U.S. patent application number 14/997078 was filed with the patent office on 2016-07-07 for system and method for integrated quantifiable detection, diagnosis and monitoring of disease using patient related time trend data.
The applicant listed for this patent is General Electric Company. Invention is credited to Gopal Biligeri Avinash, Zhongmin Steve Lin, Ananth Mohan, Saad Ahmed Sirohey.
Application Number | 20160196393 14/997078 |
Document ID | / |
Family ID | 44531914 |
Filed Date | 2016-07-07 |
United States Patent
Application |
20160196393 |
Kind Code |
A1 |
Avinash; Gopal Biligeri ; et
al. |
July 7, 2016 |
System and Method for Integrated Quantifiable Detection, Diagnosis
and Monitoring of Disease using Patient Related Time Trend Data
Abstract
A variety of systems, methods, and articles of manufacture are
disclosed. An example includes accessing of patient deviation
scores indicative of differences between patient data and reference
data representative of a population segment, wherein the patient
deviation scores are derived from longitudinal data of the patient
data such that the patient deviation scores include a plurality of
sets of patient deviation scores, each set indicative of
differences between patient data collected at a respective point in
time and the reference data; identifying a trend in the patient
deviation scores for at least one clinical parameter; generating of
a report including a visual indication of the trend; and outputting
of the report. The report includes one or more views including Z,
T, D, DT, and D feedback on T views, using image and non-image
data.
Inventors: |
Avinash; Gopal Biligeri;
(Menomonee Falls, WI) ; Sirohey; Saad Ahmed;
(Pewaukee, WI) ; Lin; Zhongmin Steve; (New Berlin,
WI) ; Mohan; Ananth; (Waukesha, WI) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
General Electric Company |
Schenectady |
NY |
US |
|
|
Family ID: |
44531914 |
Appl. No.: |
14/997078 |
Filed: |
January 15, 2016 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12627149 |
Nov 30, 2009 |
9271651 |
|
|
14997078 |
|
|
|
|
Current U.S.
Class: |
705/2 |
Current CPC
Class: |
A61B 5/00 20130101; G16H
30/20 20180101; G06F 19/321 20130101; G16H 40/63 20180101; G06F
19/324 20130101 |
International
Class: |
G06F 19/00 20060101
G06F019/00 |
Claims
1. An apparatus comprising: a processor particularly configured to:
calculate a plurality of deviation scores (Z-scores) for each
parameter based on the patient population, each z-score determined
based on a deviation between a value for the parameter measured
from the patient population and a reference value for the parameter
from a reference population; calculate a time trend score (T-score)
for each parameter from the plurality of Z-scores over time;
calculate a distribution score (DT-score) as a distribution on the
T-score for each parameter based on a respective shift in the
patient population from the reference population for the respective
parameter; and visualize the DT-score via a graphical interface DT
viewer on a display, the visualizing including a graphical
visualization of parameter shift from the patient population to the
reference population based on the DT-score, the visualization of
the DT-score providing a visualization of the disease signature for
the patient population.
2. The apparatus of claim 1, wherein the DT viewer provides a
graphical visualization of parameter shift with respect to brain
image information in the patient population.
3. The apparatus of claim 1, wherein the plurality of parameters
include image-related parameters and non-image related
parameters.
4. The apparatus of claim 1, wherein the processor is configured to
identify one or more of the plurality of parameters as one or more
parameters of interest and to weigh the one or more parameters of
interest more than the remaining of the plurality of
parameters.
5. The apparatus of claim 4, wherein the DT viewer is to highlight
the one or more parameters of interest in contrast to the remaining
one or more parameters.
6. The apparatus of claim 1, wherein the graphical visualization of
parameter shift includes a color-based graphical visualization.
7. The apparatus of claim 1, wherein the processor is further
configured to feed information from the graphical interface DT
viewer into a T viewer, wherein the T viewer is to monitor disease
progression over time for a particular patient in the patient
population based on the disease signature.
8. The apparatus of claim 7, wherein the processor is further
configured to form a disease profile from a sequence of the disease
signature over time.
9. The apparatus of claim 7, wherein the processor is further
configured to provide feedback to the DT viewer based on
interaction with the T viewer and updating the T viewer based on
changes in the DT viewer.
10. A method comprising: calculating, using a processor, a
plurality of deviation scores (Z-scores) for each parameter based
on the patient population, each z-score determined based on a
deviation between a value for the parameter measured from the
patient population and a reference value for the parameter from a
reference population; calculating, using the processor, a time
trend score (T-score) for each parameter from the plurality of
Z-scores over time; calculating, using the processor, a
distribution score (DT-score) as a distribution on the T-score for
each parameter based on a respective shift in the patient
population from the reference population for the respective
parameter; and visualizing, using the processor, the DT-score via a
graphical interface DT viewer on a display, the visualizing
including a graphical visualization of parameter shift from the
patient population to the reference population based on the
DT-score, the visualization of the DT-score providing a
visualization of the disease signature for the patient
population.
11. The method of claim 10, wherein the plurality of parameters
include image-related parameters and non-image related
parameters.
12. The method of claim 10, further including: identifying one or
more of the plurality of parameters as one or more parameters of
interest; and weighing the one or more parameters of interest more
than the remaining of the plurality of parameters.
13. The method of claim 12, further including: highlighting, via
the DT viewer, the one or more parameters of interest in contrast
to the remaining one or more parameters.
14. The method of claim 10, wherein the graphical visualization of
parameter shift includes a color-based graphical visualization.
15. The method of claim 10, further including: feeding information
from the graphical interface DT viewer into a T viewer; and
monitoring, via the T viewer, disease progression over time for a
particular patient in the patient population based on the disease
signature.
16. The method of claim 15, wherein the processor is further
configured to form a disease profile from a sequence of the disease
signature over time.
17. The method of claim 15, further including: providing feedback
to the DT viewer based on interaction with the T viewer and
updating the T viewer based on changes in the DT viewer.
18. A non-transitory computer-readable medium including
instructions which, when executed by a processor, cause the
processor to: calculate a plurality of deviation scores (Z-scores)
for each parameter based on the patient population, each z-score
determined based on a deviation between a value for the parameter
measured from the patient population and a reference value for the
parameter from a reference population; calculate a time trend score
(T-score) for each parameter from the plurality of Z-scores over
time; calculate a distribution score (DT-score) as a distribution
on the T-score for each parameter based on a respective shift in
the patient population from the reference population for the
respective parameter across the plurality of parameter T scores to
generate a disease signature from the patient population; and
visualize the DT-score via a graphical interface DT viewer on a
display, the visualizing including a graphical visualization of
parameter shift from the patient population to the reference
population based on the DT-score, the visualization of the DT-score
providing a visualization of the disease signature for the patient
population.
19. The computer-readable medium of claim 18, wherein the
instructions further cause the processor to: feed information from
the graphical interface DT viewer into a T viewer; and monitor, via
the T viewer, disease progression over time for a particular
patient in the patient population based on the disease
signature.
20. The computer-readable medium of claim 10, wherein the
instructions further cause the processor to form a disease profile
from a sequence of the disease signature over time.
Description
RELATED APPLICATIONS
[0001] The present application claims the benefit of priority to
U.S. patent application Ser. No. 12/627,149, filed on Nov. 30,
2009, entitled "System and Method for Integrated Quantifiable
Detection, Diagnosis and Monitoring of Disease using Patient
Related Time Trend Data", which is herein incorporated by reference
in its entirety for all purposes.
BACKGROUND
[0002] The present disclosure relates generally to detecting and
monitoring trends in data and, more particularly in some
embodiments, to the diagnosis and monitoring of medical conditions
from patient deviation data. The present invention relates
generally to medical diagnosis and, more particularly, to the
diagnosis of medical conditions from patient deviation data.
[0003] One type of medical condition or disease that is of interest
to the medical community is neurodegenerative disorders (NDDs),
such as Alzheimer's disease and Parkinson's disease. Alzheimer's
disease currently afflicts tens of millions of people worldwide,
and accounts for a majority of dementia cases in patients. Further,
there is not, as of yet, any known cure. The economic and social
costs associated with Alzheimer's disease are significant, and are
increasing over time.
[0004] However, NDDs may be challenging to treat and/or study
because they are both difficult to detect at an early stage, and
hard to quantify in a standardized manner for comparison across
different patient populations. In response to these difficulties,
investigators have developed methods to determine statistical
deviations from normal patient populations. For example, one
element of the detection of NDDs is the development of age and
tracer segregated normal databases. Comparison to these normals can
only happen in a standardized domain, e.g., the Talairach domain or
the Montreal Neurological Institute (MNI) domain. The MNI defines a
standard brain by using a large series of magnetic resonance
imaging (MRI) scans on normal controls. The Talairach domain
references a brain that is dissected and photographed for the
Talairach and Tournoux atlases. In both the Talairach domain and
the MNI domain, data must be mapped to the respective standard
domain using registration techniques. Current methods that use a
variation of the above method include tracers NeuroQ.RTM.,
Statistical Parametric matching (SPM), 3D-sterotactic surface
projections (3D-SSP), and so forth.
[0005] Once a comparison has been made, an image representing a
statistical deviation of the anatomy is displayed, allowing a
viewer to make a diagnosis based on the image. Making such a
diagnosis is a very specialized task and is typically performed by
highly-trained medical image experts. However, even such experts
can only make a subjective call as to the degree of severity of the
disease. Due to this inherent subjectivity, the diagnoses tend to
be inconsistent and non-standardized.
[0006] Additionally, in numerous medical contexts including but not
limited to NDD detection, analysis and reporting of results often
takes place in separate informational "silos" that are distinct
from one another. For instance, PET & MR exams are read and
interpreted by an imaging expert, while blood and cerebro-spinal
fluid results are read and interpreted by a laboratory physician.
Consequently, in many such instances any diagnosis made by the
imaging expert or the laboratory physician may be based on only a
portion of relevant patient information available.
BRIEF DESCRIPTION
[0007] Certain aspects commensurate in scope with the originally
claimed invention are set forth below. It should be understood that
these aspects are presented merely to provide the reader with a
brief summary of certain forms the invention might take and that
these aspects are not intended to limit the scope of the invention.
Indeed, the invention may encompass a variety of aspects that may
not be set forth below.
[0008] According to one embodiment, a system includes a memory
device having a plurality of routines stored therein, and a
processor configured to execute the plurality of routines stored in
the memory device. The plurality of routines may include a routine
configured to effect accessing of a patient image deviation score
indicative of a difference between patient image data and reference
image data representative of a population segment. Further, the
plurality of routines may include a routine configured to effect
accessing of a patient non-image deviation score indicative of a
difference between patient non-image data and reference non-image
data representative of the population segment. Additionally, the
plurality of routines may further include routines configured to
effect generating of a report having visual indications of
deviations of the patient image and non-image data from the
respective reference image and non-image data, and to effect
outputting of the report.
[0009] According to another embodiment, a computer-implemented
method includes accessing at least one patient image deviation
score derived through a comparison of patient image data to
standardized image data representative of a population of
individuals. The method may also include accessing one or more
patient non-image deviation scores derived through a comparison of
patient non-image data to standardized non-image data
representative of the population of individuals. Still further, the
method may include processing the image and non-image deviation
scores to generate a visual output indicative of differences
between the patient data and the standardized data, and may include
displaying the visual output to facilitate diagnosis of a patient
medical condition.
[0010] According to a further embodiment, a computer-implemented
method includes accessing an image deviation score of a patient
calculated from a comparison of patient image data from at least
two different imaging modalities to standardized image data. The
method may also include processing the image deviation score to
generate a visual output including a graphical representation
indicative of a difference between the patient image data and the
standardized image data. Still further, the method may include
displaying the visual output.
[0011] According to yet another embodiment, a computer-implemented
method includes accessing patient non-image deviation scores
calculated from a comparison of longitudinal patient non-image data
with standardized non-image data. The method may also include
processing the patient non-image deviation scores to generate a
visual output including a graphical representation indicative of a
difference between at least a subset of the longitudinal patient
non-image data and the standardized non-image data. Still further,
the method may include displaying the visual output.
[0012] According to yet another embodiment, a manufacture includes
a computer-readable medium having executable instructions stored
thereon. The executable instructions may include instructions
adapted to access a patient image deviation score derived from a
comparison of patient image data to reference image data. The
executable instructions may also include instructions adapted to
access a patient non-image deviation score derived through a
comparison of patient non-image data to reference non-image data.
Further, the executable instructions may include instructions
adapted to generate, based at least in part on the image and
non-image deviation scores, and to display a visual output
indicative of a difference between the patient image data and the
reference image data, and of a difference between the patient
non-image data and the reference non-image data.
[0013] Various refinements of the features noted above may exist in
relation to various aspects of the present invention. Further
features may also be incorporated in these various aspects as well.
These refinements and additional features may exist individually or
in any combination. For instance, various features discussed below
in relation to one or more of the illustrated embodiments may be
incorporated into any of the above-described aspects of the present
invention alone or in any combination. Again, the brief summary
presented above is intended only to familiarize the reader with
certain aspects and contexts of the present invention without
limitation to the claimed subject matter.
BRIEF SUMMARY OF THE PREFERRED EMBODIMENTS OF THE INVENTION
[0014] The preferred embodiments of the present invention may be
summarized as follows: a report of non-alphanumeric visual indicia
generated by a method for integrated quantifiable detection,
diagnosis and monitoring of a medical condition using a plurality
of medical diagnosis test results, comprising: a plurality of
different time dependent metrics corresponding to medical diagnosis
test results of an identified patient at a plurality of time points
including a first set of at least one time dependent metric
corresponding to medical diagnosis test results of an identified
patient at a plurality of time points, and a second set of time
dependent reference metrics corresponding to medical diagnosis test
results of an identified population at a plurality of time, wherein
each of a plurality of quantified deviations between the first and
second set of metrics are aggregated to generate a visual
representation corresponding to a medical condition when considered
collectively to generate the report of visual indicia
therefrom.
[0015] The report further comprises at least one representation of
a medical image. The first and second sets include data from more
than one medical diagnosis test. The second set of metrics includes
data corresponding to more than one de-identified patient, or a
plurality of different tests wherein the plurality of different
tests, is a single test type taken repetitively over time. The
second set of metrics further includes a normal reference data
corresponding to a predefined normal sample standard, and/or
abnormal reference data corresponding to a predefined abnormal
sample standard.
[0016] Alternatively at least a portion of the plurality of time
dependent metrics aggregated to generate a visual representation
corresponding to a medical condition represented by the quantified
deviations when considered collectively to generate the report of
visual indicia therefrom, further includes comparing the medical
diagnosis test results data corresponding to a selected test type
that is present in both the first set and second set, and
generating at least some of the plurality of metrics.
[0017] The first set and second set, may further comprise data of
the type selected from the following group of data types including:
image, numeric, waveform, enumerated, Boolean logic, or text. The
plurality of quantified deviations between the first and second set
of of metrics, preferably further includes at least one test type
common to both the first and second set, and more than one test
type common to both the first and second set.
[0018] The present invention may also be summarized as follows: a
report of non-alphanumeric visual indicia generated by a method for
integrated quantifiable detection, diagnosis and monitoring of a
medical condition using a plurality of medical diagnosis test
results, comprising: at least one time dependent metric
corresponding to medical diagnosis test results of an identified
patient at a plurality of time points; a reference set of time
dependent metrics corresponding to medical diagnosis test results
of an identified population at a plurality of time points; and
wherein each of a plurality of quantified deviations between more
than one of the time dependent metrics corresponding to the
identified patient and the reference set of time dependent metrics
are aggregated to generate a visual representation corresponding to
a medical condition when considered collectively to generate the
report of visual indicia therefrom.
[0019] The present invention may also be summarized as follows: a
report of non-alphanumeric visual indicia generated by a method for
integrated quantifiable detection, diagnosis and monitoring of a
medical condition using a plurality of medical diagnosis test
results, comprising: a plurality of different time dependent
metrics; each metric corresponds to a measure of time trend of
distinct quantified deviation between a first data set of medical
diagnosis test results corresponding to an identified patient at a
plurality of time points, and provided for comparison with a second
data set of medical diagnosis test results corresponding to at
least one de-identified patient, and the data corresponding to the
test results within either of the first data set and second data
set is not included in the other; wherein at least some of the
plurality of time dependent metrics are aggregated to generate a
visual representation corresponding to a medical condition and used
to observe the medical condition represented by the plurality of
different metrics when considered collectively to generate the
visual report therefrom.
[0020] Of course, the medical test results can be derivations of
the results themselves, or the raw data forming the results, such
that medical test and any associated results means the raw data or
manipulated raw data such as by weighting, truncation, or the
application of some mathematical function applied thereto to
generate derived results and still be considered test results
according to the present invention(s).
[0021] The various views can be summarily described as follows and
each comprises a distinct invention:
[0022] Z-score is calculated for each patient type, for each
patient, for each test, For each time point:
Z-score=(test-m_reference)/s_reference, i.e., deviation of test
result of a patient with respect to the reference population.
[0023] T-Score is calculated for each patient type, for each
patient, for each test, for all time points: T-score=time trend
metric of Z-scores at all time points.
[0024] D-Score is calculated for each patient type, for each test,
for each time point: D-score=separation between two distributions
wherein the first distribution is a different patient type than the
second distribution; and each is a Z-score distribution of
respective patient type for a given test.
[0025] DT Score is calculated for each patient type, for each test,
for all time points and reveals disease signatures:
DT-score=separation between two distributions wherein the first
distribution is a different patient type than the second
distribution; and each is a T-score distribution of respective
patient type for a given test.
[0026] D Score feedback on the T Score is calculated by weighting
the T Scores with disease signature data to create a disease
profile.
DRAWINGS
[0027] These and other features, aspects, and advantages of the
present invention will become better understood when the following
detailed description is read with reference to the accompanying
drawings in which like characters represent like parts throughout
the drawings, wherein:
[0028] FIG. 1 is a block diagram of an exemplary processor-based
device or system in accordance with one embodiment of the present
invention;
[0029] FIG. 2 is a block diagram of an exemplary data acquisition
and processing system in accordance with one embodiment of the
present invention;
[0030] FIG. 3 is a flow chart of an exemplary method for preparing
image data for feature extraction in accordance with one embodiment
of the present invention;
[0031] FIG. 4 is a flow chart of an exemplary method for creating a
cortical thickness map from brain image data in accordance with one
embodiment of the present invention;
[0032] FIG. 5 is a flow chart of an exemplary method for generating
deviation maps in accordance with one embodiment of the present
invention;
[0033] FIG. 6 is an exemplary visual mapping of cortical thickness
data on an inflated brain surface in accordance with one embodiment
of the present invention;
[0034] FIG. 7 is a block diagram representative of the division of
reference data into standardized databases in accordance with one
embodiment of the present invention;
[0035] FIG. 8 is a flow chart of an exemplary diagnosis method in
accordance with one embodiment of the present invention;
[0036] FIG. 9 is a flow chart of an exemplary method for creating
and analyzing deviation data in accordance with one embodiment of
the present invention;
[0037] FIG. 10 is a flow chart of a method for diagnosing a patient
based on comparison of a patient deviation map to reference
deviation maps in accordance with one embodiment of the present
invention;
[0038] FIG. 11 is a flow chart of an exemplary method for
generating a composite deviation map indicative of both structural
and functional deviation in accordance with one embodiment of the
present invention;
[0039] FIG. 12 is a flow chart of a method for generating image
deviation scores for a patient in accordance with one embodiment of
the present invention;
[0040] FIG. 13 is a flow chart of a method for generating non-image
deviation scores for a patient in accordance with one embodiment of
the present invention;
[0041] FIG. 14 is a flow chart of an exemplary method for
generating a visual representation of patient deviation data based
on deviation scores in accordance with one embodiment of the
present invention;
[0042] FIG. 15 illustrates an exemplary visual representation of a
variety of patient deviation data in accordance with one embodiment
of the present invention;
[0043] FIG. 16 is a flow chart of an exemplary visualization method
in accordance with one embodiment of the present invention;
[0044] FIG. 17 is a flow chart of a different exemplary
visualization method in accordance with one embodiment of the
present invention;
[0045] FIG. 18 is a diagram of an automatic comparison workflow to
determine a severity index in accordance with one embodiment of the
present invention;
[0046] FIG. 19 is a flow chart of an exemplary method for
calculating a combined disease severity score in accordance with
one embodiment of the present invention;
[0047] FIG. 20 is a block diagram generally illustrating a process
for comparing patient data to standardized data for a plurality of
disease types and severity levels in accordance with one embodiment
of the present invention;
[0048] FIG. 21 illustrates a plurality of representative reference
deviation maps that may be contained in a reference library or
database of such deviation maps in accordance with one embodiment
of the present invention; and
[0049] FIG. 22 illustrates additional representative reference
deviation maps that may be contained in a reference library or
database of deviation maps in accordance with one embodiment of the
present invention.
[0050] FIG. 23 illustrates multiple longitudinal trends of numerous
clinical parameters.
[0051] FIG. 24 illustrates the T viewer utilizing the Z score and T
score color map.
[0052] FIG. 25 illustrates a single patient Z score taken over
numerous discrete time events.
[0053] FIG. 26 illustrates the preferred embodiment of the T score
holistic viewer of the present invention.
[0054] FIG. 27 illustrates the T viewer color map utilizing colors
and brightness to indicating relative densities of occurrence for
as given medical condition over time.
[0055] FIG. 28 illustrates an exemplary non-image date map utilized
with the views of the holistic viewer of the present
inventions.
[0056] FIG. 29 illustrates an exemplary image date map utilized
with the views of the holistic viewer of the present
inventions.
[0057] FIG. 30 illustrates an exemplary non-image date depiction of
relative population densities of occurrence for as given medical
condition over time.
[0058] FIG. 31 illustrates an exemplary depiction of relative
shifted distributions of the comparative data associated with
different populations for a given view.
[0059] FIG. 32 illustrates an exemplary depiction of relative
overlapping distributions of the comparative data associated with
different populations for a given view.
[0060] FIG. 33 illustrates an exemplary depiction of relative
slight overlapping distributions of the comparative data associated
with different populations for a given view.
[0061] FIG. 34 illustrates an exemplary depiction of relative
overlapping distributions of the comparative data associated with
different populations for a given view, and threshold value
correction imposed thereon.
[0062] FIG. 35 illustrates an exemplary depiction of relative
overlapping distributions of the comparative data associated with
different populations for a given view, and percentile based
correction imposed thereon.
[0063] FIG. 36 illustrates the D viewer utilizing the Z score and D
score color map.
[0064] FIG. 37 illustrates an exemplary holistic viewer for the
normal population.
[0065] FIG. 38 illustrates an exemplary holistic viewer for the
abnormal population.
[0066] FIG. 39 illustrates the D score color map utilizing colors
and brightness to indicating relative densities of occurrence for
as given medical condition over the populations compared.
[0067] FIG. 40 illustrates multiple trends of numerous clinical
parameters across populations or groups.
[0068] FIG. 41 illustrates the DT viewer utilizing T score
distributions and DT score color map.
[0069] FIG. 42 illustrates an exemplary holistic viewer for the
normal population, and three corresponding disease signature data
views for the same test, wherein the sequence of the disease
signature data forms the disease profile.
[0070] FIG. 43 illustrates the weighting of the T scores data
metrics by the DT score data metrics as feedback to develop the
disease profile of progression for the longitudinal T score data,
as well as weighting of the Z scores data metrics by the D score
data metrics as feedback to develop the disease profile for the Z
score data
[0071] FIG. 44 illustrates an exemplary embodiment of the non-image
data segments associated any of the holistic viewers.
[0072] FIG. 45 illustrates an exemplary embodiment of overlaying
the non-image data segments associated any of the holistic
viewers.
[0073] FIG. 46 illustrates, by exemplary comparison, the color maps
for weighted (standard) and non-weighted Z viewer color map
reports.
[0074] FIG. 47 illustrates, by exemplary comparison, the non-image
data mapping with selective suppression of user defined data.
[0075] FIG. 48 illustrates, by exemplary depiction, a preferred
embodiment of the non-image data mapping for a given viewer in
conjunction with the corresponding mapping key.
[0076] FIG. 49 illustrates, by exemplary depiction, a preferred
embodiment of the non-image data mapping for a given viewer in
conjunction with the corresponding mapping key.
[0077] FIG. 50 illustrates a technique for calculating a T-score,
according to an embodiment of the invention.
[0078] FIG. 51 illustrates a technique for calculating a T-score,
according to another embodiment of the invention.
[0079] FIG. 52 illustrates a technique for calculating a T-score,
according to another embodiment of the invention.
[0080] FIG. 53 illustrates a technique for calculating a T-score,
according to yet another embodiment of the invention.
DETAILED DESCRIPTION
[0081] One or more specific embodiments of the present invention
will be described below. In an effort to provide a concise
description of these embodiments, all features of an actual
implementation may not be described in the specification. It should
be appreciated that in the development of any such actual
implementation, as in any engineering or design project, numerous
implementation-specific decisions must be made to achieve the
developers' specific goals, such as compliance with system-related
and business-related constraints, which may vary from one
implementation to another. Moreover, it should be appreciated that
such a development effort might be complex and time consuming, but
would nevertheless be a routine undertaking of design, fabrication,
and manufacture for those of ordinary skill having the benefit of
this disclosure.
[0082] When introducing elements of various embodiments of the
present invention, the articles "a," "an," "the," and "said" are
intended to mean that there are one or more of the elements. The
terms "comprising," "including," and "having" are intended to be
inclusive and mean that there may be additional elements other than
the listed elements. Moreover, while the term "exemplary" may be
used herein in connection to certain examples of aspects or
embodiments of the presently disclosed technique, it will be
appreciated that these examples are illustrative in nature and that
the term "exemplary" is not used herein to denote any preference or
requirement with respect to a disclosed aspect or embodiment.
Further, any use of the terms "top," "bottom," "above," "below,"
other positional terms, and variations of these terms is made for
convenience, but does not require any particular orientation of the
described components.
[0083] Turning now to the drawings, and referring first to FIG. 1,
an exemplary processor-based system 10 for use in conjunction with
the present technique is depicted. In one embodiment, the exemplary
processor-based system 10 is a general-purpose computer, such as a
personal computer, configured to run a variety of software,
including software implementing all or part of the presently
disclosed techniques, including the methods and functionality
described throughout the instant disclosure. Alternatively, in
other embodiments, the processor-based system 10 may comprise,
among other things, a mainframe computer, a distributed computing
system, or an application-specific computer or workstation
configured to implement all or part of the present techniques based
on specialized software and/or hardware provided as part of the
system. Further, the processor-based system 10 may include either a
single processor or a plurality of processors to facilitate
implementation of the presently disclosed functionality.
[0084] In general, the exemplary processor-based system 10 includes
a microcontroller or microprocessor 12, such as a central
processing unit (CPU), which executes various routines and
processing functions of the system 10. For example, the
microprocessor 12 may execute various operating system instructions
as well as software routines configured to effect certain processes
stored in or provided by a manufacture including a computer
readable-medium, such as a memory 14 (e.g., a random access memory
(RAM) of a personal computer) or one or more mass storage devices
16 (e.g., an internal or external hard drive, a solid-state storage
device, CD-ROM, DVD, or other storage device). In addition, the
microprocessor 12 processes data provided as inputs for various
routines or software programs, such as data provided in conjunction
with the present techniques in computer-based implementations.
[0085] Such data may be stored in, or provided by, the memory 14 or
mass storage device 16. Alternatively, such data may be provided to
the microprocessor 12 via one or more input devices 18. As will be
appreciated by those of ordinary skill in the art, the input
devices 18 may include manual input devices, such as a keyboard, a
mouse, or the like. In addition, the input devices 18 may include a
network device, such as a wired or wireless Ethernet card, a
wireless network adapter, or any of various ports or devices
configured to facilitate communication with other devices via any
suitable communications network, such as a local area network or
the Internet. Through such a network device, the system 10 may
exchange data and communicate with other networked electronic
systems, whether proximate to or remote from the system 10. It will
be appreciated that the network may include various components that
facilitate communication, including switches, routers, servers or
other computers, network adapters, communications cables, and so
forth.
[0086] Results generated by the microprocessor 12, such as the
results obtained by processing data in accordance with one or more
stored routines, may be stored in a memory device, may undergo
additional processing, or may be provided to an operator via one or
more output devices, such as a display 20 and/or a printer 22.
Also, based on the displayed or printed output, an operator may
request additional or alternative processing or provide additional
or alternative data, such as via the input device 18. As will be
appreciated by those of ordinary skill in the art, communication
between the various components of the processor-based system 10 may
typically be accomplished via a chipset and one or more busses or
interconnects which electrically connect the components of the
system 10. Notably, in certain embodiments of the present
techniques, the exemplary processor-based system 10 may be
configured to facilitate patient diagnosis, as discussed in greater
detail below.
[0087] An exemplary system 30 for acquiring and processing data is
illustrated in FIG. 2 in accordance with one embodiment of the
present invention. The system 30 includes a data processing system
32 configured to provide various functionality. It should be noted
that, in one embodiment, the data processing system 32 may include
a processor-based system, such as system 10, having any suitable
combination of hardware and/or software code, routines, modules, or
instructions adapted to perform the presently discussed
functionality, including performance of various steps of the
methods described elsewhere herein. It should be noted that such
software routines may be embodied in a manufacture (e.g., a compact
disc, a hard drive, a flash memory, RAM, or the like) and
configured to be executed by a processor to effect performance of
the functionality described herein.
[0088] The system 30 may also include one or more data acquisition
systems 34 for collecting data from, or regarding, a patient 36.
The patient data may include one or both of image data and
non-image data, and may include any of static data, dynamic data,
and longitudinal data. In various embodiments, the data acquisition
systems 34 may include patient monitors, imaging systems of various
modalities, computers, or any other suitable systems capable of
collecting or receiving data regarding the patient 36. For
instance, the data acquisition systems 34 may include, among
others, an X-ray system, a computed tomography (CT) imaging system,
a magnetic resonance (MR) imaging system, a positron emission
tomography (PET) imaging system, a single photon emission computed
tomography (SPECT) imaging system, a digital tomosynthesis imaging
system, an electroencephalography (EEG) system, an
electrocardiography (ECG or EKG) system, an electromyography (EMG)
system, an electrical impedance tomography (EIT) system, an
electronystagmography (ENG) system, a system adapted to collect
nerve conduction data, or some combination of these systems.
[0089] Various components of the system 30, including the data
processing system 32 and the data acquisition systems 34, may be
connected to one another via a network 38 that facilitates
communication between such components. The system 30 may also
include one or more databases, such as databases 40 and 42, for
storing data, such as data collected by the data acquisition
systems 34 and data used by or generated from the data processing
system 32, including both patient data and standardized reference
data, as discussed in greater detail below. Additionally, the data
processing system 32 may receive data directly from the data
acquisition systems 32, from the databases 40 and 42, or in any
other suitable fashion.
[0090] In some embodiments, it may be desirable to analyze one or
more features of interest from image data to facilitate diagnosis
of a patient with respect to one or more disease types or disease
severity levels. Accordingly, an exemplary method 48 for preparing
image data for feature extraction is generally illustrated in FIG.
3 in accordance with one embodiment of the present invention. Image
data 50 may be obtained from various sources, such as one or more
of the data acquisition systems 34, the databases 40 or 42, or the
like. Further, such image data may be related to a particular
patient, such as the patient 36, or to one or more reference
individuals of population sample. The method 48 may include various
steps, such as steps 52, 54, 56, 58, and 60, for processing,
registering, and extracting features of interest.
[0091] In the presently illustrated embodiment, the method 48
includes a step 52 of preprocessing the image data. Such
preprocessing may include a host of sub-processes, such an
intensity correction, resembling, filtering, and so forth. In steps
54 and 56, anatomical markers in the image data 50 may be detected,
and an image grid may be created. Based on the anatomical markers
and the image grid, the data may undergo registration in a step 58.
Following registration, features of interest in the image data 50
may be extracted in a step 60. While certain exemplary steps of the
method 48 are presently described, it should be noted that the
image data 50 may undergo registration or feature extraction
through fewer, different, or additional steps in full accordance
with the present technique.
[0092] In one embodiment, the image data 50 includes one or more
images of a human brain that may be mapped to a Talairach
coordinate system. In such an embodiment, the image data of the
human brain, which may include an MR image or some other image, may
be normalized to correct intensity variations and resampled, such
as to a 256.times.256.times.128 internal matrix, for further
processing. Also, in such an embodiment, the anterior and posterior
commissures (AC-PC) of the brain image and other anatomical
reference points may be identified to facilitate Talairach
registration. The brain images of the image data 50 may be
elastically registered, such as through warping, to the Talairach
coordinate system to facilitate later representation, analysis, and
diagnostics.
[0093] It should be noted that the particular features that are of
interest in the image data may vary depending on a particular
disease or condition of interest. For example, in diagnosing
neurological conditions, it may be useful to extract certain
features of brain image data to facilitate diagnosis. Further, in
some embodiments, it may be desirable to determine the thickness of
the cerebral cortex of a patient or of one or more reference
individuals. Accordingly, an exemplary method 64 for determining
the cortical thickness of a brain from patient image data or
reference image data, and for generating a cortical thickness map,
is provided in FIG. 4 in accordance one embodiment of the present
invention.
[0094] The method 64 may include a step 68 of segmenting brain
tissue in image data 66 from other anatomical structures outside
the brain, such as the skull. Further, in step 70, white matter of
the brain and subcortical regions, such as ventricles may be
segmented from the gray matter of the cerebral cortex. As the
relative image intensities of the brain white matter and the other
soft tissues may be very close or overlapped, in one embodiment the
segmented brain may be manually edited to remove unwanted remaining
tissue, or to restore inadvertently deleted cortical tissue,
generally corresponding to a step 72. Further white matter
segmentation, surface fitting, and smoothing may be performed in
steps 74 and 76. In a step 78, the pial surface (i.e., the outside
surface of the brain gray matter) may be detected. It should be
noted that the pial surface generally includes numerous gyri and
sulci, but may be considered to be smooth regionally to facilitate
processing. The pial surface may be detected in various matters,
such as through use of a deformable model or dilation from the
surface of the white matter. The thickness of the cerebral cortex
(i.e., the cortical thickness) may be calculated in a step 80, and
a cortical thickness map visually depicting the cortical thickness
may be created in a step 82.
[0095] In some embodiments, standardized reference cortical
thickness maps may be calculated from image data collected from
other persons or groups of persons (e.g., normal persons, persons
diagnosed with Alzheimer's disease (AD), persons diagnosed with
Parkinson's disease (PD), persons diagnosed with frontotemporal
dementia (FTD), and so forth), and stored in large databases, such
as those collected by the Alzheimer's Disease Neuroimaging
Initiative (ADNI). Such standardized maps may serve as reference
image data with respect to patient cortical measurements, and may
be grouped and standardized according to any desired
characteristic. For instance, in one embodiment, such data may be
standardized based on a demographic characteristic, such as the
race, gender, or age of the persons from which the data was
collected. Such standardized data allows for the computation of
average cortical thickness of normal patients and the thickness
distribution across different function regions of the brain that
affect memory, movement, speech, language, hearing, vision,
sensation, emotion, and so forth. The average cortical thickness
maps may be created from the reference image data, and also
standardized according to age, gender, or race distributions, or
according to any other characteristic of interest. While certain
presently disclosed embodiments are described with respect to brain
features, such as cortical thickness, it will appreciated that the
present techniques may be applied more generally to any features of
interest, including those of image data of other anatomical regions
besides the brain.
[0096] In some instances, it may be desirable to also generate
anatomical deviation maps, such as cortical thickness deviation
maps, indicative of differences between a patient anatomical region
and a reference anatomical region. As such, an exemplary method 88
for generating deviation maps from standardized reference data is
illustrated in FIG. 5 in accordance with one embodiment of the
present invention. In the presently illustrated embodiment,
reference image data 90 is standardized in a step 92. As noted
above, reference image data may be collected from a population of
individuals and grouped or standardized according to one or more
desired characteristics, such as age, gender, or race. While the
presently illustrated embodiment is described with respect to image
data, it is noted that reference non-image data and patient
non-image data may also, or instead, be used to generate the
deviation maps discussed herein in full accordance with the present
technique.
[0097] The method 88 may include a step 94 of selecting a subset of
the standardized reference image data based on a patient
characteristic. For instance, if a patient is a sixty-five-year-old
woman, a subset of the standardized reference image data grouped to
include reference images pertaining to women between sixty and
seventy years of age may be more relevant for comparative purposes
than a group of standardized reference images composed of data
collected from men between twenty and thirty years of age. Once a
desired group of standardized image data is selected, the matched
standardized image data 96 may be compared to image data 100 of the
patient in a step 98. In other embodiment, non-image data of the
patient may instead or also be compared to matched standardized
non-image data, as described above. Additionally, the various data
may be processed and standardized in any suitable manner to
facilitate such comparisons.
[0098] Based on such comparison, a patient deviation map
representative of the difference between the patient image data 100
and the standardized image data 96 may be generated in step 102.
For example, with respect to cortical thickness, a patient cortical
thickness map may be obtained through a comparison of the patient
cortical thickness map with a standardized cortical thickness map
based on a representative population of normal individuals.
Consequently, in one embodiment, the patient cortical thickness
deviation map may generally illustrate differences of the cortical
thickness of the patient with respect to normal people of similar
age, sex, or race. The deviation maps described herein may be
generated through any suitable techniques. In one embodiment, a
deviation map is a visual representation in which each point of the
map represents a z-score generally corresponding to the number of
standard deviations (based on a population) in the difference
between a patient value and the average value (of the population)
for that point. Although such deviation maps may be calculated from
image data, it is noted that deviation maps may be created using
one or more of numerical data, text data, waveform data, image
data, video data, or the like.
[0099] The various anatomical region maps and deviation maps
described herein may be visualized to facilitate further analysis
or diagnosis. For instance, any or all of the standardized cortical
thickness maps, the patient cortical thickness maps, the patient
cortical thickness deviation maps, or standardized cortical
thickness deviation maps (as described below) may be expressed as
surface matrices, and can be displayed or overlaid on a
three-dimensional (3D) brain surface, a pial surface, or an
inflated brain surface.
[0100] By way of further example, such an expression is illustrated
in FIG. 6 in accordance with one embodiment of the present
invention. Particularly, cortical thicknesses or deviations may be
depicted on an inflated brain surface 108, as illustrated within
window 110. Various regions of the brain 108 may be color coded
according to a scale 112 to represent the cortical thickness, or
deviation from normal thickness, to facilitate user-understanding
of the represented anatomical information.
[0101] Additionally, reference data may be classified and sorted
into standardized databases, such as through an exemplary method
118 generally depicted in FIG. 7 in accordance with one embodiment
of the present invention. The method 118 may include accessing
reference data 120, which may include known population image data,
and classifying such data in a step 122. For example, the reference
data 120 may be classified into various groups, such as data 124
for normal patients; data 126 for patients clinically diagnosed
with a first condition, such as Alzheimer's disease (AD); data 128
for patients diagnosed with a second condition, such as
frontotemporal dementia (FTD); and data 130 for patients diagnosed
with other conditions, such as Parkinson's disease (PD),
Huntington's disease (HD), multi-infarct dementia (MID), diffuse
cortical Lewy body disease (DLBD), normal pressure hydrocephalus,
progressive supranuclear palsy (PSP), or the like. While certain
brain disorders, brain image data, and brain deviation maps are
presently discussed for the sake of explanation, it is again noted
that the use of the present techniques with other, non-neurological
data and disorders is also envisaged. The data 124, 126, 128, and
130 may be stored in respective databases 132, 134, 136, and 138.
Such databases may be stored in one or more memory devices or in
other suitable media.
[0102] An exemplary method 144 for diagnosing a patient based at
least in part on the foregoing data is illustrated in FIG. 8 in
accordance with one embodiment of the present invention. The method
144 may include creating a patient map of a structural feature in a
step 146, based on received patient data 148. In one embodiment
related to brain disorders, the patient map created in step 146 may
include a patient cortical thickness map. In a step 150, a
normalized map of a structural feature is created based on the data
124 for normal patients. For instance, a standardized cortical
thickness map for normal patients may be generated in this step.
Although the presently illustrated embodiment is discussed with
reference to maps of structural features, it is noted that maps of
other features, such as functional or metabolic features, may also
or instead be used in full accordance with the presently disclosed
technique.
[0103] In a step 152, reference condition maps (e.g., average maps
or other reference maps) of the structural feature may be created
for each diagnosed condition or disorder, based on the reference
data 126, 128, and 130 collected with respect to individuals of a
population diagnosed with such conditions. For example, in one
embodiment, representative average cortical thickness map may be
calculated for each brain disorder of interest, such as AD, FTD,
PD, or the like. Additionally, average maps (or other reference
maps) corresponding to various severity levels within a disease
type may also be generated. Thus, multiple representative or
average maps may be created for each diagnosed condition or disease
type.
[0104] The method 144 may also include a step 154 of comparing the
patient and normal maps, and a step 156 of comparing the reference
condition and normal maps. In one embodiment, the method 144 may
include a step 158 of comparing one or more patient deviation maps
(which may be generated from the comparison of step 154) with one
or more disease reference deviation maps (which may be generated
from the comparison of step 156). It is noted that the
above-referenced maps, as well as other maps and data described
herein, may be standardized into one or more common or similar
formats to facilitate analysis and comparison. Also, it will be
appreciated that the various maps described herein may be stored in
one or more databases to facilitate subsequent data analysis.
Additionally, any or all of the foregoing comparisons may be
performed either automatically by a data processing system (e.g.,
system 32), or by a healthcare provider (e.g., a doctor), or by
some combination thereof, to facilitate automatic or manual
diagnosis of the patient in a step 160. Such diagnosis may also be
based on additional data, such as clinical data 162, laboratory
data, patient history, patient vital signs, results of various
tests (e.g., functional tests, cognitive tests, neurological tests,
or genetic tests), and so forth. Additionally, in a step 164 of the
method 144, a report 166 may be output to a database 168 for
storage, or to a user 170 in a human-readable format.
[0105] Based on the patient and reference data and maps discussed
above, numerous reference and patient deviation data and maps may
be created. By way of example, an exemplary method 172 for creating
and analyzing such deviation data is depicted in FIG. 9 in
accordance with one embodiment of the present invention. The method
172 includes accessing reference cortical thickness data for:
normal patients without diagnosed brain disorders (data 174),
patients clinically diagnosed with AD (data 176), patients
diagnosed with FTD (data 178), and patients diagnosed with PD (data
180). The method 172 may also include accessing patient cortical
thickness data 182. It will be appreciated that, in other
embodiments, the method 172 may access reference cortical thickness
data for other brain disorders, which may be processed in a manner
similar to those explicitly discussed in the present example.
Indeed, the present processing techniques may also be applied to
other disorders unrelated to the brain.
[0106] In a step 184, the normal data 174 may be compared to each
of the other data 176, 178, 180, and 182, to generate patient
deviation data 186, AD deviation data 188, FTD deviation data 190,
and PD deviation data 192, all of which may represent deviations
from the normal data 174. Such deviation data may include
structural deviation maps, such as cortical thickness deviation
maps, representative of differences between the patient data and
the disease type reference data, on the one hand, and the normal
reference data on the other. Additionally, the deviation data may
also include functional deviation maps indicative of functional,
rather than structural, differences between the patient (or
reference data indicative of reference disease types) and normal
individuals. In some embodiments, structural deviation maps may
include cortical thickness deviation maps, and functional deviation
maps may include cerebral blood flow rate deviation maps or
metabolic rate deviation maps.
[0107] In step 194, such deviation data may be analyzed. For
instance, in one embodiment, a patient cortical thickness deviation
map may be compared to representative reference cortical thickness
deviation maps for each of the above noted brain disorders to
facilitate diagnosis of the patient with respect to one or more of
such brain disorders. Additionally, reference clinical data 196,
patient clinical data 198, and other data 200 may also be analyzed
by a data processing system or a user to facilitate diagnosis. In
one embodiment, such analysis may include pattern matching of
patient maps and reference maps, and confidence levels of such
matching may be provided to a user. Finally, results 202 of the
analysis may be output to storage or to a user.
[0108] A method 194 for analyzing the data discussed above and
diagnosing a patient is illustrated in FIG. 10 in accordance with
one embodiment of the present invention. In a step 208, one or more
patient deviation maps, which may include a structural deviation
map (e.g., a cortical thickness deviation map) or some other
deviation map (e.g., a functional deviation map), may be compared
to one or more reference deviation maps, such as those previously
described. Notably, the reference deviation maps may include
deviation maps (e.g., functional deviation maps or metabolic
deviation maps or structural deviation maps) representative of one
or more disease types, as well as various severity levels of the
one or more disease types.
[0109] Based on such comparisons, one or more patient disease types
and/or disease severity levels may be identified in a step 210 and
diagnosed in a step 212. In some embodiments, such as a fully
automated embodiment, steps 210 and 212 may be combined. In other
embodiments, however, the identification and diagnosis may be
performed as separate steps. For instance, the data processing
system 32 may identify various potential disease types or severity
levels and present the identified disease types or severity levels
to a user for diagnosis. A report 214 may include an indication of
the identified patient disease types or severity levels, the
diagnosis, or both.
[0110] In some embodiments, it may be desirable to combine
indications of functional deviations and structural deviations of a
patient with respect to reference data and to output such
deviations in a visual manner that facilitates efficient diagnosis
of a patient by a healthcare provider. Accordingly, an exemplary
method 218 for generating a composite deviation map indicative of
both structural and functional deviation is depicted in FIG. 11 in
accordance with one embodiment of the present invention. In the
presently illustrated embodiment, the method 218 includes steps 220
and 222 for accessing structural and functional data, respectively,
for a patient. The patient structural and functional data may
include various image and non-image data with respect to an
anatomical region of the patient. In one embodiment, the anatomical
region may include the cerebral cortex of the patient.
Additionally, the patient structural and functional data may
include image data obtained from different imaging modalities.
[0111] The patient structural and functional data may be compared
to standardized reference structural and functional data,
respectively, in steps 224 and 226. As noted previously, reference
data may be standardized according to any desired characteristics,
such as, but not limited to, age, gender, or race. Based on such
comparisons, one or more structural deviation maps for the patient
may be generated in a step 228, and one or more patient functional
deviation maps may be generated in a step 230. In one embodiment,
the patient structural deviation map may indicate deviation of
patient cortical thickness at one or more particular locations of
the patient cerebral cortex with respect to expected thickness
represented by the standardized reference data. In another
embodiment, the patient structural deviation map may be generated
via comparison of MR images of the patient and of the standardized
reference data. Also, in a neurological context, the patient
functional deviation map may indicate deviation of patient brain
functioning, such as a cerebral blood flow rate or a metabolic
rate, from standardized rates. It will, however, be appreciated
that the deviation maps may be generated based on a wide array of
image data and/or non-image data, as discussed above.
[0112] It is again noted that the patient structural deviation map
may generally represent structural differences of an anatomical
region of the patient with respect to standardized reference data
for a similar anatomical region. For instance, in one embodiment,
the patient structural deviation map may include a cortical
thickness deviation map for the patient with respect to
standardized cortical thickness data, such as described above. In
turn, the patient functional deviation map may represent
non-structural differences between a patient anatomical region and
a corresponding anatomical region of standardized data. For
example, in some embodiments, the patient functional deviation map
may be indicative of differences in metabolic activity or other
functional activity between the patient and standardized reference
data. To facilitate easy and efficient communication of such
differences to a user, a composite patient deviation map,
indicative of both the functional and structural differences
discussed above, may be created in a step 232.
[0113] The patient structural deviation map and the patient
functional deviation map, along with any other additional deviation
maps, may be combined in any suitable fashion to create the
composite patient deviation map. For instance, in one embodiment,
the individual patient deviation maps may be overlain to create a
single composite patient deviation map indicative of multiple
deviations of the patient with respect to standardized data. In
another embodiment, the individual patient functional and
structural deviation maps may be combined through an image fusion
process. Particularly, in one embodiment, the patient structural
deviation map may be generated through comparison of patient image
data and standardized image data each of a first imaging modality,
while the patient functional deviation map is generated from image
data (of both the patient and standardized reference sources)
obtained through a second imaging modality different than the
first. For example, structural deviations identified through
comparison of MR images may be combined with functional deviations
obtained from PET image data to generate a single composite patient
deviation map indicative of both functional and structural
deviations. In another embodiment, the patient structural deviation
map based on a first criterion (e.g., cortical thickness from MRI
images) can be combined with the patient structural deviation map
based on a second criterion (e.g., medial temporal lobe atrophy
from CT images). In yet another embodiment, the patient functional
deviation map based on a first criterion (e.g., FDG, a well known
PET tracer uptake in PET images) can be combined with the patient
functional deviation map based on a second criterion (e.g., uptake
of PIB, a well known tracer for beta-amyloid in PET images).
[0114] Additionally, different colors may be used to indicate and
contrast structural differences and functional differences. For
example, in one embodiment, functional deviations may generally be
depicted in a composite patient deviation map by the color red,
while structural deviations may generally be indicated through use
of the color blue. Additionally, the magnitude of such deviations
may be represented by various shades of red or blue to allow a
doctor or other user to quickly ascertain patient deviations and
the magnitudes of such deviations, as well as to facilitate
diagnosis of the patient. It will be appreciated, however, that
other or additional colors may also be used to indicate the
different types of deviations and their relative magnitudes.
[0115] The method 218 may also include outputting the composite
patient deviation map in a step 234. In some embodiments,
outputting the composite patient deviation map may include storing
the composite patient deviation map in a memory device. In other
embodiments, outputting the composite patient deviation map may
also, or instead, include providing the composite map to a user in
a human-readable format, such as by displaying the composite
patient deviation map on a display or printing a physical copy of
the composite patient deviation map. Also, the presently
illustrated embodiment is currently represented as a parallel
process with respect to the generation of separate patient
structural and functional deviation maps. It is noted that, while
the present exemplary method is described for explanatory purposes
as a parallel process, the steps of any of the methods described
herein may be performed in any suitable manner, and are not limited
to being performed in any particular order or fashion.
[0116] The extent of patient deviation from standardized data may
also be translated into one or more deviation scores, which may, in
one embodiment, be generated through the methods generally depicted
in FIGS. 12 and 13. An exemplary method 240 of FIG. 12 may include
accessing patient image data 242 and reference image data 244. Such
image data may be received from any suitable source, such as a
database or an imaging system. Indeed, the image data 242 and 244
may include image data from a variety of modalities and collected
from a wide range of sources. The reference image data 244 may be
standardized according to any desired characteristics. For
instance, in one embodiment, the reference image data 244 may
generally represent features of normal individuals with certain
demographic characteristics (e.g., characteristics similar to the
patient). In a step 246, the patient image data 242 and the
reference image data 244 may be compared to determine deviations of
the patient image data 242 from the reference image data 244. In
one embodiment, such differences may generally represent deviation
(e.g., structural or functional differences) of the patient from
normal individuals.
[0117] The method 240 may also include a step 248 of calculating
one or more patient image deviation scores for differences between
the patient image data 242 and the reference image data 244. Such
deviation scores may be indicative of an array of functional or
structural deviations of the patient with respect to reference
image data, including deviations in metabolic activity (e.g.,
fluorodeoxyglucose (FDG) metabolism, which may be observed in PET
images), physical anatomy (e.g., cortical thickness, which may be
measured in MR images), and functional activity (e.g., Pittsburgh
Compound-B (PIB) measure, which may be determined from PET images),
to name but a few. The patient image deviation scores may be
calculated in various manners, such as based on projection
deviation, single pixel (2D) deviation, single voxel (3D)
deviation, or on any other suitable technique. The calculated
patient image deviation scores 250 may then be stored in a database
252, output to a user, or may undergo additional processing in one
or more further steps 254.
[0118] Turning to FIG. 13, an exemplary method 260 for calculating
non-image deviation scores may include accessing patient non-image
data 262 and reference non-image data 264. The non-image data may
be received from any suitable source, such as a database, a
computer, or patient monitor. The patient non-image data 262 may
include any non-image information collected for the purpose of
diagnosing the patient, such as clinical data, laboratory data,
patient history, patient vital signs, and the like, and may also
include results of functional tests, cognitive tests, neurological
tests, genetic tests, and so forth. The non-image data 264 may
include similar data, which may be standardized based on one or
more population or sample characteristics. Further, in one
embodiment, the patient non-image data 262 and reference non-image
data 264 may include one or both of numeric data and enumerated
data (each of which may be continuous or discrete). The reference
non-image data 264 may be data representative of features of normal
persons with desired demographic characteristics (e.g.,
characteristics similar to the patient). In a step 266, the patient
non-image data 262 may be compared to the reference non-image data
264 to identify differences between the data. In one embodiment,
such differences may generally represent deviation (e.g.,
structural or functional differences) of the patient from normal
individuals.
[0119] Additionally, the method 260 may include a step 268 of
calculating one or more patient non-image deviation scores for
differences between the patient non-image data 262 and the
reference non-image data 264. It is noted that various techniques
may be used to calculate the patient non-image deviation scores,
including z-score deviation or distribution analysis. Of course, it
will be appreciated that other calculation techniques may also or
instead be employed in other embodiments. The calculated patient
non-image deviation scores 270 may be stored in a database 272,
output to a user, or may undergo additional processing in one or
more further steps 274.
[0120] An exemplary method 280 for accessing patient deviation
scores and generating one or more visual representations to
facilitate patient diagnosis is generally provided in FIG. 14. The
method 280, in one embodiment, includes accessing one or more
patient image deviation scores and one or more patient non-image
deviation scores in steps 282 and 284, respectively. These
deviation scores may be processed, in a step 286, to generate a
visual representation of the differences represented by the patient
deviation scores. In one embodiment, patient deviation scores may
be derived from dynamic data (e.g., video) or longitudinal data
(e.g., data acquired at discrete points in time over a given
period), and multiple visual representations corresponding to
deviations at different points of time may be generated in step
286. The one or more visual representation may then be output, in a
step 288, to facilitate diagnosis of the patient in a step 290. For
deviations derived from dynamic or longitudinal data, multiple
visual representations may be output simultaneously or
sequentially.
[0121] In some embodiments, the visual representation generally
includes a combination and visualization of the various differences
represented by the deviation scores, thus providing a holistic view
of patient variance with respect to standardized data. By way of
example, an exemplary visual representation 296 is depicted in FIG.
15 in accordance with one embodiment of the present invention. It
is noted, however, the presently illustrated embodiment is provided
merely for explanatory purposes, and that other visual outputs may
take different forms.
[0122] In the presently illustrated embodiment, the visual
representation 296 includes a region 298 for visualization of
patient non-image deviation data maps, a region 300 for
visualization of patient image data deviation maps or other image
data, and a control panel 302. In various embodiments, numerous
display techniques may be used to make the visualized deviation
maps or other results more intuitive to a user, and to more clearly
convey the extent of deviation (i.e., abnormality) of the results
of the specific patient under review. Such display techniques, may
include, as depicted in the presently illustrated embodiment, color
mapping of image pixels or voxels, and color coding of individual
cells in a table, wherein the color-coded cells each correspond to
a particular clinical test result and the color of the cell
corresponds to the magnitude of deviation of the patient result in
comparison to standardized data. Additional display techniques may
also include temperature gauges, spider graphs, dials, font
variations, annotation, and the like.
[0123] The exemplary visual representation 296 includes a plurality
of cells 304, at least some of which include patient non-image
deviation maps associated with respective clinical test results and
are color-coded to give a visual indication of the extent of
deviation of the patient from reference data for each test. For
instance, cell 306 may be associated with a functional test and
shaded in a color that generally represents the magnitude of the
deviation of the result of the functional test for the patient in
comparison to standardized results for the functional test.
Similarly, cells 308 and 310 may be associated with a cognitive
test and a blood sugar test, respectively, and may be filled with
particular colors to indicate the magnitude of deviations of the
patient results for such tests from standardized result data.
Although the present illustration depicts discrete color shades for
the various cells, it will be appreciated that a continuous color
range may instead be used, and that any one or more desired colors
may be employed to efficiently communicate the extent of deviation
of various clinical tests to a user. Additionally, it is noted that
the patient deviation maps displayed in the cells 304 may be based
on any suitable patient test having numerical or enumerated results
that can be compared to standardized data, and such maps are non
limited to those explicitly discussed herein.
[0124] Various image data may be displayed in a region 300 of the
exemplary visual representation 296. In the presently illustrated
embodiment, a plurality of structural patient deviation maps 314
and functional patient deviation maps 316 are illustrated in the
top and bottom portions, respectively, of the region 300. These
patient deviation maps may include various coloring or shading to
denote deviation of a patient anatomical region with respect to
standardized data. For instance, regions 318, 320, and 322 in the
structural patient deviation maps 314 may generally correspond to
portions of the patient brain exhibiting no or little deviation
from the standardized data, portions exhibiting moderate deviation,
and portions exhibiting severe deviation, respectfully. In
embodiments pertaining to the human brain, such structural patient
deviation maps 314 may include patient cortical thickness deviation
maps, which may be generated from MR image data. It is again noted,
however, that the presently disclosed techniques are not limited to
cortical thickness deviation data, or to brain images. Rather, the
presently disclosed techniques may be broadly applied to facilitate
quantification, visualization, and diagnosis of a wide array of
diseases and conditions.
[0125] The functional patient deviation maps 316 may also include
variously colored regions to indicate the magnitude of deviation of
that region for the patient with respect to standardized data. The
functional patient deviation maps 316 may include, among other
things, cerebral blood flow deviation or metabolic rate deviation
of patient data from the standardized data, and may, in one
embodiment, be generated from PET image data. In these maps 316,
regions 328 may correspond to no or little deviation from the
standardized data, while regions 330 and 332 may signify minor and
major deviations, respectively, of the patient from the
standardized data. The use of three different illustrative regions
in the structural patient deviation maps 314 and functional patient
deviation maps 316 is used for the sake of clarity and for
explanatory purposes. It should be appreciated that other colors or
shading may be used instead of or in addition to those illustrated
herein, and such coloring or shading may be provided in a
continuous range or provided at discrete levels.
[0126] The control panel 302 may facilitate presentation of other
data to a user and user-control of certain visualization processes.
For instance, in the presently illustrated embodiment, patient
information may be displayed in a region 340, population
information and selection control may be provided in a region 342,
and various system parameters, test data, or other information may
be provided in a region 344. In one embodiment, the population
region 342 may allow a user to select a particular set of
standardized data from a library of standardized data groups based
on a desired characteristic. For instance, a user may enter one or
more of a desired age range, gender, or race, and the system may
then display visual representations of patient deviations from the
selected standardized data set. In other words, in such an
embodiment, the user may select demographic characteristics of the
population segment of the standardized data to which the patient
will be compared for purposes of visualizing deviation.
Consequently, in one embodiment, the user may chose to visualize
patient results as a measure of deviation from a particular
standardized data set demographically matched to the patient.
[0127] Although the exemplary visual representation 296 includes
graphical representations of structural and functional deviations
in image data, as well as deviations with respect to non-image data
(e.g., clinical tests, laboratory tests, and so forth), other
visual representations having different data, or only subsets of
the deviation data visualized above, may be generated in other
embodiments. For instance, in certain embodiments the generated
visual representation may only include representations of deviation
with respect to either image data or non-image data, rather than
both.
[0128] For example, a visualization method 360 is illustrated in
FIG. 16 in accordance with one embodiment of the present invention.
The method 360 may include a step 362 of accessing patient image
deviation scores for multiple imaging modalities, such as CT, MR,
PET, SPECT, digital tomosynthesis, or the like. The patient image
deviation scores may be calculated through a comparison of patient
image data to standardized reference image data pertaining to a
population of individuals, as generally described above. Further,
in various embodiments, the patient image deviation scores may be
computed through comparison of patient static image data or patient
dynamic image data (e.g., video) acquired in a single imaging
system, or of patient longitudinal image data acquired over
multiple imaging sessions, to reference image data of a similar or
different type (i.e., static, dynamic, or longitudinal). The
accessed patient image deviation scores may be processed in a step
364 to generate a visual representation of patient deviation with
respect to the standardized image data, as generally discussed
above. The generated visual representation may be output in a step
366 to facilitate diagnosis of the patient in a step 368.
[0129] An additional exemplary visualization method 370 is
generally depicted in FIG. 17. The method 370 may include accessing
patient non-image deviation scores for dynamic or longitudinal data
in a step 372. Dynamic non-image data may include a substantially
continuous series of clinical test results over a given period of
time, while non-image longitudinal data may include test results
(or groups of test results) obtained in a staggered fashion (e.g.,
such as at 3 month intervals) over multiple data acquisition
sessions. As generally noted above, the patient non-image deviation
scores for such data may be calculated based on comparison of
patient non-image data to standardized non-image data. In some
embodiments, the patient non-image data on which the deviation
scores are based may include non-image data from different
modalities (e.g., cognitive data, neurological data, and the like).
The patient non-image deviation scores may be processed in a step
374 to generate one or more visual outputs indicative of deviation
of the patient non-image data from the standardized non-image data.
For instance, in one embodiment, a plurality of visual outputs may
be generated based on comparison of a sequence of longitudinal
patient non-image data to standardized non-image data. The visual
representations may then be output in a step a 376 to facilitate
diagnosis of the patient in a step 378. Multiple generated visual
representations may be output simultaneously or sequentially.
[0130] FIG. 18 is an exemplary diagram of an automatic comparison
workflow 400 generally depicting the automatic generation of a
severity index for various anatomical features of interest. The
automatic comparison workflow 400 may encompass a number of
anatomical features, such as structural or functional features of a
brain, a heart, or the like. To depict the possibility of such a
multitude of anatomical features in a comparison, the automatic
comparison workflow 400 is depicted as including a first anatomical
feature "A" 402, a second anatomical feature "B" 404, a third
anatomical feature "C" 406, an "N'th" anatomical feature "N" 408,
and so forth. The automatic workflow comparison of FIG. 18
represents a specific implementation of the more generalized
matching and presentation techniques described in U.S. Patent
Application Publication No. 2007/0078873 A1, published on Apr. 5,
2007, and entitled "COMPUTER ASSISTED DOMAIN SPECIFIC ENTITY
MAPPING METHOD AND SYSTEM," which is hereby incorporated by
reference in its entirety. For example, in this specific
implementation the various anatomical features 402, 404, 406, 408
represent various axes while the disease severity deviation maps
410, 412, 414, 416 discussed below represent different labels
associated with each axis, and so forth.
[0131] For each anatomical feature, a number of deviation maps
having variations in the extent, or severity level, of a disease or
a condition are provided. For example, for anatomical feature "A"
402, a number of reference deviation maps 410 having variations in
the extent of a disease or a condition associated with anatomical
feature "A" are provided. Similarly, sets of reference deviation
maps 412, 414, and 416 are provided, which exhibit the variations
in the extent of a disease or condition for each of the remaining
respective anatomical features through the Nth feature. As will be
appreciated by those of ordinary skill in the art, each of the
disease severity reference deviation maps within the respective map
sets 410, 412, 414, 416 are generated for the respective anatomical
feature 402, 404, 406, 408 and, in the case of image data (rather
than non-image data) reference deviation maps, may be further
categorized by a tracer or tracers (if one was employed) and by the
imaging technology employed. For example, reference deviation maps
within the respective deviation map sets 410, 412, 414, 416 may be
generated by magnetic resonance (MR) imaging, positron emission
tomography (PET), computed tomography (CT), single photon
emission-computed tomography (SPECT), ultrasound, optical imaging,
or other conventional imaging techniques and by using suitable
tracers in appropriate circumstances. As discussed above, the
reference deviation maps may also or instead be generated from
non-image data, including clinical data.
[0132] For each anatomical feature, the disease severity reference
deviation maps 410, 412, 414, 416 of the anatomical features are
ordered, as generally indicated by arrow 418, according to the
severity of the disease or condition or otherwise associated with a
severity of the disease or condition. For example, for anatomical
feature "A" 402, the disease severity reference deviation maps 410
may be ordered in ascending order from the least extent or amount
of the disease or condition, to the highest amount or extent of the
disease or condition.
[0133] In the depicted embodiment, eight reference deviation maps
are depicted in each of disease severity deviation map groups 410,
412, 414, 416 as representing the various disease severity levels
associated with each anatomical feature 402, 404, 406, 408. As will
be appreciated by those of ordinary skill in the art, however, the
number of reference deviation maps in the sets of disease severity
deviation maps 410, 412, 414, 416 is arbitrary and can be increased
or decreased depending on the implementation and the
characteristics of the reviewer. For example, in exemplary
embodiments where the comparison process is automated, the number
of reference maps within each of the groups of disease severity
deviation maps 410, 412, 414, 416 may contain more than eight maps,
such as ten, twenty, one hundred, and so forth. Further, though a
single disease severity reference deviation map is presently
depicted as corresponding to each ordered severity level for each
anatomical feature, each degree of severity for each anatomical
feature may actually have one or more than one disease severity
reference deviation map provided for comparison. For example, in
exemplary implementations where the comparison process is
automated, each severity level or severity index for an anatomical
feature 402, 404, 406, 408 may be represented by more than one
disease severity reference deviation map.
[0134] Various patient deviation maps 420 may then be evaluated
relative to the respective disease severity reference deviation
maps 410, 412, 414, 416 to determine an extent of disease or
condition in the patient deviation maps 420 in comparison to the
respective disease severity reference deviation maps. Each patient
deviation map 420 for an anatomical feature may be generated by
comparing acquired patient data to normative standardized
anatomical data for the respective anatomical feature. As will be
appreciated by those of ordinary skill in the art, the patient
deviation maps 420 may be derived from images acquired using one or
more suitable tracers (e.g., when needed to capture desired
functional information), from images acquired through other
techniques, or from non-image data, as described above. Therefore,
in an exemplary embodiment, the patient deviation maps 420 based on
image data are not only compared to a set of disease severity
reference deviation maps 410, 412, 414, 416 corresponding to the
same anatomical feature 402, 404, 406, 408, but also to those
reference maps in the set of disease severity reference deviation
maps 410, 412, 414, 416 generated from image data acquired using
the same or a comparable tracer or tracers, if present, and using
the same or a comparable imaging technology. In an exemplary
embodiment, the comparison between the one or more maps of patient
deviation maps 420 and the respective set of disease severity
reference deviation maps 410, 412, 414, 416 is performed
automatically, such as by pattern matching or other suitable
comparison techniques and routines.
[0135] For example, in one implementation patient deviation maps
420 generated from image data corresponding to the anatomical
feature "A" 402 may be automatically compared to the corresponding
set of ordered disease severity reference deviation maps 410 that
were generated from data acquired using the same tracer or tracers,
if a tracer was employed, and using the same imaging modality, such
as MR or PET. As will be appreciated by those of ordinary skill in
the art, patient deviation maps 420 and the respective disease
severity reference deviation maps 410, 412, 414, 416 to which they
are compared may vary depending on patient specific factors (such
as patient history, patient symptoms, and so forth) as well as
clinical factors (such as standard practice for the attending
physician and for the medical facility, preliminary diagnoses,
years of practice, and so forth).
[0136] In the presently illustrated example, each comparison
generates a severity index 422 that expresses or represents the
extent of disease in the respective patient deviation map 420, as
determined by comparison to the anatomical feature-specific disease
severity reference deviation maps 410, 412, 414, 416. As will be
appreciated by those of skill in the art, in those embodiments in
which the comparison is performed automatically, the severity index
422 may also be generated automatically. In such embodiments, a
reviewer or evaluator may simply be provided with a severity index
422 for each anatomical feature of interest or for which patient
deviation maps 420 were generated or processed.
[0137] In some embodiments, an aggregate patient severity score 424
is generated from the severity indices 422 using statistical
analysis 426, such as a rules-based aggregation method or
technique. In an exemplary embodiment, the aggregate severity score
424 is generated automatically, such as by automatic implementation
of the analysis 426 using suitable routines or computer-implemented
code. In such embodiments, a reviewer or evaluator may simply be
provided with an overall or aggregate severity score for the
patient.
[0138] In addition to calculating disease severity scores or
indices for a patient with respect to a single disease type, the
presently disclosed data processing system may also calculate a
combined disease severity score based on a plurality of different
disease types and severity levels. For instance, an exemplary
method 430 for determining a combined disease severity score for a
patient based on multiple disease types and severity levels is
depicted in FIG. 19 in accordance with one embodiment of the
present invention. The method 430 may include a step 432 of
accessing reference deviation data (such as reference deviation
maps or other data) for multiple disease types. Such reference
deviation maps may be standardized according to a demographic (or
other) characteristic. Additionally, the step 432 may also include
accessing reference deviation maps or data with respect to a
plurality of severity levels for one or more of the disease types.
In one embodiment, the reference deviation data may include
functional or structural deviation maps indicative of differences
between normal individuals and individuals diagnosed with
particular disease types, or diagnosed with severity levels of the
different disease types. Disease severity scores may be associated
with subsets of the reference deviation data, such as the different
reference deviation maps associated with various severity levels,
as generally discussed above. These individual disease severity
scores may also be accessed in a step 434.
[0139] The method 430 may also include selecting patient disease
severity levels in a step 436. Selection of patient disease
severity levels may be performed in a variety of manners. In one
embodiment, a user may compare a patient deviation map to a library
or database of known deviation maps indicative of functional or
structural deviation associated with various disease types and/or
severity levels. An exemplary visual reference library 484 of
known, standardized deviation maps corresponding to normal patients
and patients diagnosed with various disease types, is generally
illustrated in, and discussed in greater detail below with respect
to, FIGS. 21 and 22. In such an embodiment, the user may compare a
patient deviation map to those reference deviation maps included in
the library 484 to diagnose the patient as having one or both of a
particular disease type and severity level. To facilitate such
manual analysis and comparison, in one embodiment, one or more of
the reference deviation maps or patient deviation maps may be
displayed by a computing system, and a user may indicate (via a
user-interface) a selection of a particular severity level for each
disease type corresponding to the reference deviation map closest
to the patient deviation map.
[0140] In another embodiment, a computing system, such as the data
processing system 34, may be programmed to automatically compare
the patient deviation map to reference deviation maps in the
library of reference deviation maps and to automatically select the
closest matches. Alternatively, various disease scores may be
calculated based on given diseases and severity levels and compared
to a patient disease score to automatically determine and select
the closest match. In yet another embodiment, a computing system
may apply an algorithm to select a subset of the reference
deviation maps, from which a user may make the final
selections.
[0141] Following selection of patient severity levels for a
plurality of disease types, a combined disease severity score may
be automatically calculated in step 438. Finally, a report
including or based on the combined disease severity score may be
output in a step 440. As generally noted above, outputting of the
report, as well as other reports and data described herein, may
include outputting the report to memory, outputting the report to a
user, or outputting the report to a different software routine for
further processing.
[0142] The method 430 described above may be employed in connection
with a variety of anatomical regions and disease types, including,
but not limited to, brain disorders. An exemplary process for
evaluating such brain disorders may be better understood with
reference to block diagram 450, which is illustrated in FIG. 20 in
accordance with one embodiment of the present invention. Patient
image data 454 and patient non-image data 456 may be collected from
a patient 452. As noted elsewhere herein, such patient image data
may include images obtained through any of various imaging
modalities, and may include patient cortical thickness maps,
patient cortical thickness deviation maps or any other desired
image data. As also previously discussed, the patient non-image
data 456 may include numerous data types and information, such as
results of clinical tests and laboratory tests, family history,
genetic history, and so forth. Based on the patient image data 454,
it may be determined that the patient 452 has a vascular disease,
as generally indicated in block 458. Such a determination or
diagnosis may be output in a report 460. The patient image data 454
and the patient non-image data 456 may also be used to determine
whether the patient 452 has a neurodegenerative disease, as
generally indicated in block 462.
[0143] Block 464 generally represents a work flow for determining
patient severity levels for a plurality of brain disorders or
disease types 466. Separate pluralities of reference deviation maps
468 may be associated with each disease type, and each plurality
may generally represent different severity levels of its respective
disease type. Further, each reference deviation map may be
associated with a disease severity score (e.g., of the series S1 .
. . SN for each disease type). For example, in one embodiment, the
reference deviation map representative of the lowest severity level
of a particular disease may be associated with the lowest disease
severity score (i.e., 51) for that disease type, while other
reference deviation maps indicative of increasing severity levels
of the disease type may be associated with increasing disease
severity scores (i.e., S2, S3, . . . , SN). Either or both of
patient image data 454 and patient non-image data 456 may be
compared (block 470) to the sequence of reference deviation maps
for disease type A to determine a patient severity level 472 for
disease type A. The individual patient severity score X.sub.A for
disease type A may equal the disease severity score associated with
the reference deviation map for disease type A closest to the
patient data to which it is compared. Alternatively, if the patient
data suggests that the patient severity falls somewhere between two
of the reference deviation maps for disease type A, the individual
patient severity score X.sub.A may be computed from the two disease
severity scores associated with the individual reference deviation
maps closest to the patient data. The individual severity scores
for other disease types may be calculated in a similar manner based
on their own associated reference deviation maps.
[0144] Once the individual patient severity scores 472 for each
disease type is calculated, such individual scores may be utilized
to calculate a combined patient disease severity score, as
generally shown in block 474. The combined patient disease severity
score may be calculated through addition of the individual patient
severity scores, averaging of the individual patient severity
scores (which may be weighted as desired), or in any other suitable
fashion. Further, the combined patient disease severity score may
also indicate the relative contribution of each disease type to a
patient condition. For instance, the combined patient disease
severity score may indicate that Alzheimer's disease is the primary
contributing factor to patient dementia or some other condition. In
another embodiment, the combined patient disease severity score may
indicate the relative contribution of each of a plurality of
disease types to a patient condition. By way of example, the
combined patient disease severity score may indicate the relative
contribution of various brain disorders to patient dementia (e.g.,
40% AD, 30% FTD, 30% other). A report 476 based on or indicative of
the combined patient disease severity score may be output to a user
or to storage.
[0145] As noted above, reference images and deviation maps of an
exemplary visual reference library 484 are depicted in FIGS. 21 and
22 in accordance with one embodiment of the present invention. It
is noted that the presently depicted representations are merely
provided for illustrative purposes, and that an actual
implementation of a visual reference library may include different
or additional images. Indeed, various embodiments of a visual
reference library 484 may include a significantly greater number of
images, such as tens, hundreds, or even greater numbers of
reference images or maps, which may be standardized in various
embodiments as discussed above. It will be further appreciated that
images within the visual reference library 484 may be obtained via
one or any number of imaging modalities, and may include original
images, deviation maps such as those discussed above, or any other
suitable reference images. In the presently illustrated embodiment,
the reference images generally denote metabolic rate deviations
between normal individuals and individuals diagnosed with various
brain disorders. In other embodiments, however, other deviation
maps, such as cortical thickness deviation maps, cerebral blood
flow rate deviation maps, or even deviation maps for non-brain
anatomies, may be included in the visual reference library 484.
[0146] In the presently illustrated embodiment, the visual
reference library 484 is depicted as including a set of reference
images 486 for normal persons, and reference deviation maps 488 and
490 corresponding to patients clinically diagnosed with mild and
severe forms, respectively, of Alzheimer's disease (AD). The visual
reference library 484 may also include deviation maps 492
corresponding to patients diagnosed with diffuse cortical Lewy body
disease (DLBD) and deviation maps 494 representative of patients
clinically diagnosed with frontotemporal dementia (FTD). The visual
reference library 484 may also include additional deviation maps,
such as maps 496 associated with progressive supranuclear palsy
(PSP), maps 498 associated with multi-infarct dementia (MID), and
maps 500 associated with normal pressure hydrocephalus (NPH).
[0147] Technical effects of one or more embodiments of the present
invention may include the diagnosis of various patient disease
types and severity levels, as well as providing decision support
tools for user-diagnosis of patients. In one embodiment, technical
effects include the visualization of patient clinical image and
non-image information together in a holistic, intuitive, and
uniform manner, facilitating efficient diagnosis by an observer. In
another embodiment, technical effects include the calculation of
patient cortical deviation maps and reference cortical deviation
maps of known brain disorders, the calculation of additional
patient and reference deviation maps, and the combination of such
maps with other clinical tests, to enable quantitative assessment
and diagnosis of brain disorders.
[0148] In some embodiments, a system may be programmed or otherwise
configured to gather clinical information and create integrated
holistic views of the progression of statistical deviations of
clinical data of an individual patient from one or more normal
patient populations over time from longitudinal data. In addition,
methods for providing structured integrated holistic views of the
deviation of the clinical information across a given diseased
patient population when compared against a cohort of normal
controls, both at a single point in time and across multiple time
points (longitudinally) are also disclosed. Holistic viewers
described herein may display a normative comparison to thousands of
standardized and normalized data values concurrently. The resulting
holistic view can provide patterns of deviations from normal that
may indicate a characteristic pattern corresponding to known
diseases or abnormalities.
[0149] Also, various embodiments of the present disclosure may
provide a combined non-image (clinical, neurological, laboratory,
etc.) data and image data deviation view that produces results for
observing: a holistic patient-time-view (e.g., a single patient's
deviation evolving over time); a holistic population-view (e.g., a
set of patient cohort deviation compared to a normal cohort
deviation); and a holistic population-time-view (e.g., a set of
patient cohort deviation compared to a normal cohort deviation
evolving over time). In additional embodiments, one or more of
these views are employed to refine a normal cohort database using
holistic patient-time view; to refine the holistic patient-view
information displayed by highlighting clinical markers useful for
detection of a disease, based on an analysis of the salient
clinical data points observed in the respective holistic
population-view; and to refine the holistic patient-time-view
information displayed by highlighting clinical markers useful for
monitoring of a disease, based on an analysis of the salient
clinical data points observed in the respective holistic
population-time-view.
[0150] It is noted that each of these holistic viewers may be used
independently or together cohesively. These new additions may
assist in the establishment of appropriate clinically relevant
statistical hypotheses based on the holistic understanding. These
integrated holistic viewers may be used to compare deviations
across the different diverse parameters visualized. Using the
presently disclosed techniques, a user may be able to easily
compare the results of one parameter with another, and draw
conclusions therefrom. To facilitate such analysis, the various
parameters may be standardized and normalized. For instance, the
z-score space, illustrated by the formula below, provides a way to
compute a "z-score" deviation of the result of a particular
parameter from the results obtained from a cohort of age-matched
normals. The presently disclosed holistic viewers described in this
disclosure may visualize clinical deviation data in the
z-space.
z i = x i - .mu. n .sigma. n ##EQU00001##
While z-score space is just one technique, a number of different
ways may be used to normalize the data prior to visualization in
order to compare across parameters, and any suitable technique that
normalizes the relationship between parameters may be used in full
accordance with the present techniques. It is also noted that the
presently disclosed techniques may include transforming data (e.g.,
image data, non-image data, Z-scores, other scores discussed below,
and the like) representative of physical attributes of a patient
into other states that may facilitate detection and/or monitoring
of a disease in a patient or in a population. Further, while
various examples are provided herein within the context of NDD
detection, it is noted that the present techniques may also or
instead be used for analysis of other types of data for detecting
and/or monitoring other, non-NDD disease states, as well as in
other contexts unrelated to healthcare.
[0151] T-Viewer (Single Patient Over Time)
[0152] In one embodiment, an integrated holistic view of an
individual patient's clinical data trends over time is provided.
The view may include disparate types of clinical data, including
both image and non-image data in a manner that makes it easy for
humans to distinguish. Although graphs may be used to analyze a
longitudinal trend for a single clinical parameter, they are all
quite cumbersome when it comes to viewing multiple points and
monitoring their trends over time. FIG. 23 is one such example,
where the results of numerous parameters over three time points
(i.e., month 0, month 6, and month 12) are plotted on a chart,
making it increasingly difficult to identify and distinguish
clinically relevant trends as the number of parameters
increases.
[0153] In one embodiment of the present techniques, however, a
Patient-Time Holistic Viewer (T-viewer) may use the longitudinal
clinical data of an individual patient, with each parameter in the
standardized and normalized space to allow easy comparison from one
to another. A time trend score may then be calculated for each
parameter (T-score) from its respective longitudinal result data
points. The time score is visualized in the integrated viewer in a
manner such that a user can easily identify and distinguish trends,
in both the upward AND downward (negative) directions. For
instance, a bi-directional color scale could be used to color each
parameter presented, with the colors indicating the direction and
extent of the time trend. For example, for each parameter
presented, various shades of blue and yellow may be used to
represent negative and positive trends, respectively, of varying
magnitude (e.g., paler colors may represent smaller trends while
more intense colors may represent trends of greater magnitude).
FIG. 24 illustrates an example of such an embodiment. FIGS. 25-29
provide enlargements of various portions of FIG. 24 for the
convenience of the reader and further illustrating elements similar
to those of the Z score view previously described.
It should be noted that numerous methods might be used to calculate
the T-score prior to visualization. For example:
[0154] a) Total shift: Difference between the last time point
result and first time point result
t.sub.i=z.sub.i.sub._.sub.final-z.sub.i.sub._.sub.initial
As illustrated in FIG. 50, this technique simply provides the net
shift in deviation (and direction of the shift) of the specific
test score over the points in time that the clinical data was
collected. b) Weighted Shift: Sum of the differences between
successive time points, each weighted for the time elapsed
in-between the respective visits
t i = z visit - z prev_visit weighted_time _between _visits
##EQU00002##
As illustrated in FIG. 51, this is simply the sum of the shifts
observed from each time-point in clinical data collected to the
next, with care taken to weight each shift based on the amount of
time elapsed between the respective time-points. c) Initial
momentum: Averaged shift from the first time point
t i = n = 2 N ( z n - z 1 ) N - 1 ##EQU00003##
As illustrated in FIG. 52, this method provides the average shift
observed over all the time-points in clinical data collection, but
nothing that the shift at each time-point is always calculated
relative to the first (baseline) time-point. In essence this
provides the average shift observed for the test score over time,
relative to a baseline visit. d) Shifted momentum: Averaged shift
from initial time points, say for example the first three time
points
t i = n = 4 N ( z n - i = 1 3 z i 3 ) N - 3 ##EQU00004##
As illustrated in FIG. 53, this method is similar to the previous
methods described, except that the base-line score is an average of
a set number of initial visits. In essence, this provides an
average shift observed between a set number of initial visits and a
set number of subsequent visits. e) Other methods to calculate
trending/momentum of parameters over time, for example those used
to calculate shifting financial stock strengths.
[0155] D-Viewer (Multiple Patient Population Distributions) In
another embodiment, an integrated holistic view of specific patient
population's clinical data with respect to a population of normal
cohorts is provided. The view may include disparate types of
clinical data, including both image and non-image data in a manner
that makes it easy for humans to distinguish the distribution of
clinical parameter results across disease populations. Although
various graphs can be used to analyze results for a single clinical
parameter across population groups, they are all quite cumbersome
and impractical when it comes to visualizing and analyzing a large
number of parameters. FIG. 30 is one such example, where the
results of a single parameter are plotted over three population
groups, one of them being the normal control population. The
candlestick bar graph shows the shift of parameter values from one
population to another by highlighting the mean, upper and lower 95
percentiles, and maximum and minimum for each population. It is
easy to picture the increasing difficulty to identify and
distinguish clinically relevant trends as the number of parameters
increases.
[0156] In one embodiment, however, a Population Distribution
Holistic Viewer (D-viewer) uses the clinical data from multiple
patients categorized into population groups, with each parameter in
the standardized and normalized space to allow easy comparison from
one to another. A distribution score may then be calculated for
each parameter (D-score), based on its respective shift in the
specific population group under review from the normal population.
Finally, the distribution score can be visualized in the integrated
viewer, which may include views of parameters based on image and
non-image data, in a manner such that a user can easily identify
and distinguish parameter shifts from the normal population to the
specific population under review. For instance, a color scale could
be used to color each parameter presented, with the colors
indicating the extent of the distribution shift from the normal
population.
[0157] As with the T-score in the Patient-Time Holistic Viewer,
numerous methods may be used to calculate the D-score for each
parameter in the standardized and normalized space. These might
include the following: [0158] 1) Mean shift: Distance between the
mean scores of the two distributions [0159] 2) Weighted mean shift:
Distance between the mean scores of the two distributions weighted
using the distribution standard deviations, max/min, lower/upper 95
percentiles, etc. or a combination thereof [0160] 3) Other methods
to calculate the distance between two distributions
[0161] In one embodiment, the population holistic viewer (D-viewer)
provides the ability for a user to easily visualize a large number
of imaging and non-imaging clinical parameters, and assess the
distributions across a specific disease population group relative
to the normal population. For example, a single score may be
extracted from each parameter's shift across the two populations
and visualized in the integrated viewer.
[0162] While numerous techniques could be used to extract this
D-score as highlighted above, it may be desirable to assess the
clinical relevance of the distribution shift. As a result, it could
be argued that the actual extent of the shift is not as important
as the relative overlap between the two distributions. The greater
the extent of overlap between the two distributions, the greater
the number of patients in the indistinguishable `overlapping area`
and therefore the less clinically relevant the parameter. An ideal
parameter would show two distinct distribution curves with no
overlap, indicating that diseased patients demonstrate test results
clearly separable from those demonstrated by normal patients. FIGS.
31 and 32 provide examples of two overlapping distributions (such
as one distribution of normal patients and one distribution of
patients with Alzheimer's disease (AD)), while FIG. 33 generally
provides an example in which two distributions with very little
overlap.
[0163] Thus, in one embodiment, a distribution overlap score is
used as the D-score to visualize the extent to which a parameter
deviates in a specific population group when compared to a group of
age-matched normals. This technique could be used to compare a
plethora of different distributions with respect to normal
distribution.
[0164] Numerous methods could be applied in the actual calculation
of the extent of overlap between distributions. One example might
be a score ranging from 0 to 1, with 0 signifying 100% overlap
(least relevant parameters) and 1 signifying no overlap (most
relevant parameters). Note that the two distributions may first be
normalized to ensure that the area under each distribution is the
same, i.e., variation in the actual number of patients in each
population distribution should not inadvertently cause relative
weighting in the areas under their distributions. In this
technique, the D-score may be calculated as:
d i = 1 - Area_of _overlap i Total_area _between _two
_distributions i ##EQU00005##
where the two distributions are first normalized to ensure equal
area under each curve.
[0165] In addition, numerous techniques may be used to apply
further corrections to the overlap calculation used to determine
the D-scores. For example:
1) Threshold-based correction: Use of thresholds to remove
deviation scores belonging to groups of insignificant outliers. In
FIG. 34, a threshold is specified to exclude scores in a
distribution, where the relative proportion of patients with those
scores falls below a fixed amount. This enables the exclusion of
relatively insignificant regions in the distribution prior to
comparison. Note that these regions need not be at the extremities
of the distribution (as shown in the figure), but could also lie
in-between regions of significantly higher proportion. 2)
Percentile-based correction: Use of percentile based maxima and
minima to remove scores belonging to the outliers in each
distribution. In FIG. 35, maxima and minima are used to exclude
outliers from the extremities of the distributions. This enables us
to exclude a small portion of outlying patient scores that are
either extremely high or extremely low in their deviation, relative
to the general population's deviation spread.
[0166] In one embodiment, a Population Distribution Holistic Viewer
(D-viewer) may include representations of parameter deviations
between various groups of people, such as a normal population group
and some other population group, as generally depicted in FIG. 36.
Various portions of FIG. 36 are magnified in FIGS. 37-39 for the
convenience of the reader, and have specific elements similar to
the Z score view as previously described.
[0167] DT-Viewer (Multiple Patient Populations Over Time)
[0168] In another embodiment, an integrated holistic view of a
specific patient population's clinical data trends over time is
provided. The view may include disparate types of clinical data,
including both image and non-image data in a manner that makes it
easy for humans to distinguish. In one embodiment, this view may
combine aspects of the Patient-Time and Population Viewers
described above to show longitudinal trends in the clinical data of
a patient population group compared to the longitudinal trends of a
cohort population of age-matched normals. While graphs may be used
to analyze longitudinal trends of multiple parameters across
population groups, they are extremely cumbersome and impractical
especially as the number of parameters increases. FIG. 40 is one
such example, where the results of numerous parameters over three
time points (e.g., month 0, month 6, and month 12) are plotted for
three distinct population groups. Using such graphs, a user can
attempt to compare the trends observed in two disease populations
(e.g., the two charts in the center and the right side of FIG. 40)
with the trends observed in the normal population (e.g., the chart
on the left side of FIG. 40), although the number of parameters
involved may make such a comparison quite difficult.
[0169] In one embodiment of the present disclosure, however, a
Population-Time Holistic Viewer (DT-viewer) uses the longitudinal
clinical data of a specific patient population, with each parameter
in the standardized and normalized space to allow easy comparison
from one to another. A time trend score is calculated for each
parameter (T-score) from its respective longitudinal Z-scores. A
distribution score is then calculated on each parameter's time
trends (DT-score) for its respective shift in the specific
population group under review from the reference population.
Finally, the DT-score is visualized in the integrated viewer, which
may include views of parameters based on image and non-image data,
in a manner such that a user can easily identify and distinguish
parameter shifts from the reference population to the specific
population under review. For instance, a color scale could be used
to color each parameter presented, with the colors indicating the
extent of the distribution shift from the normal population.
[0170] Numerous techniques can be applied to calculate the T-scores
(as described in the T-viewer section) across time points for the
clinical parameters in the populations under review, following
which numerous techniques can be used to extract the DT-scores (as
described in the D-viewer section) from the longitudinal T-scores.
In one embodiment, generation of DT-views may be performed in a
manner similar to that of the D-view described above (see, for
example, FIG. 36), with individual T-scores used to calculate the
time trends instead of Z-scores for each population. An example of
such an embodiment is depicted in FIG. 41.
[0171] T-Viewer Application: True Normal Selection
[0172] In another embodiment of the present disclosure, a normal
cohort database may be refined using the holistic patient-time
viewer. A previously considered normal person's holistic view will
continue to show no change during the time course if the person is
truly normal. Persons that do not exhibit this "true normal"
behavior are then removed from the normal cohort population in the
database. This technique can be applied either manually or
automatically.
[0173] Manual--The user manually reviews the holistic patient-time
view of each person in the normal cohort population. Persons that
show longitudinal changes in their deviation data are noted, and
subsequently removed from the normal cohort.
[0174] Automatic--An automated algorithm is used to scan through
the patient-time views of each person in the normal cohort
database. Persons that show longitudinal changes in their deviation
data across the various clinical data points (individually or any
combinations thereof) above pre-specified thresholds are
automatically removed from the normal cohort.
[0175] D-Viewer Application: Extract Key `Detection` Parameters to
Refine Z-Viewer
[0176] In another embodiment, the holistic patient-view information
displayed may be refined by highlighting clinical markers useful
for detection of a disease, based on an analysis of the salient
clinical data points observed in the respective holistic
population-view. As may be appreciated from the present disclosure,
the holistic viewers may be used to identify the vital parameters
sufficient for the detection and/or monitoring of a disease. Ideal
candidates for the former may be identified in the D-viewer, and
may be fed back so that they can be elevated/highlighted in the
Z-viewer.
[0177] As described in the holistic population-viewer (D-viewer)
section, a parameter that demonstrates little or no overlap between
the distribution of disease population scores and the corresponding
normal cohort scores clearly indicates that the disease scores for
this parameter are distinct and separable from the corresponding
normal scores. As a result, any deviation from normal for this
parameter, even if relatively minor compared to other parameters
(i.e., relatively lower z-score than other parameters), could be
considered significant for diagnosis of the disease. In this
manner, "disease signatures" may be identified from population data
by identifying those parameters in which variations between a
normal population and a disease population are indicative of a
particular disease state. The actual use of results obtained from
the D-viewer to augment and refine the Z-viewer could be
accomplished in numerous ways, as described below in Appendix A.
Further, once identified, such disease signatures may be used to
diagnose patients based on deviations of patient data from that of
a group of normals with respect to the significant parameters of
the disease signature.
[0178] DT-Viewer Application: Extract Key `Monitoring` Parameters
to Refine T-Viewer
[0179] In an additional embodiment, the holistic patient-time-view
information displayed may be refined by highlighting clinical
markers useful for monitoring of a disease, based on an analysis of
the salient clinical data points observed in the respective
holistic population-time-view. As noted above, the holistic viewers
may be used to identify the vital parameters sufficient for the
detection and/or monitoring of a disease. Ideal candidates for the
latter may be identified in the DT-viewer, and may be fed back so
that they can be elevated/highlighted in the T-viewer.
[0180] As described in the holistic population-time-viewer
(DT-viewer) section, this view identifies a specific patient
population's clinical parameter trends over time and provides key
insights into parameter time-trend scores to be expected. Feeding
result information into the T-viewer may facilitate monitoring of
disease progression in an individual patient when his or her
clinical data is reviewed. When such data is compared or otherwise
reviewed in the context of the disease signatures, or in
conjunction with any of the overlapping or comparative methods
described herein, a "disease profile" is generated. The actual use
of results obtained from the DT-viewer to augment and refine the
T-viewer could be accomplished in numerous ways, as described below
in Appendix A. FIG. 42 illustrates an exemplary holistic viewer for
the normal population, and three corresponding disease signature
data views for the same test, wherein the sequence of the disease
signature data forms the disease profile.
[0181] The use of the population viewers (i.e., D-viewer and
DT-viewer) to refine the output of the patient viewers (i.e.,
Z-viewer and T-viewer) is generally depicted in FIG. 43.
Feeding Results from the Population Viewers to Augment the Patient
Viewers
[0182] The actual feeding of results obtained from the holistic
population and population-time viewers to augment the holistic
patient and patient-time viewers could be accomplished in any
suitable manner, such as in the following ways: [0183] I) Visual
highlighting of ideal candidates in the patient and patient-time
viewers--This could be done using a range of visual techniques such
as enlarging the key parameters (see FIG. 44), creating outline
overlays (see FIG. 45), physical separation of key parameters, use
of brighter colors, flashing icons, different shapes, etc. [0184]
Manual--The user manually selects parameters in the D and
DT-viewers to be visually highlighted in the Z and T-viewers
respectively. The applicable visual highlighting technique is then
applied in the Z and T-viewers. [0185] Automatic--The D and
DT-viewers automatically identify parameters that match certain
pre-defined deviation criteria, and apply the appropriate visual
highlighting technique in the Z and T-viewers. [0186] II) Weighting
of selected ideal candidates in the patient and patient-time
viewers--This is accomplished by weighting the z-scores and
t-scores of the more significant parameters, as generally depicted
in FIG. 46. All the z-score data in the Z-viewer is reprocessed and
weighted based on results from the D-viewer prior to visualization.
This results in a weighted color (or any other visualization
scheme) scale, where key parameters `light up` just as much or more
than other parameters even with relatively lower deviations.
Correspondingly, the technique is applied to t-score data in the
T-viewer based on results from the DT-viewer. [0187] Manual--The
user manually selects parameters in the D and DT-viewers, and
specifies weighting factors to be used in the Z and T-viewers. The
applicable weighting factor is used in the coloring/visualization
of the key parameters in the Z and T-viewers. [0188] Automatic--The
D and DT-viewers automatically identify parameters that match
certain pre-defined deviation criteria, and apply the appropriate
weighting factors in the Z and T-viewers. [0189] III) Visualizing
of only selected ideal candidates, and suppression of all
others--In this method, only the key parameters identified by the D
and DT-viewers are visualized in the Z and T-viewers, and all other
parameters are simply suppressed, as generally illustrated in FIG.
47. [0190] Manual--The user manually selects parameters in the D
and DT-viewers, and specifies these factors to be visualized in the
Z and T-viewers. All other parameters are suppressed in the Z and
T-viewers. [0191] Automatic--The D and DT-viewers automatically
identify parameters that match certain pre-defined deviation
criteria, and specify them for visualization in the Z and
T-viewers. All other parameters are suppressed in the Z and
T-viewers.
[0192] As described above, all of the feedback mechanisms could be
implemented manually, i.e., from a user visually identifying select
parameters in the D and DT-viewers and manually specifying them as
key candidates in the Z and T-viewers, or automatically, i.e., from
the use of automated algorithms to identify and select key
parameters in the D and DT-viewers based on their respective scores
and automatically specify them in the Z and T-viewers.
User Interface & Usage
[0193] As described above, the presently disclosed holistic viewers
enable the visualization of large amounts of diverse clinical data
in a unified space in a single view. Consequently, such viewers may
provide the user with a high-level view of the data in order to
identify areas of deviation from normal expected behavior, i.e.
clinical abnormalities. Details of the individual tests may be
abstracted at this level.
[0194] In some embodiments, the viewers may also tools that enable
a user to "drill-down" in the data and analyze the details of
individual abnormalities observed at the high-level holistic view.
Selection of tests for further analysis can be accomplished with a
range of User-Interface techniques, such as: [0195] 1) Moving the
mouse (hovering the cursor) over a selected test of interest [0196]
2) Mouse clicking on one or more test of interest [0197] 3)
Dragging and dropping tests of interest into a specific area on the
screen [0198] 4) Other menus, buttons and UI techniques used to
select specific tests for more detailed review and analysis
[0199] In various embodiments, numerous tools may be provided to
the user for further analysis and drill-down into specific tests,
and might include: [0200] 1) Reporting tools that display test
score statistics, and calculation details of z-scores, t-scores,
d-scores and dt-scores depending on the specific viewer [0201] 2)
Trending, graphing and plotting tools that visualize deviation of
the specific test score relative to baseline scores, time,
population distributions, etc. [0202] 3) Highlighting tools that
identify test score deviations that fit into user-specified
thresholds, categories or limits [0203] 4) Other analysis tools
that enable a user to drill-down into an abnormal test score and
identify the extent and potential causes of the abnormality
[0204] FIG. 48 below demonstrates such an interface. As the user
hovers over a specific test score, the table underneath dynamically
updates with the details (value, normal mean, normal standard
deviation, total number of normals, z-score etc.) that were used in
the calculation of the test score deviation. As depicted in FIG.
49, the user may click on a specific test score to generate a
time-trend plot showing the longitudinal variation of that test
score's deviation.
* * * * *