity of 0.95 and NPV 0.98
when the test-negative rate was 0.4 (Table 12; FIG. 22). In
contrast, the G INDEX only achieved sensitivity of 0.86 and an NPV
of 0.96 when the test-negative rate was 0.4 (Table 12).
TABLE-US-00013 TABLE 12 Performance Characteristics of Each Model
When Thresholds Are Set For Varying Test Negative Rates Threshold
Test-negative NPV Sensitivity Specificity (logOR) rate (95% CI)
(95% CI) (95% CI) (95% CI) P INDEX -2.54 0.25 (0.21-0.28) 0.97
(0.92-0.99) 0.93 (0.85-0.98) 0.27 (0.23-0.31) -2.52 0.38
(0.34-0.42) 0.95 (0.91-0.97) 0.83 (0.74-0.91) 0.41 (0.37-0.45)
-2.39 0.42 (0.37-0.45) 0.95 (0.91-0.97) 0.82 (0.72-0.90) 0.45
(0.40-0.49) -1.95 0.51 (0.47-0.54) 0.92 (0.89-0.95) 0.68
(0.56-0.78) 0.53 (0.49-0.57) -1.93 0.51 (0.46-0.55) 0.92
(0.89-0.95) 0.68 (0.56-0.78) 0.53 (0.49-0.58) -1.73 0.82
(0.79-0.85) 0.90 (0.87-0.92) 0.32 (0.22-0.43) 0.84 (0.81-0.87) G
INDEX -3.46 0.20 (0.17-0.23) 0.94 (0.88-0.97) 0.90 (0.80-0.95) 0.22
(0.18-0.25) -3.23 0.30 (0.26-0.34) 0.95 (0.91-0.98) 0.89
(0.80-0.95) 0.33 (0.28-0.37) -3.04 0.40 (0.36-0.44) 0.96
(0.92-0.98) 0.86 (0.77-0.93) 0.44 (0.40-0.48) -2.86 0.50
(0.46-0.54) 0.97 (0.94-0.98) 0.86 (0.77-0.93) 0.55 (0.51-0.59)
-2.63 0.60 (0.56-0.63) 0.96 (0.94-0.98) 0.82 (0.71-0.90) 0.66
(0.62-0.69) -2.41 0.70 (0.66-0.73) 0.96 (0.94-0.98) 0.78
(0.65-0.86) 0.77 (0.73-0.80) G + P INDEX -4.02 0.20 (0.17-0.23)
0.97 (0.93-0.99) 0.96 (0.88-0.99) 0.22 (0.19-0.26) -3.67 0.30
(0.26-0.33) 0.98 (0.94-0.99) 0.94 (0.87-0.99) 0.33 (0.29-0.37)
-3.33 0.40 (0.36-0.44) 0.98 (0.95-1.00) 0.95 (0.86-0.98) 0.45
(0.40-0.49) -2.99 0.50 (0.46-0.54) 0.98 (0.96-0.99) 0.92
(0.83-0.97) 0.56 (0.52-0.60) -2.71 0.60 (0.56-0.64) 0.97
(0.95-0.99) 0.86 (0.76-0.93) 0.67 (0.63-0.71) -2.37 0.70
(0.66-0.73) 0.97 (0.94-0.98) 0.80 (0.70-0.88) 0.77 (0.73-0.80)
[0484] Application of the G+P INDEX in Patients with
Microhematuria
[0485] While the G+P INDEX was developed using data from patients
with macrohematuria, its robustness was tested in a further 40
samples from patients with microhematuria (Hfreq=0). A higher
test-negative rate was expected in a microhaematuria population as
the incidence of UC is lower in this population, and using a test
negative rate of 0.4, 32 (80%) patients tested negative and would
be correctly triaged out, therefore not requiring a full urological
work-up for the determination of UC.
Discussion
[0486] This study defines a clinical tool that offers clinicians
and physicians the ability to effectively triage-out patients
presenting with hematuria from the need to have a full urological
work-up for the detection of UC. The study presents an internally
validated genotypic-phenotypic model, G+P INDEX, with
bootstrap-based CI estimates, that offers a combination of high
sensitivity and high NPV (i.e. a low probability of an individual
patient with UC providing a false-negative result and a high
proportion of all negative results being true) that is not offered
by models derived exclusively from genotypic or phenotypic data
alone. This provides clinicians and physicians with a unique
opportunity to triage out patients with both micro- and
macrohematuria, in particular by identifying patients with a low
risk of having UC who do not require a full urological work up.
[0487] A high test-negative rate in the context of high sensitivity
is an important consideration for an effective triage-out test that
aims to direct patients with a low probability of UC away from a
full clinical work-up [37]. Accordingly, at a test-negative rate of
0.4 the sensitivity of the G+P INDEX presented here maximizes both
the sensitivity and NPV (0.95 and 0.98, respectively). This can be
compared with the best fit selected from the genotypic model
published in [33] (sensitivity=0.82; NPV=0.97) and is also
comparable with the sensitivity and NPV of both cystoscopy
(sensitivity=0.89-0.98; NPV=0.99) and virtual cystoscopy using
computed tomography (CT) scans or magnetic resonance imaging (MRI)
(sensitivity=0.94 and 0.91, respectively) [38-40].
[0488] It is acknowledged that the sample population used to derive
the G INDEX, P INDEX and G+P INDEX in this instance consisted of
patients with macrohematuria. However, the high sensitivity of this
test at a test negative rate of 0.4 in patients with macrohematuria
allows the G+P INDEX to be applied across both macro- and
microhematuria populations. Presuming that patients with and
without UC are similarly distributed amongst the micro- and
macrohematuria patient populations, but with an expected UC
prevalence of 4% in the microhematuria population, a high NPV can
also be expected in the microhematuria patient population.
[0489] By applying the G+P INDEX to the sample population of
patients with microhematuria who do not have UC it was shown that
80% of the patients would have been triaged out on the basis of the
result. Only 20% would be referred for a full urological work-up.
This compares with conventional guidelines that would currently see
all of the patients (100%) with microhematuria that cannot be
attributed to a benign cause undergoing a full urological work-up,
incurring significant unnecessary costs and negatively impacting
patient QoL.
[0490] Severity of hematuria is correlated with the probability of
a patient having UC, but not the stage or grade of any tumour, and
an estimated 96% and 77-88% of patients with micro- and
macrohematuria, respectively, referred to a urologist will not have
UC [2-6]. Therefore, avoiding potentially unnecessary urological
work-ups for patients with hematuria has several benefits.
Cystoscopy may be associated with adverse effects, such as pain on
voiding, bleeding, UTIs, male sexual dysfunction and the anxiety
that accompanies an inconclusive or unconfirmed UC diagnosis
[9-11]. Most notably, this novel approach has the potential to
reduce the burden on resources and the financial cost associated
with a full urological work-up on UC-negative patients. For
example, in the UK, avoiding cystoscopy in patients with hematuria
with an initial negative cytology and/or tumor biomarker test has
been estimated to save approximately US$770 per patient (.English
Pound.483 per patient) evaluated [13]. The G+P INDEX described here
provides an effective alternative to the use of urine cytology when
used in a primary evaluation setting. This is particularly relevant
in settings where primary evaluation is carried out by primary care
physicians.
[0491] On this basis, if we assign an arbitrary `nominal cost` of
US$4,500 for each full urological work up, the total cost for
working up 1,000 patients with microhematuria would approach US$4.5
million. In contrast, if 80% of patients with microhematuria are
triaged out using the G+P INDEX at an arbitrary nominal cost of
US$2,500, the total direct cost of testing and full urological
work-ups for the remaining 20% of patients would total US$3.4
million. This provides a notional net saving in direct costs of
approximately US$1.1 million per 1,000 patients with
microhematuria.
[0492] While the genotypic algorithm developed by O'Sullivan et al.
[33] comprised the same genotypic constituents as the G+P INDEX
presented here, the balance between sensitivity and specificity was
calibrated for the optimal primary detection of UC in symptomatic
patients (i.e. presenting with hematuria) who were undergoing a
full urological work-up. In contrast, the G+P INDEX in this study
also incorporated phenotypic variables and has been optimized for
high sensitivity and high NPV, to segregate out those patients with
hematuria who do not require a full urological work-up for
suspected UC. No attempt is made to define or select patients with
UC. Instead the aim is to confidently rule out those who do not
have UC, and as such, all patients not segregated out would
progress for a full urological work-up.
[0493] While several studies have previously sought to develop
predictive models that consider phenotype when assessing the risk
of UC in patients presenting with hematuria, the accuracy of
phenotype-dependent models alone appears to be limited. For
example, Loo et al. [21] prospectively investigated whether
phenotypic parameters could be used to identify patients with
microhematuria who may not have required a urological referral and
full work-up and concluded that age, male gender and a recent
diagnosis of macrohematuria were significant predictors of UC. A
history of smoking and >25 RBCs/HPF in a recent urinalysis were
not statistically significant predictors of UC, in isolation, but
even when included in their `Hematuria Risk Index` to improve
predictive accuracy, this index resulted in an AUC of 0.809 [21].
Interestingly, the phenotypic ORs in this study and those
identified by Loo et al. are comparable, with overlapping 95% CIs
for smoking history and gender, and while age, gender and smoking
history have similar weightings in each model, the influence of the
genotypic component of the G+P INDEX presented here is likely to
account for the higher AUC [21].
[0494] Likewise, Cha et al. [20] reported that age, smoking history
and degree of hematuria, but not gender, were significantly
correlated with the presence of UC in patients with asymptomatic
hematuria and used a multivariate model to develop a nomogram
comprised of phenotypic and urine cytology data for predicting UC.
As with Loo et al. [21], the reported phenotypic ORs are comparable
to those reported here, but even after incorporating urine cytology
into the nomogram, the AUC of 0.831 reported in [20] was lower than
that of the G+P INDEX.
[0495] In another study, Tan et al. [22] retrospectively stratified
patients with hematuria who had been referred to a specialist
urology clinic into high- and low-risk groups using a nomogram
derived from patient age, gender, smoking history and the degree of
hematuria. While comparisons with this study must be made with
caution given the high proportion of patients who were excluded due
to an absence of data (80 out of 405 patients), the AUC of 0.804,
sensitivity of 0.900 and NPV of 0.953 were all lower than the G+P
INDEX described here.
[0496] Several attempts have also been made to improve the accuracy
of phenotypic models by supplementing them with the results of
urinary biomarker tests. When the nuclear matrix protein NMP22
point of care proteomic assay is used in isolation to detect UC it
has a sensitivity of 0.557 and NPV of 0.968 [17]. Lotan et al. [41]
published a multivariable algorithm comprising phenotypic factors,
NMP22 and urine cytology with an AUC for predicting UC of 0.826
that was then prospectively validated with an AUC of 0.802 [31].
However, it is important to note that this model attempted to
discriminate between high-risk patients who did and did not have
UC, as opposed to maximizing sensitivity and NPV to triage-out
patients with a low probability of UC.
[0497] The improved accuracy obtained with algorithms comprising
both genotypic and phenotypic data have previously been
demonstrated in breast cancer, in particular [42-45]. Likewise,
Mitra et al. [30] used a combination of molecular markers and
smoking intensity to calculate a multivariate model that was
superior to routine clinicopathological parameters in predicting
survival in patients with UC. However, the present study is the
first to demonstrate that phenotypic risk factors can be combined
with genotypic data to increase the accuracy of a model for
separating patients with haematuria into categories requiring
differential levels of urological follow up and clinical care
rather than survivorship prediction.
[0498] When phenotypic data are combined with genotypic data in a
model, the resolution of data is likely to impact the accuracy of
the model. For example, smoking is a well understood risk factor
for UC and is included in most phenotypic models for detecting UC.
In Cha et al. [20], Tan et al. [22], Lotan et al. [31,41] and the
current study, the binary discriminants never smoked and
current/ex-smoker were used, whereas Mitra et al. [30] calculated
smoking intensity on the basis of years of smoking and number of
cigarettes smoked each day and Loo et al. [21] categorized smokers
into never smoked, passive smokers, smokers who had ceased and
current smokers. While it is known that the risk of UC increases
substantially with exposure to smoking [46], arbitrarily defining
phenotypic variables may limit the overall accuracy and utility of
phenotypic models. In contrast, an interaction between a patient's
genotypic and phenotypic variables would not be unexpected.
However, combining the impact of phenotypic factors and genetic
variables in a single tool improved the accuracy of the model
described in this study. A similar principle also applies to
describing hematuria phenotype. Patients presenting with micro- or
macrohematuria are essentially on a biological continuum and have
different likelihoods of having UC [2-6,21]. Accordingly, despite
all patients with microhematuria in this study having a Hfreq score
of 0, the severity of their hematuria, in combination with other
phenotypic factors, is likely to be indirectly accounted for in the
genotypic component of the G+P INDEX.
REFERENCES
[0499] The articles recited immediately below refer to Example 4,
and are all incorporated herein fully by reference. [0500] 1. Kelly
J D, Fawcett D P, Goldberg L C: Assessment and management of
non-visible haematuria in primary care. BMJ 2009, 338:a3021. [0501]
2. Davis R, Jones J S, Barocas D A, Castle E P, Lang E K, Leveillee
R J, Messing E M, Miller S D, Peterson A C, Turk TMT, Weitzel W:
Diagnosis, evaluation and follow-up of asymptomatic microhematuria
(AMH) in adults: AUA guideline. J Urol 2012, 188(6
Suppl):2473-2481. [0502] 3. Sutton J M. Evaluation of hematuria in
adults. JAMA 1990, 263:2475-2480. [0503] 4. Khadra M H, Pickard R
S, Charlton M, Powell P H, Neal D E: A prospective analysis of
1,930 patients with hematuria to evaluate clinical practice. J Urol
2000, 163:524-527. [0504] 5. Davidson P: Re-design of a haematuria
clinic: Assessment of 2346 haematuria patients. J Urol 2011,
185(4S):e495. [0505] 6. Price S J, Shephard E A, Stapley S A,
Barraclough K, Hamilton W T: Non-visible versus visible haematuria
and bladder cancer risk: A study of electronic records in primary
care. Br J Gen Pract 2014, 64:e584-e589. [0506] 7. Buteau A,
Seideman C A, Svatek R S, Youssef R F, Chakrabati G, Reed G, Bhat
D, Lotan Y: What is evaluation of hematuria by primary care
physicians? Use of electronic medical records to assess practice
patterns with intermediate follow-up. Urol Oncol 2014, 32:128-134.
[0507] 8. Jimbo M: Evaluation and management of hematuria. Prim
Care 2010, 373:461-472. [0508] 9. Herr H W: The risk of urinary
tract infection after flexible cystoscopy in bladder tumor patients
who did not receive prophylactic antibiotics. J Urol 2014 Jul. 18.
pii: S0022-5347(14)03963-9. [0509] 10. Burke D M, Shackley D C,
O'Reilly P H: The community-based morbidity of flexible cystoscopy.
BJU Int 2002, 89:347-349. [0510] 11. Stay K, Leibovici D, Goren E,
Livshitz A, Siegel Y I, Lindner A, Zisman A: Adverse effects of
cystoscopy and its impact on patients' quality of life and sexual
performance. Isr Med Assoc J 2004, 6:474-478. [0511] 12. Rao P K,
Jones J S: How to evaluate `dipstick hematuria`: What to do before
you refer. Cleve Clin J Med 2008, 75:227-233. [0512] 13. Rodgers M,
Nixon J, Hempel S, Aho T, Kelly J, Neal D, Duffy S, Ritchie G,
Kleijnen J, Westwood M: Diagnostic tests and algorithms used in the
investigation of haematuria: systematic reviews and economic
evaluation. Health Technol Assess 2006, 10:iii-iv, xi-259. [0513]
14. Falebita O A, Lee G, Sweeney P: Urine cytology in the
evaluation of urological malignancy revisited: is it still
necessary? Urol Int 2010, 84:45-49. [0514] 15. Feifer A H,
Steinberg J, Tanguay S, Aprikian A G, Brimo F, Kassouf W: Utility
of urine cytology in the workup of asymptomatic microscopic
hematuria in low-risk patients. Urology 2010, 75:1278-1282. [0515]
16. Svatek R S, Hollenbeck B K, Holmang S, Lee R, Kim S, Stenzl A,
Lotan Y. The economics of bladder cancer: Costs and considerations
of caring for this disease. Eur Urol 2014, 66:253-262. [0516] 17.
Grossman H B, Messing E, Soloway M, Tomera K, Katz G, Berger Y,
Shen Y: Detection of bladder cancer using a point-of-care proteomic
assay. JAMA 2005, 293:810-816. [0517] 18. Friedlander D F, Resnick
M J, You C, Bassett J, Yarlagadda V, Penson D F, Barocas D A:
Variation in the intensity of hematuria evaluation: a target for
primary care quality improvement. Am J Med 2014, 127:633-640.
[0518] 19. Shinagare A B, Silverman S G, Gershanik E F, Chang S L,
Khorasani R: Evaluating hematuria: impact of guideline adherence on
urologic cancer diagnosis. Am J Med 2014, 127:625-632. [0519] 20.
Cha E K, Tirsar L A, Schwentner C, Hennenlotter J, Christos P J,
Stenzl A, Mian C, Martini T, Pycha A, Shariat S F, Schmitz-Drager B
J: Accurate risk assessment of patients with asymptomatic hematuria
for the presence of bladder cancer. World J Urol 2012, 30:847-852.
[0520] 21. Loo R K, Lieberman S F, Slezak J M, Landa L M, Mariani A
J, Nicolaisen G, Aspera A M, Jacobsen S J: Stratifying risk of
urinary tract malignant tumors in patients with asymptomatic
microscopic hematuria. Mayo Clin Proc 2013, 88:129-138. [0521] 22.
Tan G H, Shah S A, Ann H S, Hemdan S N, Shen L C, Galib NAFA,
Singam P, Kong C H C, Hong G E, Bahadzor B, Zainuddin Z M:
Stratifying patients with haematuria into high or low risk groups
for bladder cancer: a novel clinical scoring system. Asian Pac J
Cancer Prev 2013, 14:6327-6330. [0522] 23. Sultana S, Goodman C,
Bryne D, Baxby K: Microscopic haematuria: urological investigation
using a standard protocol. Br J Urol 1996, 78:691-698. [0523] 24.
Sugimura K, Ikemoto S-I, Kawashima H, Nishisaka N, Kishimoto T:
Microscopic hematuria as a screening marker for urinary tract
malignancies. Int J Urol 2001, 8:1-5. [0524] 25. Viswanath S,
Zelhof B, Ho E, Sethia K, Mills R: Is routine urine cytology useful
in the haematuria clinic? Ann R Coll Surg Engl 2008, 90:153-155.
[0525] 26. Steiner H, Bergmeister M, Verdorfer I, Granig T, Mikuz
G, Bartsch G, Stoehr B, Brunner A: Early results of bladder-cancer
screening in a high-risk population of heavy smokers. BJU Int 2008,
102:291-296. [0526] 27. Yeung C, Dinh T, Lee J: The health
economics of bladder cancer: An updated review of the published
literature. Pharmacoeconomics 2014, 32:1093-1104. [0527] 28.
Holyoake A, O'Sullivan P, Pollock R, Best T, Watanabe J, Kajita Y,
Matsui Y, Ito M, Nishiyama H, Kerr N, da Silva Tatley F, Cambridge
L, Toro T, Ogawa O, Guilford P: Development of a multiplex RNA
urine test for the detection and stratification of transitional
cell carcinoma of the bladder. Clin Cancer Res 2008, 14:742-749.
[0528] 29. Sapre N, Anderson P D, Costello A J, Hovens C M,
Corcoran N M: Gene-based urinary biomarkers for bladder cancer: an
unfulfilled promise? Urol Oncol 2014, 32:48.e9-e17. [0529] 30.
Mitra A P, Castelao J E, Hawes D, Tsao-Wei D D, Jiang X, Shi S R,
Datar R H, Skinner E C, Stein J P, Groshen S, Yu M C, Ross R K,
Skinner D G, Cortessis V K, Cote R J: Combination of molecular
alterations and smoking intensity predicts bladder cancer outcome.
Cancer 2013, 119:756-765. [0530] 31. Lotan Y, Svatek R S, Krabbe L
M, Xylinas E, Klatte T, Shariat S F: Prospective external
validation of model for bladder cancer detection. J Urol 2014,
192:1343-1348. [0531] 32. Abogunrin F, O'Kane H F, Ruddock M W,
Stevenson M, Reid C N, O'Sullivan J M, Anderson N H, O'Rourke D,
Duggan B, Lamont J V, Boyd R E, Hamilton P, Nambirajan T,
Williamson K E: The impact of biomarkers in multi-variate
algorithms for bladder cancer diagnosis in patients with hematuria.
Cancer 2012, 118:2641-2650. [0532] 33. O'Sullivan P, Sharples K,
Dalphin M, Davidson P, Gilling P, Cambridge L, Harvey J, Toro T,
Giles N, Luxmanan C, Felipe Alves C, Yoon H S, Hinder V, Masters J,
Kennedy-Smith A, Beaven T, Guilford P: A multigene urine test for
the detection and stratification of bladder cancer in patients
presenting with hematuria. J Urol 2012, 188:741-747. [0533] 34.
Epstein J I, Amin M B, Reuter V R, Mostofi F K: The World Health
Organization/International Society of Urological Pathology
consensus classification of urothelial (transitional cell)
neoplasms of the urinary bladder. Bladder Consensus Conference
Committee. Am J Surg Pathol 1998, 22:1435-1448. [0534] 35. Efron B,
Tibshirani R J: An introduction to the bootstrap. New York: Chapman
Hill; 1993. [0535] 36. DiCiccio T J, Tibshirani R J: Bootstrap
confidence intervals. Statist Sci 1996, 11:189-228. [0536] 37.
Van't Hoog A H, Cobelens F, Vassall A, van Kampen S, Dorman S E,
Alland D, Ellner J: Optimal triage test characteristics to improve
the cost-effectiveness of the Xpert MTB/RIF assay for TB diagnosis:
A decision analysis. PLoS ONE 2013, 8:e82786. [0537] 38. Qu X,
Huang X, Wu L, Huang G, Ping X, Yan W: Comparison of virtual
cystoscopy and ultrasonography for bladder cancer detection: A
meta-analysis. Eur J Radiol 2010, 80:188-197. [0538] 39. Mowatt G,
N'Dow J, Vale L, Nabi G, Boachie C, Cook J A, Fraser C, Griffiths T
R; Aberdeen Technology Assessment Review (TAR) Group: Photodynamic
diagnosis of bladder cancer compared with white light cystoscopy:
Systematic review and meta-analysis. Int J Technol Assess Health
Care 2011, 27:3-10. [0539] 40. Blick C G, Nazir S A, Mallett S,
Turney B W, Onwu N N, Roberts I S, Crew J P, Cowan N C: Evaluation
of diagnostic strategies for bladder cancer using computed
tomography (CT) urography, flexible cystoscopy and voided urine
cytology: results for 778 patients from a hospital haematuria
clinic. BJU Int 2012, 110:84-94. [0540] 41. Lotan Y, Capitanio U,
Shariat S F, Hutterer G C, Karakiewicz P I: Impact of clinical
factors, including a point-of-care nuclear matrix protein-22 assay
and cytology, on bladder cancer detection. BJU Int 2009,
103:1368-1374. [0541] 42. Tyrer J, Duffy S W, Cuzick J: A breast
cancer prediction model incorporating familial and personal risk
factors. Stat Med 2004, 23:1111-1130. [0542] 43. Dubsky P, Brase J
C, Jakesz R, Rudas M, Singer C F, Greil R, Dietze O, Luisser I,
Klug E, Sedivy R, Bachner M, Mayr D, Schmidt M, Gehrmann M C, Petry
C, Weber K E, Fisch K, Kronenwett R, Gnant M, Filipits M, on behalf
the Austrian Breast and Colorectal Cancer Study Group (ABCSG): The
EndoPredict score provides prognostic information on late distant
metastases in ER+/HER2- breast cancer patients. Br J Cancer 2013,
109:2959-2964. [0543] 44. Brentnall A R, Evans D G, Cuzick J:
Distribution of breast cancer risk from SNPs and classical risk
factors in women of routine screening age in the UK. Br J Cancer
2014, 110:827-828. [0544] 45. Filipits M, Nielsen T O, Rudas M,
Greil R, Stoger H, Jakesz R, Bago-Horvath Z, Dietze O, Regitnig P,
Gruber-Rossipal C, Muller-Holzner E, Singer C F, Mlineritsch B,
Dubsky P, Bauernhofer T, Hubalek M, Knauer M, Trapl H, Fesl C,
Schaper C, Ferree S, Liu S, Cowens J W, Gnant M, Austrian Breast
and Colorectal Cancer Study Group: The PAM50 risk-of-recurrence
score predicts risk for late distant recurrence after endocrine
therapy in postmenopausal women with endocrine-responsive early
breast cancer. Clin Cancer Res 2014, 20:1298-1305. [0545] 46.
Krabbe L M, Svatek R S, Shariat S F, Messing E, Lotan Y: Bladder
Cancer Risk: Use of the PLCO and NLST to Identify as Suitable
Screening Cohort. Urol Oncol. 20149 July 16): pii: 51078-1439
(14)00215-4.
CONCLUSIONS
[0546] In conclusion, the G+P INDEX reported here shows a
significant opportunity to change clinical utility. G+P INDEX is
able to accurately triage out patients who present to their
clinician or physician with hematuria, who have a low probability
of UC with a high overall test-negative rate, high level of
sensitivity and high NPV. This model is suitable for use during
primary evaluation of patients with hematuria to triage out
patients who do not require a full urological work-up, thereby
potentially reducing the number of patients with hematuria
requiring referral for specialist urological evaluation for UC,
helping to maintain patient QoL and helping to reduce
diagnosis-related costs.
Example 5: Triage of Patents with Hematuria Using the G+P Index III
("G2")
[0547] A further use of the G+P INDEX is to categorize patients
according to Risk of UC. The risk category is then used to
prioritize patients for follow up investigation.
[0548] This example provides an alternative method for triaging
patients using an index, herein termed the "G2+P Index." It is
similar to the G+P Index described above in Example 4.
Definition of G.sub.2
[0549] This classifier calculates the following formula:
M 1 = [ IGF ] - [ HOXA ] + [ MDK ] + [ CDC ] ; then , R = - 6.9802
+ 0.2007 * M 1 + 1.7893 * [ IL 8 R ] + 0.1552 * M 1 2 - 0.2882 * [
IL 8 R ] 2 - 0.0720 * M 1 * [ IL 8 R ] ##EQU00015##
to obtain the score:
[0550] SCORE=e.sup.R/(1+e.sup.R), where [IGF], [HOXA], [MDK], [CDC]
and [IL8R] are the logarithms of the sample concentrations for the
genes IGF, HOXA, MDK, CDC and IL8R, respectively; is ordinary
multiplication, and e=2.718282 . . . is the base of the natural
(Napierian) logarithm.
[0551] A high SCORE indicates a higher likelihood of UC being
present. As an example, we may set a threshold of 0.12 and declare
a SCORE >=0.12 as having a high likelihood of having UC, and
those scores below 0.12 has having a low likelihood of having
UC.
Example 6: Concurrent Use of [G1=P]-Index and G2 to Triage Out
Patients Who Present with Hematuria Who have a Low Probability of
Urothelial Carcinoma
[0552] A. Patents with Micro-Haematuria
[0553] Patient Data-Set [0554] 45 patient samples were assayed
[0555] 5 patients lack Smoking Status and were excluded from this
analysis [0556] All patients have completed a full urological
work-up and none have urothelial carcinoma. [0557] Phenotypic
variables: Gender, Age, SmokingStatus [0558] Patient demographics
are shown in Table 13 below.
TABLE-US-00014 [0558] TABLE 13 Gender Age NonSmoker ExSmoker Smoker
Female Age < 60 10 1 2 Age > 60 9 1 2 Male Age < 60 3 2 3
Age > 60 3 2 2
Observed Test Negative Rate:
[0559] Using a Test Negative Rate of 40% as the threshold, data is
shown below in Table 14.
TABLE-US-00015 TABLE 14 TNR 40% LR < 0.12 0.12 <= LR <
0.23 0.23 <= LR Triage Negative 28 3 1 Triage Positive 4 2 2
[0560] Using a Negative Test Rate of 50% as the threshold, data is
shown below in Table 15.
TABLE-US-00016 TABLE 15 TNR 50% LR < 0.12 0.12 <= LR <
0.23 0.23 <= LR Triage Negative 28 3 3 Triage Positive 4 2 0
[0561] Using a Negative Test Rate of 60% as the threshold gives
identical results.
[0562] The highest risk group (Male, Current or Ex-Smoker, Age
>=60) were all positive for this Triage classification.
SUMMARY AND CONCLUSIONS
[0563] 1. By deed of the clinical guidelines all 41 patients would
normally receive a full urological work-up. [0564] 2. All 41 did
receive a full-work-up and all 41 were determined to have no
urothelial carcinoma. [0565] 3. At a Test Negative Rate (TNR of 50%
85% of the micro-hematuria patients would be screened out and
therefore consequentially not receive a full urological work-up.
[0566] 4. If Cxbladder-triage was used at a TNR of 50% (Triage
Index -3.33) 85.4% of the patients would be triaged out and
consequently, correctly would not receive a full urological
work-up. [0567] 5. If Cxbladder-triage was used at a TNR of 40%
(Triage Index -3.0) 80.5% of the patients would be triaged out and
consequently, correctly would not receive a full urological
work-up.
[0568] B. Patients with Macrohematuria
Patient Data Set
[0569] 587 samples from Clinical Trial data and North Shore and
CURT product trials were used. This data set was a subset
consisting of complete data containing Age, Gender, Smoking Status
and Haematuria frequency, as well as gene concentrations for IGF,
HOXA, MDK, CDC, IL8R.
[0570] We used the same data used to develop the Cxbladder-triage
model below;
G+P INDEX=-8.46+0.79 IGF-1.60 HOXA+2.10 MDK+0.95 CDC-0.38 IL8R+0.98
SNS+0.56 Hfreq+1.11 Gender+0.64 Age
[0571] We plotted the triage score against G2 diagnostic score in
FIG. 24. The Triage thresholds -3.33, -2.99 and -2.71 correspond to
Test Negative rates of 40%, 50% and 60% respectively. The verticals
are the Cxbladder thresholds of 0.12 and 0.23. Filled circles
correspond to tumours; green are Ta, and red are all other higher
stage tumours. Table 16 shows the clinical findings.
TABLE-US-00017 TABLE 16 Stage No-cancer T1 T2 T2a T3 Ta Tis Count
515 16 10 1 2 40 3
[0572] We considered the counts in the 4 quadrants of FIG. 24 and
determined by various cutoffs for Triage Index and G2.
[0573] The threshold for the Cxbladder-triage (Triage Index) of -3
corresponds to a Test Negative rate of 50%. Table 17 shows these
results.
TABLE-US-00018 TABLE 17 Quadrant Thresholds Control TCC Bottom Left
Triage < -3 AND G2 < 0.12 272 6 Top Left Triage > -3 AND
G2 < 0.12 155 7 Bottom Left Triage < -3 AND G2 >= 0.12 16
0 Top Right Triage > -3 AND G2 >= 0.12 72 59
[0574] Using a Cxbladder-triage threshold of -3.33 (Test Negative
Rate of 40%) with the same G2 threshold of 0.12, we observed the
data shown in Table 18.
TABLE-US-00019 TABLE 18 Quadrant Thresholds Control TCC Bottom Left
Triage < -3.33 AND G2 < 0.12 222 4 Top Left Triage > -3.33
AND G2 < 0.12 205 9 Bottom Right Triage < -3.33 AND G2 >=
0.12 9 0 Top Right Triage > -3.33 AND G2 >= 0.12 79 59
SUMMARY AND CONCLUSIONS
[0575] 1. Use of a serial combination of Cxbladder Triage (G+P) and
Cxbladder Detect (G2) on the same patients in the same time
interval provided a comprehensive segregation of patients into four
key clinical groupings.
[0576] 2. For a combined population of patients presenting with
micro and macro hematuria and using a Test Negative Rate of 40%
there was a total of 235 of the 587 patients (40.0%) of the
patients triaged out. We conclude that these patients will not need
a full work-up for UC.
[0577] 3. For the same population the corresponding residual group
of 352 (60%) patients will receive a full urological work-up.
[0578] 4. This residual group contained all high grade and late
stage tumours. These patients were not triaged out and would
therefore consequently, correctly receive a full urological
work-up.
[0579] 5. A total of 4 low-grade Ta's (5.6% of the total number of
tumours) will be triaged out and will not receive the full
work-up.
[0580] 6. If the triage rules allow triage out of all patients
below the 40% Test Negative Rate and with G2 scores <0.12 then
there were a total of 13 low grade Ta's triaged out that will not
receive a full work-up and all high grade late stage tumours will
be caught and receive a full urological work-up.
[0581] 7. These modified triage rules also resulted in a total of
440 patients triaged out of a total of 587 (75%).
ADVANTAGES AND GENERAL CONCLUSIONS
[0582] In conclusion, the G+P INDEX reported here shows a
significant opportunity to change clinical utility. G+P INDEX is
able to accurately triage out patients who present to their
clinician or physician with haematuria, who have a low probability
of UC with a high overall test-negative rate, high level of
sensitivity and high NPV. This model is suitable for use by primary
care physicians to triage out patients who do not require a full
urological work up, thereby reducing the number of patients with
haematuria requiring referral for specialist urological evaluation
for UC, helping to maintain patient quality of life and reducing
diagnosis-related costs. The disclosed methods provided
unexpectedly accurate assessment of the lack of need for follow-up
investigation for those with hematuria. These represent excellent
effects that could not have been achieved without use of the
disclosures contained herein.
INCORPORATION BY REFERENCE
[0583] All patents, patent applications and non-patent literature
citations are herein incorporated fully by reference as if
separately so incorporated.
INDUSTRIAL APPLICABILITY
[0584] Embodiments of this invention are useful in the fields of
healthcare and medicine.
Technical Arts
[0585] Embodiments of this invention provide highly accurate,
sensitive, and specific computer-impolemented methods for triaging
patients to determine which patients do not require substantial
short-term procedures or follow up. The methods improve computer
operations by providing new and non-obvious computer operations
based on analysis of specific genetic and phenotypic information
from patients with hematuria that yield tangible, useful, and
concrete results to improve the quality of health care and reduce
cost.
Sequence CWU 1
1
171360PRTHomo sapiens 1Met Glu Asp Phe Asn Met Glu Ser Asp Ser Phe
Glu Asp Phe Trp Lys1 5 10 15Gly Glu Asp Leu Ser Asn Tyr Ser Tyr Ser
Ser Thr Leu Pro Pro Phe 20 25 30Leu Leu Asp Ala Ala Pro Cys Glu Pro
Glu Ser Leu Glu Ile Asn Lys 35 40 45Tyr Phe Val Val Ile Ile Tyr Ala
Leu Val Phe Leu Leu Ser Leu Leu 50 55 60Gly Asn Ser Leu Val Met Leu
Val Ile Leu Tyr Ser Arg Val Gly Arg65 70 75 80Ser Val Thr Asp Val
Tyr Leu Leu Asn Leu Ala Leu Ala Asp Leu Leu 85 90 95Phe Ala Leu Thr
Leu Pro Ile Trp Ala Ala Ser Lys Val Asn Gly Trp 100 105 110Ile Phe
Gly Thr Phe Leu Cys Lys Val Val Ser Leu Leu Lys Glu Val 115 120
125Asn Phe Tyr Ser Gly Ile Leu Leu Leu Ala Cys Ile Ser Val Asp Arg
130 135 140Tyr Leu Ala Ile Val His Ala Thr Arg Thr Leu Thr Gln Lys
Arg Tyr145 150 155 160Leu Val Lys Phe Ile Cys Leu Ser Ile Trp Gly
Leu Ser Leu Leu Leu 165 170 175Ala Leu Pro Val Leu Leu Phe Arg Arg
Thr Val Tyr Ser Ser Asn Val 180 185 190Ser Pro Ala Cys Tyr Glu Asp
Met Gly Asn Asn Thr Ala Asn Trp Arg 195 200 205Met Leu Leu Arg Ile
Leu Pro Gln Ser Phe Gly Phe Ile Val Pro Leu 210 215 220Leu Ile Met
Leu Phe Cys Tyr Gly Phe Thr Leu Arg Thr Leu Phe Lys225 230 235
240Ala His Met Gly Gln Lys His Arg Ala Met Arg Val Ile Phe Ala Val
245 250 255Val Leu Ile Phe Leu Leu Cys Trp Leu Pro Tyr Asn Leu Val
Leu Leu 260 265 270Ala Asp Thr Leu Met Arg Thr Gln Val Ile Gln Glu
Thr Cys Glu Arg 275 280 285Arg Asn His Ile Asp Arg Ala Leu Asp Ala
Thr Glu Ile Leu Gly Ile 290 295 300Leu His Ser Cys Leu Asn Pro Leu
Ile Tyr Ala Phe Ile Gly Gln Lys305 310 315 320Phe Arg His Gly Leu
Leu Lys Ile Leu Ala Ile His Gly Leu Ile Ser 325 330 335Lys Asp Ser
Leu Pro Lys Asp Ser Arg Pro Ser Phe Val Gly Ser Ser 340 345 350Ser
Gly His Thr Ser Thr Thr Leu 355 36022880DNAHomo sapiens 2aggttcaaaa
cattcagaga cagaaggtgg atagacaaat ctccaccttc agactggtag 60gctcctccag
aagccatcag acaggaagat gtgaaaatcc ccagcactca tcccagaatc
120actaagtggc acctgtcctg ggccaaagtc ccaggacaga cctcattgtt
cctctgtggg 180aatacctccc caggagggca tcctggattt cccccttgca
acccaggtca gaagtttcat 240cgtcaaggtt gtttcatctt ttttttcctg
tctaacagct ctgactacca cccaaccttg 300aggcacagtg aagacatcgg
tggccactcc aataacagca ggtcacagct gctcttctgg 360aggtgtccta
caggtgaaaa gcccagcgac ccagtcagga tttaagttta cctcaaaaat
420ggaagatttt aacatggaga gtgacagctt tgaagatttc tggaaaggtg
aagatcttag 480taattacagt tacagctcta ccctgccccc ttttctacta
gatgccgccc catgtgaacc 540agaatccctg gaaatcaaca agtattttgt
ggtcattatc tatgccctgg tattcctgct 600gagcctgctg ggaaactccc
tcgtgatgct ggtcatctta tacagcaggg tcggccgctc 660cgtcactgat
gtctacctgc tgaacctagc cttggccgac ctactctttg ccctgacctt
720gcccatctgg gccgcctcca aggtgaatgg ctggattttt ggcacattcc
tgtgcaaggt 780ggtctcactc ctgaaggaag tcaacttcta tagtggcatc
ctgctactgg cctgcatcag 840tgtggaccgt tacctggcca ttgtccatgc
cacacgcaca ctgacccaga agcgctactt 900ggtcaaattc atatgtctca
gcatctgggg tctgtccttg ctcctggccc tgcctgtctt 960acttttccga
aggaccgtct actcatccaa tgttagccca gcctgctatg aggacatggg
1020caacaataca gcaaactggc ggatgctgtt acggatcctg ccccagtcct
ttggcttcat 1080cgtgccactg ctgatcatgc tgttctgcta cggattcacc
ctgcgtacgc tgtttaaggc 1140ccacatgggg cagaagcacc gggccatgcg
ggtcatcttt gctgtcgtcc tcatcttcct 1200gctctgctgg ctgccctaca
acctggtcct gctggcagac accctcatga ggacccaggt 1260gatccaggag
acctgtgagc gccgcaatca catcgaccgg gctctggatg ccaccgagat
1320tctgggcatc cttcacagct gcctcaaccc cctcatctac gccttcattg
gccagaagtt 1380tcgccatgga ctcctcaaga ttctagctat acatggcttg
atcagcaagg actccctgcc 1440caaagacagc aggccttcct ttgttggctc
ttcttcaggg cacacttcca ctactctcta 1500agacctcctg cctaagtgca
gccccgtggg gttcctccct tctcttcaca gtcacattcc 1560aagcctcatg
tccactggtt cttcttggtc tcagtgtcaa tgcagccccc attgtggtca
1620caggaagtag aggaggccac gttcttacta gtttcccttg catggtttag
aaagcttgcc 1680ctggtgcctc accccttgcc ataattacta tgtcatttgc
tggagctctg cccatcctgc 1740ccctgagccc atggcactct atgttctaag
aagtgaaaat ctacactcca gtgagacagc 1800tctgcatact cattaggatg
gctagtatca aaagaaagaa aatcaggctg gccaacgggg 1860tgaaaccctg
tctctactaa aaatacaaaa aaaaaaaaaa attagccggg cgtggtggtg
1920agtgcctgta atcacagcta cttgggaggc tgagatggga gaatcacttg
aacccgggag 1980gcagaggttg cagtgagccg agattgtgcc cctgcactcc
agcctgagcg acagtgagac 2040tctgtctcag tccatgaaga tgtagaggag
aaactggaac tctcgagcgt tgctgggggg 2100gattgtaaaa tggtgtgacc
actgcagaag acagtatggc agctttcctc aaaacttcag 2160acatagaatt
aacacatgat cctgcaattc cacttatagg aattgaccca caagaaatga
2220aagcagggac ttgaacccat atttgtacac caatattcat agcagcttat
tcacaagacc 2280caaaaggcag aagcaaccca aatgttcatc aatgaatgaa
tgaatggcta agcaaaatgt 2340gatatgtacc taacgaagta tccttcagcc
tgaaagagga atgaagtact catacatgtt 2400acaacacgga cgaaccttga
aaactttatg ctaagtgaaa taagccagac atcaacagat 2460aaatagttta
tgattccacc tacatgaggt actgagagtg aacaaattta cagagacaga
2520aagcagaaca gtgattacca gggactgagg ggaggggagc atgggaagtg
acggtttaat 2580gggcacaggg tttatgttta ggatgttgaa aaagttctgc
agataaacag tagtgatagt 2640tgtaccgcaa tgtgacttaa tgccactaaa
ttgacactta aaaatggttt aaatggtcaa 2700ttttgttatg tatattttat
atcaatttaa aaaaaaacct gagccccaaa aggtatttta 2760atcaccaagg
ctgattaaac caaggctaga accacctgcc tatatttttt gttaaatgat
2820ttcattcaat atcttttttt taataaacca tttttacttg ggtgtttata
aaaaaaaaaa 2880318DNAHomo sapiens 3tgcaccccca agaccaaa 18426DNAHomo
sapiens 4tgattaaagc taacgagcag acagaa 26526DNAHomo sapiens
5ccttcccttt cttggctttg gccttt 26622DNAHomo sapiens 6cgttgtacct
gcccaattgt ga 22720DNAHomo sapiens 7gggacgcatc actcaacgtt
20827DNAHomo sapiens 8aagagaaagc agtgcaaacc ttcccgt 27916DNAHomo
sapiens 9gccgccgcgg aataat 161028DNAHomo sapiens 10tgtctaccct
tatacacaac tccatagg 281131DNAHomo sapiens 11agccgggatc taccataccc
attgactaac t 311220DNAHomo sapiens 12tggaacggcc aaatgtactg
201320DNAHomo sapiens 13tggcgtattc ccgttcaagt 201423DNAHomo sapiens
14actctgcccg acgtggtctc cca 231523DNAHomo sapiens 15ccttgaggca
cagtgaagac atc 231623DNAHomo sapiens 16cctgtaggac acctccagaa gag
231727DNAHomo sapiens 17tggccactcc aataacagca ggtcaca 27
* * * * *
References