Device to monitor side effects of treatment

Abstract

Systems for monitoring organ health during treatment, such as cancer, utilize physiological imaging of the type used for treatment monitoring and organ-specific processing to provide a comprehensive assessment of treatment side effects.

Description

[Technical Field]

[0001] The present invention relates to a device for monitoring side effects of medical treatment.

[0002] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0003] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Patent Application No. 63 / 021,936, filed May 8, 2020, and U.S. Nonprovisional Patent Application No. 17 / 236,715, filed April 21, 2021, which are incorporated herein by reference.

[0004] The present invention relates to medical imaging equipment that performs physiological imaging (e.g., PET imaging, functional MRI), and more particularly to an apparatus that uses physiological imaging to automatically assess toxic side effects during treatment of cancer or the like. [Background technology]

[0005] Treatments (e.g., those using immunotherapy) can have serious side effects that require the treatment to be discontinued. In one example, cancer treatment using immune checkpoint inhibitors (ICIs) can cause patients to develop immune-mediated colitis, whose symptoms include diarrhea. Grade 2 or 3 colitis requires a delay in ICI treatment, but grade 4 colitis can be life-threatening and requires permanent discontinuation of treatment.

[0006] The impact of such side effects on treatment can be reduced by careful monitoring of side effects during treatment. Unfortunately, the types of possible side effects can be wide-ranging, making it difficult to monitor for many common side effects. For example, a colonoscopy or colon biopsy may be required to diagnose immune-mediated colitis, which is impractical for routine repeated surveillance. Summary of the Invention [Problem to be solved by the invention]

[0007] The present invention has been made to solve the problems in the prior art described above. [Means for solving the problem]

[0008] The inventors recognize that physiological imaging (often performed to monitor tumor regression during cancer treatment) can be utilized in conjunction with organ-specific toxicity rules to simultaneously provide insight into the health of unrelated organs.

[0009] In one embodiment, gut health is measured by: 18 The organ may be monitored using F-FDG uptake agents, and the same molecular imaging uptake agents may be used to track tumors. The present invention uses data from a structural imaging scanner (e.g., a CT machine) that provides anatomical information to segment the organ, and then performs a focused analysis of molecular uptake data (e.g., from a PET machine) or organ function data (e.g., from an MRI machine) for that organ, and applies toxicity rules to assess organ health. Multiple organs can be analyzed simultaneously to provide a comprehensive overview of organ health.

[0010] As used herein, physiological imaging refers to imaging techniques (e.g., PET imaging, functional MRI, or MRS) that detect either molecular (e.g., metabolic) changes, functional (e.g., blood flow) changes, or chemical (e.g., local chemical composition and chemical absorption) changes, whereas structural imaging (standard CT imaging and MRI imaging) provides anatomical images.

[0011] In one embodiment, the present invention provides an apparatus for assessing organ health during patient treatment, the apparatus having an electronic computer executing a stored program to (a) receive a structural image of at least one patient organ, (b) receive a physiological image of the at least one organ indicative of organ function, (c) process the structural image to create a mask that describes the organ, (d) use the mask to select a portion of the physiological image that is relevant to the organ, (e) apply toxicity rules specific to the organ to the physiological image of the portion to provide an organ health assessment, and / or (f) output an indication of organ health based on the assessment.

[0012] Thus, one feature of at least one embodiment of the present invention is to provide an automated system that can leverage often already existing physiological imaging information to characterize side effects that may impact treatment.

[0013] The stored program may further process a portion of the physiological image to identify lesions and refine the organ mask to remove lesions before applying the toxicity rules.

[0014] Thus, a feature of at least one embodiment of the present invention is that it allows for organ health monitoring even when the organ contains lesions that are being treated. By segmenting and isolating these lesions, sensitivity to organ health is improved.

[0015] By processing the structural image to create a mask, the mask may be associated with an organ type, and identifying lesions within the organ may follow lesion identification rules associated with the organ type.

[0016] Thus, a feature of at least one embodiment of the present invention is to provide advanced quantitative analysis of uptake information with respect to organ health, which would be difficult to achieve by observing physiological images alone due to the variability in the correlation between physiological images and organ health.

[0017] The electronic computer may execute a stored program to further process the structural image to identify a lymph node region mask, which is used to select portions of the physiological scan associated with lymph nodes. Lymph node activity rules may then be applied to the portions of the physiological image associated with lymph nodes to assess lymph node activity, and the output may provide an indication of lymph node activity indicative of activation and / or stress of the lymphatic system.

[0018] Thus, one feature of at least one embodiment of the present invention is to provide an early measure of organ health by assessing whether lymph nodes are activated or overburdened.

[0019] More generally, the stored program may provide a set of toxicity rules associated with different organs, and the different toxicity rules may be applied to different organs to provide organ health assessments for a plurality of different organs, where the output displays organ health for a plurality of organs (e.g., first and second organs, first, second and third organs, first, second, third and fourth organs, etc.).

[0020] Thus, a feature of at least one embodiment of the present invention is to provide medical personnel with a comprehensive overview of organ health that would otherwise require numerous tests. By automatically segmenting organs and using organ-specific rules, simultaneous organ health assessment of multiple organs becomes practical.

[0021] The output may provide a composite measure of organ health for multiple organs.

[0022] Thus, a feature of at least one embodiment of the present invention is to provide a healthcare provider with an immediate indication of whether organ health is significantly compromised.

[0023] Alternatively or additionally, the output may provide an image based on the structural image and augmented with organ health data.

[0024] Thus, one feature of at least one embodiment of the present invention is to utilize the image data used for segmentation to provide a framework for communicating the analysis results of the system to healthcare providers.

[0025] The stored program may retain previous outputs indicating organ health associated with previous structural images and previous physiological images to provide an indication of organ health trends for multiple organs.

[0026] Thus, one feature of at least one embodiment of the present invention is to provide important trend information that allows for early intervention in toxicological conditions, information that would be difficult or impossible to assess by simply observing individual uptake scans.

[0027] The toxicity model may include thresholds or other characteristics extracted from the physiological images to predict organ health states for multiple different organs, and the output may display the predicted organ health states for the multiple organs.

[0028] Thus, a feature of at least one embodiment of the present invention is to incorporate an empirical understanding of current organ health to estimate likely prognosis.

[0029] The electronic computer may further analyze the physiological scans to assess changes in the cancerous lesions, and the output may provide an indication of changes in the cancerous lesions.

[0030] Thus, a feature of at least one embodiment of the present invention is to provide an integrated display of treatment effectiveness and side effects to improve guidance during treatment.

[0031] These particular objects and advantages may apply to only some embodiments that fall within the scope of the claims and, therefore, do not define the scope of the invention. [Brief explanation of the drawings]

[0032] [Figure 1] FIG. 1 is a block diagram of an exemplary embodiment of the present invention showing a CT scanner and a PET scanner in communication with a central processor having a display terminal and running stored programs that provide organ health information. [Figure 2] FIG. 1 is a flow block diagram showing how a central processor processes data from the CT and PET scanners to provide an organ health assessment. [Figure 3] 3 is a flowchart of a program executed by the central processor in relation to the flow block diagram of FIG. 2. [Figure 4] 4 is an example of a display generated on the display terminal of FIG. 1 by the processing of FIGS. 2 and 3. DETAILED DESCRIPTION OF THE INVENTION

[0033] Referring now to FIG. 1, a system 10 for assessing the health of an organ during disease treatment may include a structural imaging scanner 12, such as a kilovoltage CT (computed tomography) scanner or megavoltage CT scanner, an MRI (magnetic resonance imaging) scanner, or the like, capable of providing high-resolution structural image scans 14 presenting anatomical information of a patient 15.

[0034] Additionally, system 10 may include a physiological imaging scanner 16 that is capable of scanning patient 15 after administration of, for example, a molecular imaging uptake agent 17 to measure the amount of uptake of uptake agent 17 .

[0035] In one embodiment, the physiological imaging scanner 16 may be a PET (positron emission tomography) scanner. As commonly understood in the art, PET is a nuclear medicine imaging technique that produces physiological scans 18 that reveal molecular processes within the body of a patient 15, as manifested by the preferential migration of an uptake agent 17 to tumor tissue. An example of a molecular imaging uptake agent 17 is a computed tomography (CT) scanner. 18 F-FDG PET / CT) is incorporated into a physiological imaging scanner 16. This PET scanner is only one example of a physiological imaging scanner 16, and the present invention contemplates that physiological imaging scanners 16 including functional CT machines or functional MRI machines or other similar devices that measure basal tissue metabolism may be used.

[0036] Generally, the physiological scan 18 produced by the physiological imaging scanner 16 has a lower spatial resolution than the high-resolution structural image scan 14 from the structural imaging scanner 12. In each case, the structural image scan 14 and the uptake agent physiological scan 18 represent multi-dimensional information associated with volume elements (voxels) distributed in three dimensions within a volumetric region of interest within the patient 15.

[0037] In the present invention, the patient 15 may be scanned simultaneously in both the structural imaging scanner 12 and the physiological imaging scanner 16, which may in some cases be the same machine using different hardware or protocols. These scans may be repeated at various times throughout the treatment of the patient 15, for example, between sessions of treatment of the patient 15 with chemotherapy, radiation therapy, etc.

[0038] Continuing with reference to FIG. 1 , the structural imaging scans 14 and the uptake agent physiological scans 18 are received by an electronic computer 22 for processing, as described in more detail below. Generally, the electronic computer 22 includes one or more processing units 24 in communication with a memory 26, which retains data and stored programs 28 for implementing portions of the present invention. The computer 22 may communicate with a graphics display 30 that displays color output images based on the structural imaging scans 14 and the uptake agent physiological scans 18, and may further communicate with user input devices 32, such as a keyboard, mouse, or the like, each of which allows a user to enter data. As described in more detail below, the present invention provides output on the display 30 indicative of organ health and lesion progression or regression based on measurements of uptake of the uptake agent at multiple tumor sites and multiple organ sites within the patient 15.

[0039] 2 and 3, in a first step indicated by process block 34, stored program 28 receives the scan data acquired from structural imaging scanner 12 and physiological imaging scanner 16 as described above.

[0040] 3 , the structural image scan 14 from the structural imaging scanner 12 is transferred to a set of segmenters 38a-38b, which operate to generate masks 40a-40c that isolate portions of the structural image scan 14 associated with particular organs. In this regard, each segmenter 38 may be tailored to a particular organ (e.g., heart, lungs, brain, liver, kidneys, bone, spleen, stomach, pancreas, pharynx, larynx, blood vessels, muscle, gallbladder, intestines, lymph nodes, bone marrow, bladder, etc.) such that the mask 40 defines the specific organ volume of each organ for a particular patient and is readily identifiable as the respective organ. While only three segmenters 38 are shown for clarity, more or fewer segmenters 38 may be used as needed or desired to cover all organs or organ regions of interest.

[0041] In one embodiment, segmenter 38 may be implemented as a trained convolutional neural network (CNN), for example, using a training set of structural image scans 14 acquired from various patients that have been preprocessed (e.g., normalized) and segmented to identify the organs of those patients. Using the training set, the neural network is trained to generate sets of weights 42a-42c that are specific to various organ types and capable of generating segmentations of those organs, as commonly understood in the art. In one embodiment, the neural network architecture may be that of DeepMedic, described in Kamnitsas K., Ledig C., Newcombe VF, et al., "Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation," Med Image Anal., 2017, Vol. 36, pp. 61-78, which is incorporated herein by reference. The present invention contemplates that other segmentation systems that provide automatic segmentation of organ volumes may also be used. Furthermore, although multiple independent segmenters 38 are shown, it should be understood that common hardware may be used, and the process may be performed sequentially by switching individual weights 42.

[0042] The mask 40 information for each organ is then received by a corresponding lesion identifier 44a-44c, which in turn receives the physiological scan 18 and / or structural scan 14 (shown by the dashed lines). The lesion identifiers 44a-44c are also organ-specific, having lesion identification rules 45a-45c associated with each organ that describe, for example, the uptake characteristics associated with lesions in the respective organ, as well as the size and shape characteristics of the lesions, to aid in automatically segmenting and identifying lesions within the received mask 40, in accordance with processing block 48 of Figure 3. An automated lesion identification approach can look for localized areas of excessive uptake, as described, for example, in U.S. Patent Application Publication No. 2016 / 0100795, which is assigned to the assignee of the present invention and incorporated by reference.

[0043] The output of lesion classifier 44 is a set of lesion volumes 46a-46c that can be subtracted (on a spatial basis) from mask 40 to produce refined masks 40'a-40'c that exclude the lesion volumes.

[0044] The physiological scan 18 and refined mask 40' are then passed to toxicity evaluators 50a-50c, which are associated with organ-specific rules 52a-52c for the individual organs in the mask 40'. Generally, the toxicity evaluator 50 analyzes the data from the physiological scan 18 limited to the scope of the augmented mask 40' to enhance sensitivity and specificity. In one embodiment, the organ-specific rules 52 extract standardized uptake value (SUV) histograms from the physiological scan 18 within the mask 40' and store these values ​​in a historical record 54, which records these values ​​for subsequent sessions of scanning for a particular patient 15. Trends in these historical SUV values ​​from the record 54 may then be related to organ health. For example, the inventors have determined that for the intestinal organ, the uptake of an 18We have observed that the increase in SUV95, defined as the 95th percentile of the bowel SUV histogram, from a patient's pretreatment baseline scan on F-FDG PET / CT is significantly greater in patients who subsequently develop colitis compared to patients who do not. In this case, an organ-specific rule 52 for the bowel may state that if SUV95 increases by more than 40% from baseline, the patient is in a pre-colitis state (i.e., a state prone to colitis) and treatment adjustment or other measures should be considered. Our research has shown that the above-mentioned signs of colitis, which occur an average of 115 days before the clinical diagnosis of colitis, are a useful predictor of deteriorating organ health. This measure has a sensitivity of 75% and a specificity of 88%. Similar empirical studies may be used to develop and validate additional rules 52 for other organs.

[0045] Each toxicity assessor 50 using the organ-specific rules 52 may then output an organ health value 56 to the combiner 58. An example of the organ health value 56 may output an associated change in a value, such as the SUV 95 mentioned above, along with an interpretation value, such as the threshold (40%) used to define the pre-colitis state. These values ​​are then used by the combiner 58, as described in more detail below. More generally, the combiner 58 combines information from each of the toxicity calculators 50 and the organ-specific rules 52. The combiner 58 also receives raw CT structural image scans 14, as described below.

[0046] 2 and 3, the structural image scan 14 may also be provided to a lymph node neighborhood segmenter 60, which operates similarly to segmenter 38 but identifies the extent of lymph node clusters within the patient. Segmenter 60 may also, in one embodiment, utilize a convolutional neural net trained as described above using a set of manually segmented structural scans to identify lymph node neighborhoods (e.g., around the base of the neck) to generate a set of weights 62 for the neural net. Segmenter 60, operating in conjunction with weights 62, then generates a neighborhood mask 64 that defines the extent of the lymph node cluster, in accordance with processing block 65 of FIG. 3.

[0047] This neighborhood mask 64 information is sent to the lymph node evaluator 66, which, in accordance with processing block 67 of FIG. 3, uses the physiological scan 18 to identify lymph nodes (e.g., by thresholding or as described in the patient above) and determine whether they are enlarged, indicating possible stress on the lymph node system (e.g., due to detectable or undetectable organ toxicity from treatments, including immunotherapy).

[0048] 3, the lymph node evaluator 66 may generate an indication 68 of lymph node enlargement, which is a proxy for lymph node stress, that is also output to the synthesizer 58. This indication 68 may be, for example, the percentage increase in lymph node size since the baseline scan, as indicated by uptake of an uptake agent.

[0049] Additionally, the structural imaging scans 14 and physiological scans 18 may also be used in a lesion tracking circuit 73, which may provide time-series monitoring of disease state and thus provide an indication of the effectiveness of primary therapy. This tracking may implement the methods described in the above-referenced U.S. Patent Application Publication No. 2016 / 0100795, and outputs treatment information to the combiner 58.

[0050] 2, 3, and 4, the synthesizer 58, in accordance with processing block 71, can provide an output display 79 presenting a comprehensive overview of organ health, which would be difficult or impossible if it were to rely individually on analyzing various organs according to different analytical criteria dictated by the organ-specific rules 52 and aggregated from various empirical studies. In one embodiment, the display 70 can present scan images 72, derived, for example, from a CT structural image scan 14, depicting each image 78 being analyzed, highlighted using masks 40′ and 64 created by previously removing areas of pathology 75. Each of these organ images can include text annotations 74 and / or diagrams identifying the organ and measurements 79 obtained, and can present a chart 76 showing trends in the organ health measurements 79 and thresholds 77 of the health values ​​56 according to the organ-specific rules 52, the thresholds 77 indicating empirically derived thresholds at which side effects are likely to prove treatment-limiting. Thus, for example, in the case of the intestine, the change in SUV 95 value may give a measurement 79 and the 40% threshold may be displayed as the threshold 77 .

[0051] Each of these various health state values ​​56 may be combined and weighted or normalized to provide a single overview 80 of the patient's organ health state, intended to alert the physician to a possible deterioration in the condition, which would prompt the physician to carefully review the underlying data.

[0052] In addition to monitoring the patient's organ health, the present invention can use the data from the lesion tracking circuitry 73, as described above with respect to the aforementioned U.S. Patent Application Publication No. 2016 / 0100795, to provide an assessment of disease progression by combining chart 82 with disease location scan 84. In this case, the present invention provides a more complete view of treatment and its effectiveness and side effects.

[0053] It will be appreciated that the present invention may provide a user interface that provides the user with the ability to select and modify system parameters explicitly described above or implicitly required, which may include parameters required to select one or more organs, as well as parameters that add filters that control the amount and type of information provided.

[0054] The functionality of the present invention may be accessed or used remotely; data collected by an imaging system may be transmitted over a (e.g., secure, HIPAA-compliant) communications network to a remote computer system that analyzes the data, and the results of that analysis may be returned (or transmitted elsewhere) for use.

[0055] In this regard, the present invention may include a database component that collects and stores data from one or more patients, and this information may be used to modify the effective thresholds used in the present invention and / or to provide predictions or recommendations regarding historical outcomes that may apply to a particular query subject.

[0056] As used herein, the term organ generally refers to any group of tissues adapted to perform a specific function, including but not limited to the liver, lungs, lymph nodes, intestines, etc.

[0057] Certain terminology is used herein for reference purposes only and is not intended to be limiting. For example, terms such as "upper," "lower," "above," and "below" refer to directions in the drawings to which reference is made. Terms such as "front," "back," "rear," "bottom," and "side" indicate the orientation of portions of a component within a consistent but arbitrary frame of reference that becomes clear with reference to the text and associated drawings that describe the component under discussion. Such terminology may include the words specifically mentioned above, their derivatives, and synonyms. Similarly, the terms "first," "second," and other such numerical terms referring to structures do not imply a particular sequence or order unless clearly dictated by context.

[0058] When introducing elements or features of the present disclosure and exemplary embodiments, the articles "a," "an," "the," and "said" are intended to mean that there are one or more of such elements or features. The words "comprising," "including," and "having" are intended to be inclusive, meaning that there may be additional elements or features other than the specifically stated element or feature. It should further be understood that method steps, processes, and operations described herein should not be construed as requiring them to be performed in the particular order described or illustrated, unless specifically identified as such. It should also be understood that additional or alternative steps may be employed.

[0059] References to "a microprocessor" and "a processor," or "the microprocessor" and "the processor," may be understood to encompass one or more microprocessors capable of communicating in a standalone and / or distributed environment and thus configurable to communicate with other processors via wired or wireless communication, and such one or more processors may be configured to operate in one or more processor-controlled devices, which may be similar or different devices. Furthermore, references to memory, unless otherwise specified, may encompass one or more memory elements and / or components readable and accessible by the processor, which may be internal to the processor-controlled device, external to the processor-controlled device, or accessed via a wired or wireless network.

[0060] It is expressly intended that the present invention is not limited to the embodiments and examples contained herein, and that the claims be understood to encompass modifications of those embodiments, including portions of the embodiments and combinations of elements of separate embodiments, as falling within the scope of the following claims. All publications mentioned herein, including patent and non-patent publications, are hereby incorporated by reference in their entirety.

[0061] In order to assist the Patent Office, and all readers of any patent that may issue on this application, in interpreting the claims appended to this application, applicants wish to note that, unless the phrase "means for" or "step for" is expressly used in a particular claim, no appended claim or claim element is intended to invoke 35 U.S.C. 112(f).

Claims

1. 1. A device for assessing organ health during patient care, comprising: an electronic computer, the electronic computer executing a stored program, (a) receiving a structural image of at least one patient organ; (b) receiving a physiological image of the at least one organ indicative of organ function; and (c) processing the structural image according to the organ function shown in the physiological image to create a mask that fits the at least one organ; (d) using the mask to select a portion of the physiological image associated with the at least one organ; (e) applying toxicity rules specific to said portion to provide an organ health assessment; (f) outputting an indication of organ health status of the at least one organ based on the assessment; and To do Device.

2. The electronic computer executes the stored program and further processing the portion of the physiological image associated with the at least one organ to identify lesions in the at least one organ and refining the mask to remove the lesions prior to application of the toxicity rules; 10. The apparatus of claim 1.

3. 3. The apparatus of claim 2, wherein the mask for an organ type is identified by processing the structural image to create a mask, and the identifying of lesions in the at least one organ follows a lesion identification rule associated with the organ type.

4. the electronic computer executing the stored program further processes the structural image to identify a lymph node region mask; the lymph node region mask is used to select portions of the physiological image that are associated with lymph nodes; applying lymph node activity rules to the portion of the physiological image associated with a lymph node to assess activity of the lymph node; The output also provides an indication of lymph node activity indicative of activation of the lymph node system.

10. The apparatus of claim 1.

5. the stored program provides a set of toxicity rules associated with various organs; said processing said structural image to generate a set of masks for different organs, each mask associated with a particular organ type; the different masks are used to select different portions of the physiological image that relate to the different organs, the different portions being associated with specific attribute types depending on the mask used; applying the different toxicity rules to different parts associated with the toxicity rules and the organs according to the organs to provide an organ health assessment for a plurality of different organs; the output being indicative of organ health status for a plurality of organs.

10. The apparatus of claim 1.

6. The device of claim 1 , wherein the output provides a composite measure of organ health for multiple organs.

7. The apparatus of claim 1 , wherein the output provides an image based on the structural image and augmented with organ health data.

8. 10. The apparatus of claim 1, wherein the stored program maintains previous outputs indicating organ health associated with previous structural images and previous physiological images to provide a display of organ health trends for multiple organs.

9. the toxicity rules provide thresholds for predicting organ health status for different organs; the output indicating predicted organ health status for a plurality of organs.

10. The apparatus of claim 1.

10. The electronic computer further analyzes the physiological images to assess disease progression; The output provides an indication of disease (cancer lesion) changes.

10. The apparatus of claim 1.

11. The apparatus of claim 1 , wherein the physiological image is a PET scan or a functional CT scan or a functional MRI / MRS scan.

12. The apparatus of claim 1 , wherein the structural image is selected from the group consisting of a CT scan and an MRI scan.

13. 10. The apparatus of claim 1, wherein the electronic computer processes the structural image to create the mask depicting at least one organ using machine learning trained on various organ types.

14. The at least one organ is an intestine, and the physiological image is １８ The device of claim 1, which indicates the amount of F-FDG uptake.

15. The apparatus of claim 1 further comprising a device in communication with the electronic computer for acquiring structural and physiological images.

16. 1. A method of assessing organ health, comprising: The apparatus according to any one of claims 1 to 15, further comprising analyzing images of patient information. method.

17. The device of claim 1, further comprising modifying a treatment regimen for the patient based on the analyzing step.

17. The method of claim 16.

Images