U.S. patent application number 16/123894 was filed with the patent office on 2019-04-04 for systems and methods for determining hepatic function from liver scans.
The applicant listed for this patent is HEPATIQ, INC.. Invention is credited to Dipankar Ghosh, John Carl Hoefs.
Application Number | 20190099147 16/123894 |
Document ID | / |
Family ID | 51263560 |
Filed Date | 2019-04-04 |
View All Diagrams
United States Patent
Application |
20190099147 |
Kind Code |
A1 |
Ghosh; Dipankar ; et
al. |
April 4, 2019 |
SYSTEMS AND METHODS FOR DETERMINING HEPATIC FUNCTION FROM LIVER
SCANS
Abstract
Systems and methods described herein determine an objective
metric for analyzing health of a patient's liver. In some
embodiments, the system may include a scanner that can detect
radiation counts responsive to administration of radioactive
compound to a patient. Further, the system may include an image
detection module that can access image data responsive to the
detected radiation counts by the scanner. The image detection
module can programmatically identify a first region of interest
corresponding to a liver of the patient from the image data. A
parameter calculator module can programmatically determine a first
attribute associated with the first region of interest and
calculate a first parameter indicating health of the liver of the
patient based at least in part on the first attribute associated
with the first region of interest.
Inventors: |
Ghosh; Dipankar; (Irvine,
CA) ; Hoefs; John Carl; (Irvine, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
HEPATIQ, INC. |
Irvine |
CA |
US |
|
|
Family ID: |
51263560 |
Appl. No.: |
16/123894 |
Filed: |
September 6, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14879741 |
Oct 9, 2015 |
10076299 |
|
|
16123894 |
|
|
|
|
14333370 |
Jul 16, 2014 |
9155513 |
|
|
14879741 |
|
|
|
|
61847313 |
Jul 17, 2013 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06K 9/3233 20130101;
G06T 2207/30056 20130101; G06T 7/50 20170101; G06T 7/0014 20130101;
G06F 19/00 20130101; A61B 6/032 20130101; G06T 2207/30008 20130101;
G06T 7/0012 20130101; A61B 6/5217 20130101; G06T 2207/10108
20130101; A61B 6/481 20130101; A61B 6/037 20130101; A61B 6/5205
20130101; G06T 7/13 20170101; A61B 6/469 20130101; A61B 6/50
20130101 |
International
Class: |
A61B 6/00 20060101
A61B006/00; G06T 7/13 20170101 G06T007/13; A61B 6/03 20060101
A61B006/03; G06T 7/00 20170101 G06T007/00; G06T 7/50 20170101
G06T007/50; G06K 9/32 20060101 G06K009/32 |
Claims
1.-20. (canceled)
21. A system for automating a determination of a risk of liver
disease for a patient from functional image data generated from a
functional scanner, the system comprising one or more hardware
processors configured to programmatically: access the functional
image data generated by the functional scanner; determine a first
shape of a first anatomical feature in the functional image data
corresponding to a liver of the patient; determine a location of a
first center point of the first anatomical feature; revise the
first center point of the first anatomical feature based on the
determined location of the first center point; identify a first
region of interest corresponding to the liver of the patient based
on the identified first shape and revised first center point;
determine a risk of liver disease based on the identified first
region of interest; and generate a user interface for display, the
user interface comprising: the first shape corresponding to the
liver; the first region of interest corresponding to the liver; and
the determined risk of liver disease.
22. The system of claim 21, wherein the one or more hardware
processors are further configured to: calculate a ratio based on
functional activity within the identified first region of interest
corresponding to the liver and parameters stored in picture
archiving and communication system (PACS) repository, wherein the
risk of liver disease is further determined based on the calculated
ratio.
23. The system of claim 21, wherein the determination of the risk
of liver disease further comprises using patient data stored in a
PACS repository.
24. The system of claim 21, wherein the liver disease is a presence
of hepatocellular carcinoma (HCC).
25. The system of claim 21, wherein the determination of risk
further comprises: accessing CT or MRI obtained liver region of
interest associated with the patient; and comparing the first
region of interest with the CT or MRI obtained liver region of
interest.
26. The system of claim 21, wherein the determination of risk
comprises determining a total count ratio from the functional image
data based at least on the first region of interest corresponding
to the liver.
27. An electronic method for automating a determination of a risk
of liver disease for a patient from functional image data generated
from a functional scanner, the method comprising: accessing the
functional image data generated by the functional scanner;
determining a first shape of a first anatomical feature in the
functional image data corresponding to a liver of the patient;
determining a location of a first center point of the first
anatomical feature; revising the first center point of the first
anatomical feature based on the determined location of the first
center point; identifying a first region of interest corresponding
to the liver of the patient based on the identified first shape and
revised first center point; determining a risk of liver disease
based on the identified first region of interest; and generating a
user interface for display, the user interface comprising: the
first shape corresponding to the liver; the first region of
interest corresponding to the liver; and the determined risk of
liver disease.
28. The method of claim 27, further comprising: calculating a ratio
based on functional activity within the identified first region of
interest corresponding to the liver and parameters stored in
picture archiving and communication system (PACS) repository,
wherein the risk of liver disease is further determined based on
the calculated ratio.
29. The method of claim 27, wherein the determination of the risk
of liver disease further comprises using patient data stored in a
PACS repository.
30. The method of claim 27, wherein the liver disease is a presence
of hepatocellular carcinoma (HCC).
31. The method of claim 27, wherein the determination of risk
further comprises: accessing CT or MRI obtained liver region of
interest associated with the patient; and comparing the first
region of interest with the CT or MRI obtained liver region of
interest.
32. The method of claim 27, wherein the determination of risk
comprises determining a total count ratio from the functional image
data based at least on the first region of interest corresponding
to the liver.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The present application claims the benefit of U.S.
Provisional Patent Application No. 61/847,313, filed Jul. 17, 2013,
the disclosure of which is hereby incorporated by reference in its
entirety.
BACKGROUND
[0002] Chronic liver disease is characterized by the gradual
buildup of scar tissue (fibrosis) in response to many forms of
chronic hepatic inflammation. This can lead to cirrhosis with a
decrease in hepatic function. Liver biopsy is one of the most
common methods of detecting a liver's health. The biopsy is,
however, invasive as it requires removing a portion of the liver
for analysis. Furthermore, liver biopsy analysis is subjectively
scored and may also vary depending on the location of biopsy.
SUMMARY
[0003] The systems and methods described herein can be implemented
by a computer system comprising computer hardware. The computer
system may include one or more physical computing devices, which
may be geographically dispersed or co-located.
[0004] Certain aspects, advantages and novel features of the
inventions are described herein. It is to be understood that not
necessarily all such advantages may be achieved in accordance with
any particular embodiment of the inventions disclosed herein. Thus,
the inventions disclosed herein may be embodied or carried out in a
manner that achieves or selects one advantage or group of
advantages as taught herein without necessarily achieving other
advantages as may be taught or suggested herein.
[0005] In certain embodiments, a system for detecting a liver
health parameter of a patient can include a single photon emission
computed tomography (SPECT) scanner that can obtain image data of
organs of a living patient, including a liver and a spleen of the
patient. The SPECT scanner can obtain the image data by at least
detecting radiation counts responsive to administration of a
radioactive compound to the patient. The system can further include
a memory device including an image detection module and a parameter
calculator stored thereon as computer-executable instructions. The
system can further include a hardware processor that can implement
the image detection module by executing the computer-executable
instructions to at least access the image data output by the
scanner. The instructions may further include programmatically
identify a first region of interest corresponding to the liver of
the patient from the image data, the first region of interest
comprising a bounded region around the liver of the patient and
being indicative of a size of the liver, wherein the size of the
liver is correlated with a health condition of the liver, such that
a size of the first region of interest is indicative at least in
part of the health condition of the liver. In some embodiments, the
instruction can further include programmatically identify a second
region of interest corresponding to a spleen of the patient from
the image data. Additionally, the hardware processor may also
implement the parameter calculator by executing computer-executable
instructions to at least programmatically determine a first
attribute associated with the first region of interest. The
instructions can further include programmatically determine a
second attribute associated with the second region of interest. In
addition, the instructions can include calculate a first parameter
indicative of the health condition of the liver of the patient
based at least in part on the first attribute associated with the
first region of interest and the second attribute associated with
the second region of interest. In some embodiments, the
instructions can include output, in a computer-generated graphical
user interface, an indication of the first parameter for
presentation to a clinician, enabling the clinician to make a
clinical care decision for the patient.
[0006] The system of the preceding paragraph can have any
sub-combination of the following features: wherein the first
parameter includes perfused hepatic mass; wherein the first
attribute includes a representation of radiation counts in the
first region of interest; wherein the image detection module can
compare a geometric property of the first region of interest
relative to the second region of interest; wherein the memory
further includes a user interface module that can include
additional instructions configured to generate and output a second
user interface, the second user interface can provide functionality
for the clinician to input a command to modify the first region of
interest; wherein the image detection module can combine a
plurality of frames from the image data, said frames corresponding
to planes transverse to the patient's body; wherein the image data
includes a plurality of frames; wherein the image detector module
can programmatically detect the first region of interest from a
first frame and programmatically detect the second region of
interest from a second frame, wherein said first frame corresponds
to a different plane with respect to the patient's body than the
second frame; wherein the graphical user interface includes a
display of the first region of interest.
[0007] Additionally, in certain embodiments, a method for detecting
a liver health parameter of a patient can include receiving image
data comprising a representation of detected radiation counts
corresponding to one or more organs of a patient. The method can
also include programmatically identifying a first region of
interest corresponding to a liver of the patient from the image
data. In addition, the method can include programmatically
identifying one or more additional regions of interest
corresponding to one or both of a spleen of the patient and marrow
of the patient from the image data. Further the method can include
determining a first parameter indicative of health of a patient
based at least in part on a first attribute associated with the
first region of interest and optionally also based on a second
attribute associated with the second region of interest. In some
embodiments, the method can include programmatically generating an
output responsive to the first parameter for presentation to a
clinician, wherein the output comprises one or more of a value of
the first parameter and a health report associated with the first
parameter. In some embodiments, at least said programmatically
identifying a first region of interest corresponding to a liver of
the patient from the image data is performed under control of
processing electronics.
[0008] The method of the preceding paragraph can have any
sub-combination of the following features: wherein the first
parameter includes perfused hepatic mass; wherein the first
attribute includes a representation of radiation counts in the
first region of interest; further including detecting a first
centroid associated with the first region of interest and detecting
a second centroid associated with the second region of interest;
further including programmatically identifying a third region of
interest based at least in part on the detected first and second
centroids.
[0009] In certain embodiments, a system for detecting a liver
health parameter of a patient can include a hardware processor. The
system can receive scanner output data responsive to detected
radiation from a radiation detecting scanner, said scanner output
data responsive to a radioactive compound administered to a
patient. The system can further apply image processing techniques
to detect two or more separate tissue masses in the scanner output
data, at least one of the two or more separate tissue masses
corresponding to an organ selected from a group consisting of a
liver and a spleen. The system can also determine a parameter
corresponding to function of one or more of the two or more
separate tissue masses of the patient. In some embodiments, the
system can output a graphical indication of the parameter for
presentation on a display.
[0010] The system of the preceding paragraph can have any
sub-combination of the following features: wherein the system can
further detect a bone marrow region based at least in part on the
detected one or more organs; wherein the parameter includes one of
a liver volume, a spleen volume, a perfused hepatic mass, a total
count ratio, a staging indicator, an estimated peritoneoscopic
score, a normalized liver volume, a normalized spleen volume, a
highest average concentration, liver counts, a liver spleen index,
a liver bone marrow index, a liver length, a spleen length, spleen
counts, bone marrow counts, and a hepatic activity index; wherein
the scanner output data includes at least one of or more frames
corresponding to images of the patient in a plane transverse to a
long axis of the body of the patient; wherein the system can
further combine a plurality of frames from the image data, said
frames corresponding to planes transverse to a long axis of the
body of the patient.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Embodiments disclosed herein are described below with
reference to the drawings. Throughout the drawings, reference
numbers are re-used to indicate correspondence between referenced
elements. The drawings are provided to illustrate embodiments of
the inventions described herein and not to limit the scope
thereof.
[0012] FIG. 1 illustrates an embodiment of a computing environment
including a quantitative liver spleen scan diagnostics (QLSSD)
system that can enable clinicians to quantitatively analyze health
of a patient's liver.
[0013] FIG. 2 illustrates an embodiment of an organ health
detection process.
[0014] FIG. 3 illustrates an embodiment of a liver health detection
process.
[0015] FIG. 4 illustrates an embodiment of a process for
identifying liver, spleen, and bone marrow ROI.
[0016] FIG. 5 illustrates an example graphical representation of
the embodiment of processes described with respect to FIGS. 3 and
4.
[0017] FIG. 6 illustrates liver, spleen, and bone marrow ROI
detected from the SPECT scan.
[0018] FIG. 7 illustrates an embodiment of a process to detect
three dimensional ROIs.
[0019] FIG. 8 illustrates an embodiment of a process for predicting
post-surgery liver health.
[0020] FIG. 9 illustrates an embodiment of a user interface that
can enable clinicians to generate health parameters corresponding
to liver's healthiness.
[0021] FIG. 10 illustrates an embodiment of a PACS import user
interface.
[0022] FIG. 11 illustrates an embodiment of a user interface that
allows clinician to review and modify automatically generated
ROIs.
[0023] FIG. 12 illustrates an embodiment of a user interface
showing length of a patient's spleen.
[0024] FIG. 13 illustrates an embodiment of a user interface that
can enable clinicians to adjust frame range.
[0025] FIG. 14 illustrates an embodiment of a user interface
including a report with the PHM parameter.
[0026] FIG. 15 illustrates an embodiment of a user interface that
includes a suggested impression and enables clinicians to enter
their own impressions.
[0027] FIG. 16 illustrates an embodiment of a user interface that
enables sending a report to PACS.
[0028] FIG. 17 illustrates an embodiment of a user interface
including a report showing a trend in PHM parameter over time.
[0029] FIG. 18 illustrates an embodiment of a user interface
including an example report showing liver health parameters.
DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS
I. Introduction
[0030] The liver is a vital organ with a wide range of functions,
including filtering blood coming from the digestive track before
passing it to the rest of the body. The liver also detoxifies
chemicals and metabolizes drugs. Diseases can reduce the
functionality of the liver. Long-term damage to the liver from any
cause can lead to permanent scarring, called cirrhosis. Assessing
health of a liver is critical in order to predict prognosis and for
treatment of patients. A liver biopsy is most commonly used test to
determine hepatic functionality. But, the test is invasive and
subjective, depending on the analyzing clinician. The results may
also depend on the location of the biopsy. Furthermore, the
decrease in hepatic function relates more to a patient's health
than information gathered from routine blood tests or even
fibrosis. Accordingly, in some embodiments, the system described
herein can generate quantitative measurement of hepatic function.
In addition to its ordinary meaning, hepatic can mean of or
pertaining to function of the liver. The hepatic function can
correspond to healthiness or functionality of liver.
[0031] Non-invasive method of detecting health of a patient's liver
can include analyzing images generated by a Single Photon Emission
Computed Tomography (SPECT) scanner. The SPECT scanners can
generate images responsive to administrating radioactive compound
to patients. Because one of the functions of the liver is to filter
blood, the radioactive compound is filtered by the liver and the
SPECT scanner can pick the radiation counts to detect absorption of
the radioactive compound. The absorption depends on the health of
the liver. A healthy liver absorbs the majority of the compound,
and the radiation detected by the scanner may mostly be
concentrated in the liver. However, when the liver is diseased,
more of the radioactive compound can leak out of the liver and flow
into, for example, the spleen and/or bone marrow. Accordingly, the
radioactive counts from the SPECT scan responsive to absorption of
the compound can indicate hepatic function.
[0032] Abstracting information from the image scans can be
difficult. The analysis may change for different SPECT scanners.
Further, significant training might be required for clinicians to
determine parameters from the images. In addition, the analysis may
suffer from subjective determinations of clinicians (who may, for
example, hand-draw regions of interest as described more in detail
below). The regions of interests (ROIs) calculated using a QLSSD
system (discussed below) can be robust compared to hand drawn ROIs.
For instance, there may be variations in hand drawn ROIs between
different clinician. Moreover, it may be cumbersome and
time-intensive to hand draw ROIs. Also, the clinicians may not be
able to detect counts appropriately from the image as it may depend
on the contrast levels of images and may vary between scanners. In
some instances, splenectomy, liver--spleen overlap, and anatomical
variability may also increase the difficulty in analyzing SPECT
images for the clinicians. Accordingly, the QLSSD system can
diagnose and identify stages of chronic liver diseases based on the
image scans.
[0033] This disclosure describes embodiments of a quantitative
liver spleen scan diagnostics (QLSSD) system that can provide
clinicians a tool to determine a patient's health based on
characterization of the patient's liver through one of the scanning
techniques, e.g. from a SPECT scanner. In some embodiments, the
QLSSD system can calculate one or more numerical parameters that
may be correlated with the patient's liver health. The QLSSD system
can also generate text or graphical impressions based on the
calculated numerical parameters to report results based on the
scans.
II. Example QLSSD System
[0034] FIG. 1 illustrates an embodiment of a computing environment
100 for providing clinicians with access to QLSSD system 120 to
determine a patient's health based on analyzing scans of the
patient's organ. In an embodiment, the QLSSD system 120 determines
the patient's health based on analyzing images of liver from a
SPECT scanner. The computing environment 100 can include clinician
systems 108 that can access the QLSSD system 120, which may include
one or modules to determine the patient's hepatic function.
[0035] For instance, the QLSSD system 120 can include an image
retriever module 122 that can retrieve images corresponding to
scans of a body part, e.g. liver. In an embodiment, the image
retriever 122 can receive raw images directly from the SPECT
scanner 106. In other embodiments, the image retriever 122 can
receive images from a PACS (Picture Archiving and Communication
System) repository 102. The image retriever 122 can also receive
images from a storage medium such as a compact disc (CD), a
portable hard drive, etc. The PACS system 102 may store images in a
DICOM (Digital Imaging and communication in Medicine) format. The
PACS system 102 may also include other non-image data regarding
patients. The image retriever 122 can also receive images of
different formats (e.g. jpeg, png, pdf, bmp, CT scanner raw files,
MRI raw files, PET raw files, x-ray raw files, etc.). In an
embodiment, the image retriever 122 retrieves images from the PACS
or SPECT scanner over a network 104. The image retriever 122 may
get the images from PACS 102 in response to an input from clinician
system 108. In some embodiments, the image retriever may
automatically receive the images from the SPECT scanner 106 after a
predetermined time interval.
[0036] The QLSSD system 120 can include an image detection module
124 to analyze the images retrieved by the image retriever 122. The
image detection module 124 can process the images and identify one
or more regions of interests (ROI) from the images as described
more in detail below. The regions of interests can include organs,
tissues, tissue masses, bones, etc. In an embodiment, the ROI can
include a region corresponding to a patient's liver. The image
detection module 124 can process images generated by SPECT scanner.
In some embodiments, the image detection module 124 can also
process images produced from a CT scanner or a MRI machine or from
another type of scanner. The image detection module 124 can use
information obtained from one type of image to process another type
of image for the same patient. For instance, the image detection
module 124 can use information detected from CT scan of the patient
to detect regions of interest in the SPECT scans. The image
detection module 124 may store analyzed images in patient data
repository 140 or transmit it back to PACS 102. In some
embodiments, the image detection module 124 may include internal
checks to ensure that the ROI corresponds to the particular organ.
If the detection module 124 determines that the detected ROI does
not accurate reflect a particular organ, then it can give
clinicians an option to override the automatic ROI detection as
described more in detail below.
[0037] The parameter calculator 126 of the QLSSD system 120 can
determine one or more attributes from the region of interests
identified by the image detection module 124. For instance, the
parameter calculator 126 can determine a length of a region of
interest corresponding to the spleen. The parameter calculator 126
can also determine a volume or concentration corresponding to a
region of interest in terms of total counts or voxels. In an
embodiment, a voxel represents a value (e.g. radiation count) in a
three dimensional space. In some embodiments, the parameter
calculator 126 can calculate parameters corresponding to health of
a liver. Some of the example parameters include Perfused Hepatic
Mass (PHM), Hepatic Activity Index (HAI), Total Counts Ratio (TCR),
Normalized Liver Volume (NLV), and Normalized Spleen Volume (NSV).
These parameters may quantitatively indicate hepatic function. The
calculated parameters can be stored in the patient data repository
140. In some instances, the calculated parameters can be
transmitted and stored in PACS along with other data for that
patient. The parameters can also be transmitted over a network to a
clinician system. The parameter calculator 126 can also generate a
trend based on stored parameters to allow clinicians to monitor
health of patients over time.
[0038] The user interface module 128 can interact with one or more
other modules of the QLSSD system 120 to generate one or more
graphical user interfaces. In some embodiments, the user interfaces
can be one or more web pages or electronic documents. The user
interface module 128 can also receive data such as patient
information from the clinician systems 108. In some instances, the
user interface module 128 may receive commands from the clinician
systems 108 to initiate one or more functionalities of the QLSSD
system. The data can be stored by patient data repository 140.
Embodiments of user interfaces are described in detail below.
[0039] The QLSSD system 120 can be implemented in computer hardware
and/or software. The QLSSD system 110 can execute on one or more
computing devices, such as one or more physical server computers.
In implementations where the QLSSD system 110 is implemented on
multiple servers, these servers can be co-located or can be
geographically separate (such as in separate data centers). In
addition, the QLSSD system 110 can be implemented in one or more
virtual machines that execute on a physical server or group of
servers. Further, the QLSSD system 110 can be hosted in a cloud
computing environment, such as in the Amazon Web Services (AWS)
Elastic Compute Cloud (EC2) or the Microsoft.RTM. Windows.RTM.
Azure Platform. The QLSSD system 110 can also be integrated with
scanners 106 and 110 through software or hardware plug-in or an API
(application programming interface). In some embodiments, the
clinician systems 108 may implement some or all of the modules of
the QLSSD system 120. For instance, the clinician systems 108 may
implement the user interface generator module 128, while the rest
of the modules are implemented remotely on a server. In other
embodiments, a plugin to the QLSSD system 110 may be installed on
to a third party tool.
[0040] The clinician systems 108 can remotely access the QLSSD
system 110 on these servers through the network 104. The clinician
systems 108 can include thick or thin client software that can
access the QLSSD system 110 on the one or more servers through the
network 104. The network may be a local area network (LAN), a wide
area network (WAN), such as the Internet, combinations of the same,
or the like. For example, the network 104 can include a hospital's
private intranet, the public Internet, or a combination of the
same. In some embodiments, the user software on the clinician
system 108 can be a browser software or other application software.
The clinician system 108 can access the QLSSD system 110 through
the browser or application software.
[0041] In general, the clinician systems 108 can include any type
of computing device capable of executing one or more applications
and/or accessing network resources. For example, the clinician
systems 108 can be desktops, laptops, netbooks, tablet computers,
smartphones, smartwatches, augmented reality wear, PDAs (personal
digital assistants), servers, e-book readers, video game platforms,
television set-top boxes (or simply a television with computing
capability), a kiosk, combinations of the same, or the like. The
user systems 108 include software and/or hardware for accessing the
QLSSD system 110, such as a browser or other client software.
III. Organ Health Detection Process
[0042] FIG. 2 illustrates an embodiment of an organ health
detection process 200 for calculating a parameter corresponding to
the health of the organ (or more generally a mass of tissue). The
parameter can be a numerical, graphical, or textual indicator. The
organ health detection process can be implemented by the system
described above. For illustrative purposes, the process 200 will be
described as being implemented by components of the computing
environment 100 of FIG. 1. The process 200 depicts an example
overview of calculating a parameter based on scanned images of a
patient.
[0043] The process 200 can begin at block 202 with receiving image
data responsive to imaging a patient. The image retriever module
122 can receive image data corresponding to SPECT, CT, or MRI scans
of the patient's organ. The received images may include one or more
anatomical features (e.g. liver, spleen, bone marrow, etc.). The
image detection module 124 can automatically detect these
anatomical features using one or more object detection techniques
(e.g. morphology, edge detection, centroid search, histogram, etc.)
at block 204.
[0044] The detected anatomical features can be used to extract
quantitative information from the image scan. For example, at block
206, the parameter calculator 126 can calculate an attribute
associated with the detected anatomical feature. In an embodiment,
the detected anatomical feature can include a spleen, and the
corresponding attribute can be the length of the spleen. A more
detailed example with respect to liver and spleen will be discussed
below with respect to FIGS. 3 and 4. The parameter calculator 126
can further determine an indicator or a parameter associated with
the health of the detected anatomical features or the health of the
patient at block 208. In one embodiment, the indicator includes
Perfused Hepatic Mass (PHM), which may directly correlate with
hepatic function or healthiness of the liver. The PHM can be a
numerical indicator or graphical output based on a numerical
indicator and may provide an objective standard to determine a
patient's liver health.
[0045] The process 200 can be used to determine a parameter
corresponding to health of different anatomical features of a
patient. As discussed above, the process 200 can be used to
determine a parameter corresponding to hepatic function. In another
example, the process 200 can be used to determine a parameter
corresponding to a patient's heart or cardiac output, pulmonary
nodule classification, or kidney function, or function of any other
organ or tissue mass of a patient.
IV. Liver Health Detection Process
[0046] FIG. 3 illustrates a more specific example process 300 of
the organ health detection process 200 described above. The process
300 can enable clinicians to obtain a parameter corresponding to
healthiness of the patient's liver. The liver health detection
process 300 can be implemented by the system described above. For
illustrative purposes, the process 300 will be described as being
implemented by components of the computing environment 100 of FIG.
1. The process 300 depicts an example overview of calculating a
parameter associated with hepatic function using scanned SPECT
images of a patient.
[0047] The process 300 may begin with acquiring scanned images from
a SPECT scanner at block 302. The SPECT scanner generates images by
measuring radiation counts responsive to administrating radioactive
compound to a patient. In an embodiment, the images are generated
30 minutes after administration of the compound. The time may vary
between patients, but typically it takes about half an hour for the
radioactive compound (e.g. Technetium-99 metastable) to be filtered
from the blood by the liver. For a patient with a healthy liver,
most of the radioactive compound will be found in the liver.
However, in a diseased liver, the compound can leak into the spleen
and bone marrow. Since most of the compound may be found in the
liver, spleen, bone marrow (near vertebrae) region, the SPECT
scanner output can provide mechanism for separating liver, spleen,
and bone marrow from rest of the organs. An example summarized
transaxial SPECT scan of liver, spleen, and bone marrow is shown in
FIG. 6.
[0048] As discussed above, the scanned images can be received
directly from the scanner or via PACS. The image retriever module
122 can automatically process the received images depending on the
source at block 304. The received images can include multiple
orientations. In one embodiment, the image retriever module 122 can
retrieve planar posterior and transverse SPECT images from SPECT
scanner or PACS. The transverse SPECT images may include one or
more frames corresponding to planes perpendicular to the long axis
of the patient's body. The planar posterior images may include one
or more frames taken from the perspective of the patient's backside
and may correspond to planes parallel to the long axis of the
patient's body. Other image views include anterior, oblique,
sagittal, coronal, reformatted, secondary captures, or derived
images. In some embodiments, the image retriever module 122 can
directly process raw scanner data instead of image data. The
received images may be divided into multiple frames spanning an
area of the patient's body. Accordingly, the received images can
include a three dimensional perspective of the SPECT scan.
[0049] The image detection module 124 can process the received
images at block 124. In an embodiment, the image detection module
124 can generate a combined transaxial image (CTI) from the
transverse liver-spleen images obtained from the SPECT scanner at
block 306. In an embodiment, the CTI is a summarized transaxial
image (STI), which can be a combination or a sum of each of the
transaxial images covering the liver-spleen area. In other
embodiments, the CTI is generated using through voxel-by-voxel
averaging of the transverse images. Other methods for generating
CTI that can be used include using median, mode, maximum intensity,
auto-correlation, and similar statistical techniques. In an
embodiment, maximum intensity transaxial image (MITI) is detected
by selecting the highest intensity pixel at (x,y) from some or all
the transaxial images covering the liver-spleen area. Accordingly,
the final MITI may have highest intensities from some or all the
transaxial images for each pixel. In some embodiments, MITI can
provide a better contrast than STI. The image detection module 124
can combine the transaxial images before detection of region of
interests. In some embodiments, the regions of interests may be
detected prior to combination of frames.
[0050] The image detection module 124 can use the CTI to detect
regions of interests. The regions of interest can correspond to
anatomical features of the patient. In an embodiment, the image
detection module 124 detects liver, spleen and bone marrow. The
image detection module 124 can locate the liver based on a
comparison of intensities in the CTI. FIG. 5 illustrates a
simplified diagram to show the steps for detecting and separating
liver, spleen and bone marrow ROIs.
[0051] The image detection module 124 can select frame boundaries
and conduct directional searching to locate the liver. At block
308, the image detection module 124 can identify a body contour 520
as shown in step 502 in FIG. 5 by comparing intensities. In
general, counts will be near zero outside of the body. A healthy
liver will include much higher intensity counts than surrounding
tissue. Even a diseased liver may include at least some portions
with high intensity counts.
[0052] At block 310, the image detection module 310 can determine
liver boundary threshold. Step 504 of FIG. 5 illustrates an example
for determining liver boundary threshold. The image detection
module 124 can identify a border region 522 around the body contour
520 to determine the threshold. In an embodiment, the threshold is
the average pixel intensity in the border region or some other
value. The threshold can also be the maximum or minimum pixel
intensity in the border region. In some embodiments, the threshold
can be determined using a logarithmic formula.
[0053] Using the threshold value, the image detector 310 can use
directional searching to locate a region of interest corresponding
to the liver at block 312. As an example, the image detector 124
can select the north-west corner of the body contour (although any
other starting region may be selected) and move across the rows and
columns of pixels to determine when the intensity count exceeds the
threshold. Accordingly, the image detector 310 can find a pixel
location that exceeds the threshold value indicating that the pixel
is most likely inside the liver. In one embodiment, a value of 20%
above the tissue threshold is used. After finding the pixel
location, the image detection module 124 can implement a
directional searching to identify the boundary points of the liver.
As shown in step 506 of FIG. 5, the image detector 124 can search
along eight directions in a star pattern from the pixel location
(526). In other embodiments, the image detector 124 can also
implement other methods of searching, e.g. 16-point star, etc. The
boundary points can be identified by comparing intensities of
pixels with the threshold value as discussed above. In some
embodiments, the image detection module 124 can also use sampling
to identify boundary points of the liver.
[0054] Accordingly, the image detector 124 can identify (for
example) 8 boundary points of the liver. At block 314, using these
boundary points, the image detector 124 can calculate a first
centroid location. In some embodiments, the image detector 124 can
quickly locate the centroid of the organ without using
trigonometric calculations by using straight combinations of matrix
indices. For example, the matrix indices may be combined using
summations. After finding the first centroid, the image detector
124 can repeat the process to locate the second centroid using
directional searching from the first centroid location. In some
embodiments, the image detector can rotate the star search
direction. For example, the rotation angle can be offset by 10
degrees in between centroid search. The eight point star search may
be repeated multiple times until the location of the centroid
appears to converge. In some embodiments, the number of iterations
can be fixed. Step 508 of FIG. 5 illustrates the final centroid
location 528. Accordingly, the image detector 124 can identify
boundary points of the liver and the centroid of the liver.
[0055] In blocks 316 to 318 of the process 300, the boundary points
and the centroid of spleen can be identified in a similar manner as
discussed above with respect to the liver. In most patients, the
spleen is located in a south-east direction relative to the liver.
Thus, the image detector module 124 can conduct a similar search
from the south-east corner of the body contour 520 to identify the
spleen boundary points and centroid. Step 510 of FIG. 5 illustrates
both the liver 524 and the spleen 530.
[0056] The image detector 124 can automatically validate the
detected liver and spleen. At block 324, the image detector 124 can
calculate the distance between the liver and spleen centroids and
compare with an acceptable range to validate the detected liver and
spleen. If the distance is not within a range, e.g., it is less
than the minimum or more than the maximum value, the image detector
can notify one or more modules of the QLSSD system 120. The QLSSD
system 120 can notify the clinician systems that organ detection
might have failed and may give an option to the clinician to
manually draw the ROIs. The user interface generator module 128 can
generate user interfaces for the clinician to draw the ROIs.
[0057] The image detector 124 can also validate for overlaps. For
instance, if the liver and spleen regions overlap, the QLSSD system
120 may send a message to the clinician system that enables a
clinician to select a brightest area of the liver or indicate that
liver is not visible. The clinician may also be able to select the
brightest spleen area. Based on the selection, the image detector
124 can use directional searching and centroid analysis to
re-identify the liver and spleen. In some embodiments, clinicians
may be given an option to select the liver and spleen pixels before
any of the search process. Thus, the QLSSD system 120 can validate
organ detection.
[0058] If the liver and spleen regions are validated, then at block
330, the image detector 124 can automatically draw ROIs around
liver, spleen, and bone marrow as described more in detail with
respect to FIG. 4. Once the ROIs are determined, the parameter
calculator 126 can calculate parameters corresponding to hepatic
function as discussed below. In some embodiments, the process 300
includes the step of preparing and generating a report including
impressions at blocks 332-334. The user interface generator 128 can
prepare reports based on the calculated parameters and a lookup
table. The lookup table may be stored in the data repository 140.
The lookup table can store appropriate impressions for different
ranges of parameters. An example impression is shown in FIG.
15.
V. Regions of Interest Analysis
[0059] FIG. 4 illustrates an example ROI detection process 400 for
determining ROIs based on the detected liver and spleen centroids
using the process illustrated in FIG. 3. The ROI detection process
400 can be implemented by the system described above. For
illustrative purposes, the process 400 will be described as being
implemented by components of the computing environment 100 of FIG.
1.
[0060] In an embodiment, once the liver and spleen centroids are
identified along with boundary points, the image detector 124 can
proceed with identifying region of interests to segment liver,
spleen and bone marrow. At block 402, the image detector 124 can
fit in a geometric shape such an elliptical shape over the liver
and spleen based on the respective boundary points and centroid
detected by the process 300 discussed above. As an example, the
image detector can fit an ellipse based on a least squared error
reduction method. In some embodiments, the major axis of the fitted
ellipse passes through the centroid of the detected organs.
[0061] The image detector 124 can determine a centroid-centroid
line connecting the centroid of the liver to the centroid of the
spleen at block 404. The image detector 124 can then identify a
boundary point along the centroid-centroid line that corresponds to
a valley (lowest) in count distribution. The image detector 124 can
find additional boundary points north (or above) of the
centroid-centroid line using the liver and spleen major axis as the
starting points and finding the minimum concentration in between.
The image detector 124 can use these boundary points to obtain a
demarcation line between the liver and the spleen. An example
demarcation line (or liver-spleen boundary line) 610 is illustrated
in FIG. 6. One side of the demarcation line can be the liver ROI
602 and the other side of demarcation line can be part of the
spleen ROI 606.
[0062] At block 406, the image detector 124 can detect a boundary
line 612 (see FIG. 6) between the liver and the bone marrow. In an
embodiment, the image detector 124 can attempt to calculate
centroid of the bone marrow and the liver-marrow boundary line can
be calculated as discussed above. But, in some embodiments, the
image detector 124 may not be able to identify a bone marrow
centroid. In these instances, the image detector 124 can start from
a point (below the centroid) on the major axis of the liver ellipse
and look for pixels in the direction of bone marrow where the value
of count is at a certain liver-marrow number. In one embodiment,
the liver-marrow number=threshold (as calculated above)+20%
(maximum intensity in liver-threshold). Other variations and
percentages can be also be used. Accordingly, the image detector
124 can identify boundary line 612 between liver and bone
marrow.
[0063] At block 408, the image detector 124 can connect the
liver-spleen boundary 610 with the liver marrow boundary 612 as
seen in FIG. 6. As discussed above with respect to liver, at block
410, the image detector 124 can identify a spleen-bone marrow
boundary line 614 using a spleen-marrow number. In one embodiment,
the spleen-marrow number=threshold+20% (maximum intensity in
spleen-threshold). At block 412, the image detector 124 can connect
the liver-spleen boundary 610 to the spleen-marrow boundary 614 as
illustrated in FIG. 6.
[0064] At block 414, the image detector 124 can detect outer
boundary 616 (see FIG. 6) of the marrow. In an embodiment, the
image detector 124 can connect the end point of the liver-marrow
boundary to the end point of the spleen-marrow boundary. The image
detector 124 can use the tissue threshold to connect the end
points. At block 416, the image detector 124 can find the outer
boundary 618 of the liver using the tissue threshold and body
contour as discussed above. The image detector 124 can connect the
outer boundary 618 to the end point of the liver-marrow boundary on
the bottom and the end point of the spleen-marrow boundary on the
top.
[0065] Similarly, at block 420, the image detector 124 can find the
outer spleen boundary 620. The image detector 124 can connect the
outer spleen boundary 620 to the end point of the liver-spleen
boundary on the top and the end point of the spleen-marrow boundary
on the bottom at block 422. Accordingly, the image detector can
generate liver ROI 602 and the spleen ROI 606 from the calculated
boundaries. The image detector can generate the marrow ROI 604 from
the liver-marrow, spleen-marrow, and the outer marrow
boundaries.
[0066] In some embodiments, the image detector 124 can validate the
calculated ROIs. For example, the image detector 124 can check
whether there is a large gap in liver-spleen boundary or if there
are too many voxels above tissue threshold in liver ROI connection
with liver/spleen boundary. Based on the validation, the user
interface generator 128 can generate a user interface that can
enable a clinician to modify the generated ROIs.
[0067] The ROIs calculated using the QLSSD system 120 can be robust
compared to hand drawn ROIs. For instance, there may be variations
in hand drawn ROIs between different clinician. Moreover, it may be
cumbersome and time-intensive to hand draw ROIs. Also, the
clinicians may not be able to detect counts appropriately from the
image as it may depend on the contrast levels of images and may
vary between scanners.
VI. Parameters
[0068] As discussed above, most of the existing health detection
techniques suffer from subjective analysis. There is a lack of an
objective analysis that is repeatable (within a small error range)
between clinicians and patients. The QLSSD system 120 can abstract
parameters from the scanned images to generate objective parameters
for evaluating hepatic function. The parameter calculator 126 can
use the processed images with detected ROIs to calculate one or
more of the following parameters. Other variations in calculation
of these parameters are possible as can be understood by a person
skilled in the art.
Liver, Spleen, and Bone Marrow Counts
[0069] In SPECT images, the counts correspond to detected radiation
from the compound. Accordingly, higher counts may indicate higher
concentration of the compound in a particular organ. The parameter
calculator 124 can use the detected ROIs for liver, spleen, and
bone marrow to calculate total counts in each of the respective
ROIs from one of the combined transaxial image. For combined
frames, the counts can represent volume as the stack of frames may
include multiple slices of the patient's body. In some embodiments,
a single mid-organ frame is used to compute liver and spleen
concentration. Concentration may also be computed by averaging
counts in a particular sub-region of the organ from a frame with
the highest counts. The image detector can find 3.times.3 voxel
areas in the highest count frame and average the counts to
determine a concentration (e.g. counts/minute/voxel). For three
dimensional ROIs, the counts may be summed across all frames for
respective organ ROIs. The counts can indicate to clinicians how
much of the compound is in the liver versus the other organs. As
discussed above, for a healthy liver, most of the counts might be
found in the liver ROI as compared to other organ ROIs. In some
embodiments, bone marrow counts are expressed as a ratio to the
number of frames to normalize for number of vertebral bodies
covered by the scan. In another embodiment, the number of vertebral
bodies is counted and that is used to normalize the bone marrow
counts.
Liver and Spleen Length
[0070] In an embodiment, the parameter calculator 128 can calculate
lengths of detected organs. For example, the parameter calculator
128 can detect length of the liver from an anterior planar image
for right lobe from mid-liver dome to the right inferior margin and
left love from right dome to inferior left lobe margin.
[0071] The parameter calculator 128 can measure the spleen length
as the greatest pole to pole length in posterior planar view. In
some embodiments, the spleen length is determined from the
transverse images. If there is a difference in spleen length by 10%
or more between different frames, the clinician may be warned and
manual intervention may be required via one of the user interfaces
described below.
Liver Spleen Index, Liver Bone Marrow Index (LBI), and Perfused
Hepatic Mass (PHM)
[0072] In some embodiments, the hepatic function may be understood
from distribution ratios using counts obtained from the one or more
scanned images. For instance, the liver-spleen index (LSI) can be
determined from comparing counts in the liver to the counts in
spleen. In an embodiment, the liver-spleen index (LSI) is a
function of liver to spleen ratio of total counts corrected for
spleen length. The parameter calculator 128 can calculate total
count ratio between liver and spleen as a total liver counts
divided by total liver plus spleen counts, or L/(L+S)t. In some
embodiments, the ratio is reproducible within 1%. The ratio might
be affected by spleen size independent of chronic liver disease. A
correction might be required for variation in spleen length between
patients. In one embodiment, the parameter calculator 128 can
correct for the spleen length. The parameter calculator 128 can
estimate the L/(L+S)t ratio expected from the impact of spleen
length in patients with normal livers (empirically derived formula
from patients with normal livers and varying size spleens). The
parameter calculator can then divide the measured L/(L+S)t ratio by
the estimated normal L/(L+S)t ratio and multiply it by 100 to
derive the liver spleen index (LSI).
[0073] The distribution of counts between liver and bone marrow may
be expressed as the liver-bone marrow index (LBI). In an
embodiment, the parameter calculator 128 can calculate LBI as the
log of liver count divided by bone marrow count per frame and
multiplied by 50 to produce a similar range to LSI.
[0074] In some embodiments, the parameter calculator 128 can
generate a parameter that is a function of both LSI and LBI. For
example, the parameter calculator 128 can calculate the perfused
hepatic mass (PHM) parameter by averaging of LBI and LSI, that is:
PHM=(LBI+LSI)/2.
Spleen and Liver Volume
[0075] The parameter calculator 128 can also calculate liver and
spleen volumes. Spleen and liver volumes may be calculated using
the total counts in an organ divided by a representative
concentration on the cross-sectional frame times the voxel volume.
In one embodiment, the parameter calculator 128 can use a single
mid-organ frame that is representative for the concentration. The
volumes in cc may be expressed as a ratio to the ideal body weight
(IBW) in pounds. One skilled in the art can recognize alternate
methods of obtaining representative concentrations of the organ of
interest, such as, sampling, histogram analysis, whole organ
averaging, or single organ slice. Additional example calculations
are discussed in "A Novel, Simple Method of Functional Spleen
Volume Calculation by Liver-Spleen Scan," by Hoefs et al, The
Journal of Nuclear Medicine, Vol. 40, No. 10 (October 1999),
incorporated herein by reference in its entirety. In some
embodiments, volumes do not rely on precise edge detection and are
insensitive to voxel size.
[0076] In some embodiments, the liver volume is automatically
calculated by performing a search through the scanned images to
identify a frame that contains the highest concentration areas of
the liver. The image detector can find 3.times.3 voxel areas in the
identified frame and average the counts to determine a
concentration (e.g. counts/minute/voxel). The parameter calculator
128 can use the highest average concentration value to calculate
the volume of the liver using the following formula and example
calculation:
Liver Volume=((Total Liver Counts/Highest Average Liver
Concentration)*Voxel Size*0.9562)-66.5.
Total counts liver=8600000 CPM Representative concentration=860
CPM/voxel Voxel volume (0.474 cm on a side)=0.474
cubed=0.474*0.474*0.474=0.10650 cc Raw
volume=(8600000/860).times.0.1065=1065 cc Corrected volume=raw
volume.times.correction factor=(1065.times.0.9562)-66.5=952 cc
[0077] The constants in the liver volume formula can be modified
based on calibration samples. For example, the calculated volume
can be compared to phantom volumes by linear regression analysis.
The spleen volume can be calculated using the method discussed
above with respect to liver. In an embodiment, each voxel
represents a 4.times.4.times.4 cubic millimeters. Accordingly, the
liver volume can be determined in cubic millimeters.
Normalized Liver and Spleen Volumes
[0078] Liver and spleen volumes may depend on the patient's overall
size. Thus, in some embodiments, ideal body weight is used to
normalize the organ volumes to provide clinically useful
parameters. The patient's actual body (e.g. obtained from PACS) may
also be used. The parameters can be calculated as follows:
Normalized Liver Volume=Liver Volume/Ideal Body Weight
Normalized Spleen Volume=Spleen Volume/Ideal Body Weight
The formula to calculate IBW may be different for males and
females.
Female IBW=100+{height (in inches)-60}.times.5
Male IBW=106+{height (in inches)-60}.times.6
Example: if the liver volume is 952 as calculated in the example
above, then A Female, 62 inches tall would have an IBW=110 lbs and
the Normalized Hepatic volume (corrected for body size)=8.7 cc/lb
IBW
Estimated Peritoneoscopic Score (estPS)
[0079] Peritoneoscopy can provide an indication of the degree of
"smoothness" or "granularity" or "nodularity" of the liver. The
QLSSD can calculate an Estimated Peritoneoscopic Score (estPS) by
combining several parameters. In one embodiment, the estPS is
calculated as follows:
estPS1=4.342-2.008RR-0.0206PHM+18.15/RL
where RL is the right lobe length in cm, and RR is the
redistribution ratio calculated as:
(RR)=[(Lp/Sp/2.5)+(Lp/BMp/17.5)]/2
Here Lp, Sp and BMp are pixel counts from the posterior planar view
for the liver, spleen and bone marrow respectively.
[0080] The Peritoneoscopic assessment of the liver can be a better
indicator than histologic fibrosis measurements since sampling
errors may be avoided. Thus, the estPS from QLSSD can provide a
good estimate of hepatic fibrosic stage as measured by histology
with almost no sampling error.
Staging
[0081] The parameter calculator 126 can classify severity of liver
disease by comparing one of the calculated parameters with expected
ranges. In one embodiment, patients are staged using the PHM
parameter as follows:
PHM.gtoreq.100; normal hepatic function (low risk)
95.ltoreq.PHM<100: mildly reduced hepatic function (intermediate
risk) PHM<95: reduced hepatic function (high risk)
[0082] In another embodiment, high risk patients are further
classified as moderately reduced hepatic function if the PHM>70
or severely reduced hepatic function if PHM<70. Other indicators
including colors may also be used for staging. these ranges may
vary in other embodiments, or may have more or fewer ranges (e.g. a
threshold may be used to determine if the patient's liver is
healthy or not).
Hepatic Activity Index
[0083] There may be a close correlation between LSI and LBI. A
linear regression equation drawn in a large group of patients can
define this relationship. In an embodiment, the LSI can be used in
this equation to determine an estimated LBI. The estimated LBI may
be subtracted from the Measured LBI and this difference can be
divided by the LSI to get the HAI. The HAI of less than a -0.10 can
indicate a significant departure from the usual relationship and
indicates a more rapidly progressive liver disease such as
alcoholic hepatitis.
[0084] In some embodiments, the formula for calculating HAI is:
If LSI>0.0
HAI=(LBIt-((LSI*0.665)+43.0))/LSI
Otherwise
[0085] HAI=0
VII. 3D Processing
[0086] In some of the embodiments discussed above, the frames
corresponding to slices across the patient's body can be analyzed
as combined frame (e.g. STI, MITI, etc.). In the alternative, the
frames can be processed individually to detect three dimensional
profile of organs. FIG. 7 illustrates a process for detecting ROIs
in three dimensions. The 3D ROI detection process 700 can be
implemented by the system described above. For illustrative
purposes, the process 700 will be described as being implemented by
components of the computing environment 100 of FIG. 1.
[0087] At block 702, the image detector 124 can detect 2D ROIs of
liver, spleen, and bone marrow as discussed above with respect to
FIGS. 3 and 4. The image detector 124 can continue detecting ROIs
for each frame. Some frames may have low signal content as the
organs begin to taper off. The image detector 124 can store the
ROIs from each of the frames for liver, spleen, and bone marrow. At
block 706, the image detector 124 can combine the stored liver ROIs
using morphological features to form a 3D liver volume. Similarly,
the image detector 124 can generate 3D spleen volume at block 708
and a 3D bone marrow volume at block 710. In some embodiment, the
image detector 124 can use edge following to connect ROIs between
frames. The parameter calculator 126 can calculate liver function
parameters and liver disease stages from the 3D ROIs using the
formulas as discussed above with respect to 2D frame.
[0088] In some embodiments, the image detector 124 can generate 3D
ROIs using directional searching and centroid in three dimension
instead of combining the results of 2D analysis for individual
frames. In three dimensions, the image detector 124 can rotate
direction vector of search before conducting iterative directional
searching and centroid analysis. In another embodiment, the image
detector 124 can use transform processes to map 3D volumes (such as
ellipsoids) into points in the transform space. For instance, the
liver, spleen, and marrow may be modeled as ellipsoids (or union of
ellipsoids) to use the transform techniques.
[0089] The 3D capability may also allow calculation of fibrosis
directly. In some embodiments, the functional ratios can overlap
because there might be overlap between liver, bone marrow and
spleen. Separating the frames before analyzing can reduce effects
of overlap.
[0090] The image detector 124 may also use data from CT or MRI
scan. The CT and MRI scan include information relating to outline
of organs. The image detector 124 can use the outline to map data
from SPECT scan on to a CT or MRI scan. Based on the mapping, the
image detector 124 can detect ROls from the SPECT scan.
VIII. Total Count Ratio (TCR)
[0091] The SPECT reconstruction may have a limited range of slices
that includes the entire organ being evaluated. Each slice may be
the width of the voxel (or pixel). A threshold can be designated as
to the surface voxel for inclusion of a surface voxel as a percent
of the maximal voxel concentration in the liver. As the liver
becomes more diseased, fewer of the surface voxels might have
greater than 50% of the maximal voxel concentration due to the
presence of fibrosis. Thus, the volume of the included voxels may
be smaller as the liver becomes more diseased (and total counts in
this 3-D ROI decreases). The included volume in patients with
chronic liver disease has fewer counts compared to the total counts
(TCR) compared to a similar procedure on a normal liver. The total
counts ratio can be calculated from summarized transaxial
image.
[0092] The image detector 124 can determine ROI around the organs
as discussed above. Based on the ROIs, the total counts (TC) for
each of the organs may be calculated by the parameter calculator
126. The parameter calculator 128 can select a threshold. In an
embodiment, the threshold is 50% of the maximal voxel
concentration). The parameter calculator 126 can apply the
threshold to each slice within the above ROI, therefore picking the
surface voxel to be used on each slice. The image detector 124 can
take the surface voxels selected from each slice to draw a 3-D
image for the whole organ. The parameter calculator 126 can
calculate the counts within the generated 3-D image. These counts
can represent the threshold counts. The parameter calculator 126
can calculate the total count ratio (TCR), where TCR=threshold
count/total count (TC) for each organ. Accordingly, TCR can be
calculate for the organs and included in reports.
IX. Predicting Post-Op Surgery
[0093] 3D imaging can enable pre-surgery estimates of the loss of
hepatic function after surgery for hepatocellular carcinoma (HCC)
and other hepatic masses. The expected anatomic loss from surgery
can be overlaid with the 3D ROIs and the loss of function
calculated. These factors may be used in the output impressions to
stage the liver disease, estimate the risk of complications and for
prognosis. In patients with hepatic cancer and limited hepatic
reserve, the liver health parameters may be used to estimate the
loss of function at surgery to determine surgical risk.
[0094] FIG. 8 illustrates an embodiment of a process 800 for
estimating post-resection parameters. The post-resection parameter
calculation process 800 can be implemented by any of the systems
described above. For illustrative purposes, the process 800 will be
described as being implemented by components of the computing
environment 100 of FIG. 1. At block 802, the image detector 124 can
determine 3D ROIs of the liver, spleen, and bone marrow as
discussed above. At block 804, the image retriever module 122 can
access CT or MRI images including the liver. The image detector 124
can superimpose the CT or MRI image on the 3D ROIs using, for
example, image registration techniques.
[0095] In some embodiments, superimposition may be performed using
built-in capabilities of hybrid SPECT/CT scanners. At block 806,
the user interface module 128 can generate a user interface that
can allow clinicians to select or draw resection volume on the
superimposed image. In an embodiment, the resection volume is
automatically drawn based on importing parameters from surgical
planning system. In some embodiments, the QLSSD system 120 can
display a suggested resection volume. The suggested resection
volume may be based on the differences between the SPECT liver
volume and the CT or MRI liver volume. As an example, portions of
the liver that contain hepatocellular carcinoma (HCC) will show up
in the CT or MRI images but not in the SPECT images. At block 808,
the parameter calculator can ignore the area that is part of the
planned resection volume to calculate post-resection parameters.
Accordingly, the process 800 can enable clinicians to determine
liver health post operation and determine whether more or less of
the liver should be removed. Also, the clinicians can assess risk
of surgery by reviewing the post-resection parameters.
X. User Interfaces
[0096] The example graphical user interfaces shown in FIGS. 9 to 18
may be generated by the QLSSD system 120, the QLSSD system plugins,
or a combination of both. For illustration purposes, these user
interfaces are shown primarily in application interfaces, although
it should be understood that these user interfaces can be generated
with web browsers (including mobile apps) other than application
interfaces. Further, example user interface controls (active links)
are shown, including buttons, status bars, hyperlinks or links, and
the like. Any of the user interface controls shown can be replaced
with other user interface controls, including but not limited to
radio buttons, check boxes, text boxes, select boxes or drop-down
boxes, combinations of the same, and the like.
[0097] FIG. 9 illustrates an embodiment of a user interface 900
generated by the user interface module 128 that can enable
clinicians to access functionality of the QLSSD system 120. For
example, clinicians can import scanned images for a patient from
PACS and view the scanned images in the user interface. The user
interface also include active links to change contrast and
brightness of the scanned image. Clinicians can further run
automated analysis on the scanned images by selecting an active
link. The user interface 900 can enable clinicians to navigate a
DICOM hierarchy for the selected image (patient, study, series, and
image). The navigation button and list can facilitate moving from
one image to another, to refresh patient list, etc. Moreover,
images can be imported from PACS such as SPECT transverse images
and/or static posterior planar images. In some embodiments, image
are imported automatically from PACS. The clinician can select an
image or a set of images for analysis by the QLSSD system as
discussed above.
[0098] FIG. 10 illustrates an embodiment of a user interface 1000
generated by the user interface module 128 that can enable
communications with the PACS system 102. For example, the clinician
can run queries to access data corresponding to a particular
patient. Based on the selected query, the QLSSD system 120 can
communicate with PACS and retrieve the requested data.
[0099] FIG. 11 illustrates an embodiment of a user interface 1100
that can be generated by the user interface module 128 in response
to the detected regions of interest corresponding to organs. The
illustrated embodiment of the user interface 1100 shows liver,
spleen, and bone marrow ROIs. The user interface 1100 can include
an active link 1102 for selecting whether the patient has his or
her spleen removed. For patients with no spleen, the ROI detection
may need to be repeated by selecting active link 1110 or manually
determined from clinicians using active links 1104, 1006, and 1108.
The user interface 1100 also includes link 1112 to select spleen
length view as shown in FIG. 12 and link 1114 to select frame range
view as shown in FIG. 13. The spleen length user interface 1200 can
enable clinicians to visually confirm the length of the spleen as
calculated by the QLSSD system 120. The spleen length user
interface 1200 can also enables functionality for the clinician to
override calculated spleen length. In the illustrated embodiment,
the spleen length is 9.2 cm and scanned image is showing a
posterior view of the liver and the spleen. FIG. 13 illustrates an
embodiment of a user interface 1300 that can enable clinicians to
select the frame range. In some instances, the image detection
module 124 can identify a larger portion of the bone marrow and may
include pelvis region. The clinicians 1300 can select the frame
range to guide the image detector module 124 to more accurately
calculate the ROIs. The user interface 1300 can enable clinicians
to visually confirm that the range of frames detected by QLSSD
system 120 encompasses the entire liver and spleen.
[0100] FIG. 14 illustrates an embodiment of a user interface 1400
that includes a report generated by the QLSSD system 120. The user
interface module 128 can generate a report in response to receiving
a command from a clinician. The illustrated report includes
calculation of a PHM parameter for the patient. The report can also
include additional details from the patient, such as sex, height,
weight, date of the study. In some embodiments, the report can
include a trend graph 1402 plotting a parameter over time. In the
illustrated embodiment, the trend graph includes PHM over time.
[0101] FIG. 15 illustrates an embodiment of a user interface 1500
that enables clinicians to write their impressions for a patient
based on the calculated parameter. In some embodiments, the user
interface 1500 includes automatically generated impressions that
can be used by clinicians. As discussed above, the impressions may
be generated by the QLSSD system from lookup tables based on the
calculated parameters. The clinician has the option of using the
suggested impression from the QLSSD system 120.
[0102] FIG. 16 illustrates an embodiment of a user interface 1600
that enables clinicians to communicate with PACS. In the
illustrated embodiment, the clinicians can use the interface 1600
to send reports for storage in PACS. The reports can be
automatically associated with the patient and the doctor.
[0103] FIG. 17 illustrates an embodiment of a reporting user
interface 1700 that includes a generated report for a patient. The
report includes calculated parameters, for example, PHM and a trend
in patient's PHM over the years. A clinician can review the trend
and identify whether the patient is improving or getting worse. In
addition, the clinicians can identify effects of a particular
treatment. In some embodiments, the trend can include future
prediction based on resection parameters as discussed above. The
report can also include impression selected by the clinician or
automatically generated by the QLSSD system 120. In some
embodiments, the report can illustrate SPECT scans overlaid with
detected ROIs.
[0104] In one embodiment, the report includes on the left side
images from which the raw data were derived: a base image
(anterior/posterior), summarized transaxial images, single
transaxial slice and distributions of total and planar counts. The
sections on the right side have panels for demographics, LSI, LBI
and PHM; a panel for volumes by 4 methods (the circled is the one
we use); a panel for total counts; 2 panels for posterior planar
counts; 2 panels for representative concentrations from the single
slice--one for liver and one for spleen; and liver and spleen
lengths.
[0105] FIG. 18 illustrates another embodiment of a reporting user
interface 1800 that includes a generated report for a patient.
Compared to the report illustrated in FIG. 17, the user interface
1800 includes additional parameters for clinicians.
XI. Terminology
[0106] A number of computing systems have been described throughout
this disclosure. The descriptions of these systems are not intended
to limit the teachings or applicability of this disclosure. For
example, the clinician systems and described herein can generally
include any computing device(s), such as desktops, laptops, video
game platforms, television set-top boxes, televisions (e.g.,
internet TVs), computerized appliances, and wireless mobile devices
(e.g. smart phones, PDAs, tablets, or the like), to name a few.
Further, it is possible for the clinician systems described herein
to be different types of devices, to include different
applications, or to otherwise be configured differently. In
addition, the user systems described herein can include any type of
operating system ("OS"). For example, the mobile computing systems
described herein can implement an Android.TM. OS, a Windows.RTM.
OS, a Mac.RTM. OS, a Linux or Unix-based OS, or the like.
[0107] Further, the processing of the various components of the
illustrated systems can be distributed across multiple machines,
networks, and other computing resources. In addition, two or more
components of a system can be combined into fewer components. For
example, the various systems illustrated can be distributed across
multiple computing systems, or combined into a single computing
system. Further, various components of the illustrated systems can
be implemented in one or more virtual machines, rather than in
dedicated computer hardware systems. Likewise, the data
repositories shown can represent physical and/or logical data
storage, including, for example, storage area networks or other
distributed storage systems. Moreover, in some embodiments the
connections between the components shown represent possible paths
of data flow, rather than actual connections between hardware.
While some examples of possible connections are shown, any of the
subset of the components shown can communicate with any other
subset of components in various implementations.
[0108] Depending on the embodiment, certain acts, events, or
functions of any of the algorithms, methods, or processes described
herein can be performed in a different sequence, can be added,
merged, or left out altogether (e.g., not all described acts or
events are necessary for the practice of the algorithms). Moreover,
in certain embodiments, acts or events can be performed
concurrently, e.g., through multi-threaded processing, interrupt
processing, or multiple processors or processor cores or on other
parallel architectures, rather than sequentially.
[0109] Each of the various illustrated systems may be implemented
as a computing system that is programmed or configured to perform
the various functions described herein. The computing system may
include multiple distinct computers or computing devices (e.g.,
physical servers, workstations, storage arrays, etc.) that
communicate and interoperate over a network to perform the
described functions. Each such computing device typically includes
a processor (or multiple processors) that executes program
instructions or modules stored in a memory or other non-transitory
computer-readable storage medium. The various functions disclosed
herein may be embodied in such program instructions, although some
or all of the disclosed functions may alternatively be implemented
in application-specific circuitry (e.g., ASICs or FPGAs) of the
computer system. Where the computing system includes multiple
computing devices, these devices may, but need not, be co-located.
The results of the disclosed methods and tasks may be persistently
stored by transforming physical storage devices, such as solid
state memory chips and/or magnetic disks, into a different state.
Each process described may be implemented by one or more computing
devices, such as one or more physical servers programmed with
associated server code.
[0110] Conditional language used herein, such as, among others,
"can," "might," "may," "e.g.," and the like, unless specifically
stated otherwise, or otherwise understood within the context as
used, is generally intended to convey that certain embodiments
include, while other embodiments do not include, certain features,
elements and/or states. Thus, such conditional language is not
generally intended to imply that features, elements and/or states
are in any way required for one or more embodiments or that one or
more embodiments necessarily include logic for deciding, with or
without author input or prompting, whether these features, elements
and/or states are included or are to be performed in any particular
embodiment. The terms "comprising," "including," "having," and the
like are synonymous and are used inclusively, in an open-ended
fashion, and do not exclude additional elements, features, acts,
operations, and so forth. Also, the term "or" is used in its
inclusive sense (and not in its exclusive sense) so that when used,
for example, to connect a list of elements, the term "or" means
one, some, or all of the elements in the list. In addition, the
articles "a" and "an" are to be construed to mean "one or more" or
"at least one" unless specified otherwise.
[0111] Conjunctive language such as the phrase "at least one of X,
Y and Z," unless specifically stated otherwise, is otherwise
understood with the context as used in general to convey that an
item, term, etc. may be either X, Y or Z. Thus, such conjunctive
language is not generally intended to imply that certain
embodiments require at least one of X, at least one of Y and at
least one of Z to each be present.
[0112] While the above detailed description has shown, described,
and pointed out novel features as applied to various embodiments,
it will be understood that various omissions, substitutions, and
changes in the form and details of the devices or algorithms
illustrated can be made without departing from the spirit of the
disclosure. Thus, nothing in the foregoing description is intended
to imply that any particular feature, characteristic, step, module,
or block is necessary or indispensable. As will be recognized, the
processes described herein can be embodied within a form that does
not provide all of the features and benefits set forth herein, as
some features can be used or practiced separately from others. The
scope of protection is defined by the appended claims rather than
by the foregoing description.
* * * * *