Determining quantitative measures of tumors from biomedical images

Abstract

The present disclosure is directed to systems and methods for determining measures of tumors from biomedical images. A computing system can receive (i) a first biomedical image derived via a computed tomography (CT) scan of an organ associated with a tumor in a subject and (ii) a second biomedical image derived via a positron emission tomography (PET) scan of the organ. The computing system can determine, from the first biomedical image, a region of interest (ROI) corresponding to the organ associated with the tumor in the subject. The computing system can identify a portion in the second biomedical image corresponding to the ROI. The computing system can generate a plurality of measures of the tumor in the subject based on one or more contours of the portion. The computing system can store an association between the subject and the plurality of measures of the tumor.

Description

DETERMINING QUANTITATIVE MEASURES OF TUMORSFROM BIOMEDICAL IMAGESCROSS REFERENCES TO RELATED APPLICATIONS

[0001] The present application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 519,780, titled “Determining Quantitative Measures of Tumors from Biomedical Images,” filed August 15, 2023, which is incorporated by reference in its entirety.BACKGROUND

[0002] A digital image may include one or more features within. Various computer vision techniques may be used by a computing system to automatically detect the features from within the digital image.SUMMARY

[0003] Aspects of the present disclosure are directed to systems, methods, and computer-readable media for determining measures of tumors from biomedical images. A computing system can receive (i) a first biomedical image derived via a computed tomography (CT) scan of an organ associated with a tumor in a subject and (ii) a second biomedical image derived via a positron emission tomography (PET) scan of the organ. The computing system can determine, from the first biomedical image, a region of interest (ROI) corresponding to the organ associated with the tumor in the subject. The computing system can identify a portion in the second biomedical image corresponding to the ROI determined from the first biomedical image. The computing system can generate a plurality of measures of the tumor in the subject based on one or more contours of the portion. The computing system can store, using one or more data structures, an association between the subject and the plurality of measures of the tumor.

[0004] In some embodiments, the computing system can receive, via a user interface, a threshold defining at least one of an intensity or a size, at which to differentiate the portion corresponding to the tumor from a remainder of the second biomedical image. In some embodiments, the computing system can modify the portion within the second biomedical image based on the one or more organs selected from the plurality of organs.

[0005] In some embodiments, the computing system can receive, via a user interface, a selection of one or more organs from a plurality of organs. In some embodiments, the computing system can determine the one or more contours of the portion within the second biomedical image based on the one or more organs selected from the plurality of organs. In some embodiments, the computing system can resize a dimension of the second biomedical image to correspond a dimension of first biomedical image, prior to identification of the portion in the biomedical image corresponding to the ROI from the first biomedical image.

[0006] In some embodiments, the computing system can determine, from the first biomedical image, the ROI corresponding to a subset of organs selected from a plurality of organs, the subset of organs including (i) a first organ corresponding to a primary anatomical site for the tumor and (ii) a second organ corresponding to a metastatic anatomical site to which the tumor has spread. In some embodiments, the computing system can apply a machine learning (ML) model to the first biomedical image to determine the ROI. The ML model can be established using a training dataset comprising a plurality of examples. Each of the plurality of examples can identify (i) a respective biomedical image derived from a corresponding CT scan of an organ associated with a tumor and (ii) an annotation identifying a respective ROI defining the organ.

[0007] In some embodiments, the computing system can receive (i) the first biomedical image derived via the CT scan of a first volume corresponding to a torso region of the subject and including one or more organs at least one of which has the tumor and (ii) the second biomedical image derived via the PET scan of a second volume at least partially overlapping with the first volume in the subject. In some embodiments, the computing system can register the second biomedical image with the first biomedical image to identify the portion of the second biomedical image corresponding to the ROI determined from the first biomedical image.

[0008] In some embodiments, the plurality of measures can include at least one of (i) a radiomic feature associated with the PET scan of the organ, (ii) a number of contours identified from the portion, and (iii) an uptake volume of a radioactive tracer for the PET scan by the subject. In some embodiments, the computing system can generate a report to provide for the subject based on the association between the subject and the plurality of measures of the tumor.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The foregoing and other objects, aspects, features, and advantages of the disclosure will become more apparent and better understood by referring to the following description taken in conjunction with the accompanying drawings, in which:

[0010] FIG. 1 depicts a block diagram of a system for determining quantitative measures of tumors from biomedical images in accordance with an illustrative embodiment;

[0011] FIG. 2A depicts a block diagram of a process to segment images in the system for determining quantitative measures in accordance with an illustrative embodiment;

[0012] FIG. 2B depicts a block diagram of a process to extract features in the system for determining quantitative measures in accordance with an illustrative embodiment;

[0013] FIG. 2C depicts s a block diagram of a process to generate reports in the system for determining quantitative measures in accordance with an illustrative embodiment;

[0014] FIG. 3 depicts a flow diagram of a method of determining quantitative measures of tumors from biomedical images in accordance with an illustrative embodiment;

[0015] FIG. 4 depicts a diagram of a process of segmenting images, with a screenshot of a computed tomography (CT) scan image and a screenshot of a segmented CT scan image in accordance with an illustrative embodiment;

[0016] FIG. 5 depicts a diagram of a process of transferring contours between images of different modalities, with screenshot of a segmented computed tomography (CT) scan image and a screenshot of a positron emission tomography (PET) scan image with a transfer of the segment, in accordance with an illustrative embodiment;

[0017] FIG. 6 depicts a screenshot of a user interface in the system for determining quantitative measures in accordance with an illustrative embodiment;

[0018] FIGs. 7A and 7B each depict a screenshot of positron emission tomography (PET) scan image with definition of a region of interest (RO I) therein in accordance with an illustrative embodiment;

[0019] FIG. 8 depicts a diagram of a process to sample an image, with a screenshot of a resampled positron emission tomography (PET) scan image, in accordance with an illustrative embodiment;

[0020] FIG. 9A depicts a screenshot of a positron emission tomography (PET) scan image fused with computed tomography (CT) scan image, in accordance with an illustrative embodiment;

[0021] FIG. 9B depicts a screenshot of a positron emission tomography (PET) scan image with a region of interest (RO I) identified from a corresponding computed tomography (CT) scan image, in accordance with an illustrative embodiment;

[0022] FIG. 9C depicts a screenshot of a zoomed in view of the positron emission tomography (PET) scan image with into a region of interest (ROI), in accordance with an illustrative embodiment;

[0023] FIGs. 10A and 10B each depict screenshots of a report identifying various quantitative measures of tumors determined from a computed tomography (CT) scan image and a positron emission tomography (PET) scan image, in accordance with an illustrative embodiment; and

[0024] FIG. 11 depicts a block diagram of a server system and a client computer system in accordance with an illustrative embodiment.DETAILED DESCRIPTION

[0025] Following below are more detailed descriptions of various concepts related to, and embodiments of, systems and methods for determining quantitative measures of tumors from biomedical images. It should be appreciated that various concepts introduced above and discussed in greater detail below may be implemented in any of numerous ways, as the disclosed concepts are not limited to any particular manner of implementation. Examples of specific implementations and applications are provided primarily for illustrative purposes.

[0026] Section A describes systems and methods for determining quantitative measures of tumors from biomedical images.

[0027] Section B describes a network environment and computing environment which may be useful for practicing various computing related embodiments described herein.A. Systems and Methods for Determining Quantitative Measures of Tumors from Biomedical Images

[0028] Reliable imaging biomarkers facilitating management and outcome prognostication in patients with neuroendocrine tumors (NET) are lacking. A positron emission tomography (PET) image and computed tomography (CT) image analysis may be used to assess somatostatin receptor (SSTR) expression and stage patients with NET. However, manual segmentation of avid lesions to extract quantitative parameters for response assessment and prognostication remains time-consuming and may not be practical in widely metastatic disease. To address these technical challenges, presented herein in an automated (or semi -automated) platform for tumor segmentation and extraction of radiomic features. This platform was in 38 patients with SSTR-positive gastroenteropancreatic or lung NET imaged with PET / CT before and after peptide receptor radionuclide therapy with available outcome data.

[0029] Using CT images (of the PET / CT) as blueprints, a machine learning algorithm for image segmentation automatically computed masks for normal organs with physiologic uptake (e.g., adrenal glands, kidneys), and also identified the bone and the liver. Then, the PET was resampled based on the CT. To determine the normal liver standardized uptake value (SUV) mean and standard deviation (SD), a region-of-interest can be placed in the normal liver. For liver lesions, a threshold based on normal liver SUV mean was applied, for bone lesions a fixed SUV-threshold of 3. For the other lesions, a threshold based on average normal liver SUV mean derived from a representative patient cohort was used. Manual corrections can be applied. Minimum lesion size may be arbitrarily defined as > 0.3 cm3. Multiple output parameters including radiomic features and largest distance between two lesions (Dmax) may be extracted.

[0030] In the results, interim analyses (7 patients, 14 scans) showed segmentation of 186 lesions (116 liver, 10 bone, 19 lung, 57 soft tissue) with mean whole body (wb) tumor volume TV (ml) of 687.45 (range, 5.18-2021.65), wb TL 10485.74 (range, 55.39-37893.42), wb SUVmean 15.22 (range, 7.42-29.25), wb SUVmax 52.03 (range, 20.63-90.36); thereof mean liver TV of 655.04 (range, 0-1925.81), bone TV 1.33 (range, 0-5.48), soft tissue TV 31.00 (range, 0-201.75). Implementation of machine learning masks and segmentation required about 3.5-4 min per scan, and with manual adjustments, total segmentation time was 5-9 min per case. Radiomic features for individual lesions may be extracted. The personalized output report provided lesion uptake metrics and an embedded image graphic facilitating lesion tracking. In this manner, the platform can perform rapid automatic analyses of whole-body tumor burden and the extraction of radiomic features from PET and CT scan images even in widely metastatic disease.

[0031] Referring now to FIG. 1, depicted is a block diagram of a system 100 for determining quantitative measures of tumors from biomedical images. In overview, the system 100 can include at least one image processing system 105, at least one computed tomography (CT) scanner 110, a positron emission tomography (PET) scanner 115, at least one display 120, and at least one database 125, among others, communicatively coupled with one another via at least one network 130. The image processing system 105 can include at least one dataset acquirer 135, at least one segment detector 140, at least one contour transferrer 1145, at least one feature analyzer 150, and at least one report generator 155, among others. The image processing system 105 can include or provide at least one user interface 160. Each of the components in the system 100 as detailed herein may be implemented using hardware (e.g., one or more processors coupled with memory), or a combination of hardware and software as detailed herein in Section B. Each of the components in the system 100 may implement or execute the functionalities detailed herein.

[0032] In further detail, the image processing system 105 can be any computing device comprising one or more processors coupled with memory and software and capable of performing the various processes and tasks described herein. The image processing system 105 may be in communication with the CT scanner 110, the PET scanner 115, the display 120, the database 125, and other devices, via the network 130. The image processing system 105 may be situated, located, or otherwise associated with at least oneserver group. The server group may correspond to a data center, a branch office, or a site at which one or more servers corresponding to the image processing system 105 is situated.

[0033] Within the image processing system 105, the dataset acquirer 135 can retrieve biomedical images (e.g., CT scan image and PET scan image) of an organ within a subject. The segment detector 140 can determine a region of interest (RO I) corresponding to a feature (e.g., tumor) within the organ as depicted in a biomedical image in one modality (e.g., CT scan image). The contour transferrer 145 can identify a portion of a biomedical image in another modality (e.g., PET scan image) corresponding to the ROI associated with the feature. The feature analyzer 150 can determine quantitative measures associated with contours in the portion of the biomedical image. The report generator 155 can generate an output including information about the quantitative measures. The user interface 160 can be a graphical user interface (GUI) with one or more elements to control various processes of the image processing system 105.

[0034] The CT scanner 110 (sometimes herein generally referred to as an imaging device or an image acquirer) can be any device to acquire biomedical images in CT modality of a volume (e.g., including organs and tissue) of a subject. The CT scanner 110 can include a rotating X-ray tube with a set of detectors to measure an attenuation of X-rays as these pass through a volume of the subject. The CT scanner 110 can acquire multiple X- ray measurements acquired at different angels relative to the subject. The subject may be under guidance of clinician or hospital staff while scanned by the CT scanner 110. The CT scanner 110 can use any number of CT acquisition techniques, such as a sequence CT, a spiral CT, an electron beam tomography (EBT), and a dual source CT, among others.Based on the measurement of X-rays, the CT scanner 110 (or a computing device coupled thereto) can construct the biomedical image in the CT modality.

[0035] The PET scanner 115 (sometimes herein generally referred to as an imaging device or an image acquirer) can be any device to acquire biomedical images in PET modality of a volume (e.g., including organs and tissue) of a subject. To facilitate the PET scan, the subject can be injected by a clinician with a tracer (e.g., a radiopharmaceutical drug, such as 68Ga-DOTATATE, 18F-Fluorodeoxy glucose (18F-FDG), 1 IC-Methionine (11C-MET), 18F-Sodium Fluoride (18F-NaF), 18F-Florbetapir (18F-AV-45), and gallium- 68 PSMA-11 (68Ga-PSMA-l 1), among others). As the tracer undergoes beta plus decay, the material may emit positrons and the positrons may in turn collide with electrons, causingannihilation and emission of gamma rays. The PET scanner 115 can monitor for the emission of gamma rays from within the subject. The PET scanner 115 can be equipped with a set of detectors to measure the emission of gamma rays from various angles relative to the subject. Using the measurements, the PET scanner 115 (or a computing device coupled thereto) can construct the biomedical image in the PET modality.

[0036] The display 120 can be communicatively coupled with the image processing system 105 or any other computing device comprising one or more processors coupled with memory and software and capable of performing the various processes and tasks described herein. The display 120 may display, render, or otherwise present any information provided by the image processing system 105 or the images of subjects acquired via the CT scanner 610 or PET scanner 115, or both. The information may be used by a clinician examining a subject to define an administration of a treatment to the subject. The database 125 can include any be data storage to store and maintain data, such as metadata associated with subjects and biomedical images of subjects acquired via the CT scanner 110 and the PET scanner 115. The data stored on the database 125 can be accessible to the image processing system 105, the CT scanner 110, and the PET scanner 115, among others, via the network 130.

[0037] Referring now to FIG. 2A, depicted is a block diagram of a process 200 to segment images in the system 100 for determining quantitative measures. The process 200 can correspond to or include one or more operations performed within the system 100 to acquire and segment images. Under the process 200, the dataset acquirer 135 executing on the image processing system 105 can retrieve, identify, or otherwise receive biomedical images derived from a scan of one or more organs in a volume 210 of a subject 205. In some embodiments, the dataset acquirer 135 can receive the biomedical images from imaging devices, such as the CT scanner 110 and the PET scanner 115. In some embodiments, the dataset acquirer 135 can retrieve the biomedical images stored on the database 125.

[0038] The subject 205 can be a human or an animal at risk of cancer or suffering from cancer. The cancer may include, for example, a bone cancer (e.g., osteosarcoma, chondrosarcoma, chordoma, or Ewing sarcoma), a lung cancer (e.g., non-small cell lung cancer (NSCLC) or Small cell lung cancer (SCLC)), a breast cancer (e.g., Ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC), Lobular carcinoma in situ (LCIS), andInvasive lobular carcinoma (ILC)), prostate cancer (e.g., Adenocarcinoma, small cell, neuroendocrine, transitional cell carcinomas, sarcomas), melanoma (e.g., superficial spreading, nodular, lentigo maligna, and acral lentiginous), and colon cancer (e.g., adenocarcinoma, Gastrointestinal carcinoid tumors, lymphomas), among others.(0039] The cancer can be present in the form at least one tumor in one or more organs (e.g., gastrointestinal tract, brain, trachea, adrenal gland, spleen, gallbladder, kidney bone, liver, lung, breast, or colon) in the subject 205, within the scanned volume 210. The volume 210 can correspond to any portion of a body of the subject 205, such as a torso region, a peripheral region (e.g., one or both arms or legs), a neck region, head region, or an entire body, among others. Out of the organs, at least one organ can correspond to a primary anatomical site at which the cancer originated. At least one other organ corresponding to a metastatic (or secondary) anatomical site to which the cancer spread from the primary anatomical site. Different tracers can be used and injected in the subject 205 depending on the type of cancer in the subject to be detected. For example, for prostate cancer, the tracer 68Ga-PSMA can be used; for breast cancer, 18F-FES can be used; and for lung cancer or gastroenteropancreatic neuroendocrine tumors, 68GA-DOTATATE can be used.

[0040] The CT scanner 110 can scan, obtain, or otherwise acquire at least one CT image 215 of the one or more organs within the volume 210 in the subject 205. The CT scanner 110 can generate the CT image 215 of the volume 210 including the organs in accordance with a CT acquisition technique. For example, the CT scanner 110 can measure X-rays passing through the volume 210 of the subject 205 at multiple angles, and then use the X-ray measurements to produce the CTS image 215. The CT image 215 can be two- dimensional or three-dimensional.

[0041] In addition, the PET scanner 115 can scan, obtain, or otherwise acquire at least one PET image 220 of the one or more organs in the volume 210 in the subject 205. The volume 210 scanned by the PET scanner 115 can at least partially overlap (e.g., 50- 100%) with the volume 210 scanned by the CT scanner 110. The PET scanner 115 can generate the PET image 220 of the volume 210 including the organs of the subject 205. For instance, the PET scanner 115 can acquire emission of gamma rays from the beta decay of the radiopharmaceutical tracer injected in the subject 205. The PET image 220 can be two- dimensional or three-dimensional. Relative to the CT image 215, the PET image 220 canlack certain definition or details associated with the organs in the subject 205. Conversely, relative to the PET image 220, the CT image 215 can lack certain definition or details associated with the organs in the subject 205. As such, there may be details unascertainable from the one image modality but ascertainable from the other modality.[0042J With the acquisition, the dataset acquirer 135 can retrieve, identify, or otherwise receive the CT image 215 derived via the CT scanning or the PET image 220 derived via the PET scanning of the one or more organs in the volume 210 of the subject 205. The CT image 215 can be received from the CT scanner 110 or the database 125. The PET image 220 can retrieved from the PET scanner 115 or the database 125. In some embodiments, the dataset acquirer 135 can retrieve, identify, or otherwise receive information associated with the CT scanning of the subject 205 or information associated with the PET scanning of the subject 205. The information can include metadata, such as an identifier for the subject 205 (e.g., an anonymized identifier), scan date including start and end times, an identifier for the scanner (e.g., device serial number), date of injection of radiopharmaceutical tracer, amount of radiopharmaceutical tracer, and identifier on organs scanned in image, among others. The information can be retrieved from the CT scanner 110, the PET scanner 115, or the database 125.

[0043] The segment detector 140 executing on the image processing system 105 can detect, recognize, or otherwise determine at least one region of interest (RO I) 225 within the CT image 215. The ROI 225 can correspond to at least one organ in the subject 205. The organ can be associated with the tumor, for example, as the primary anatomical site for the cancer or the metastatic anatomical site to which the cancer spread. The ROI 225 can identify the organ associated with the tumor in terms of coordinates (e.g., pixel coordinates) within the CT image 215. In some embodiments, the ROI 225 can identify at least a portion of the organ corresponding to a feature in the organ, such as the tumor (e.g., malignant tumor) and lesion (e.g., abnormal tissue or other injury), among others, associated with the cancer. The ROI 225 can identify the tumor (or lesion) on the organ in terms of coordinates (e.g., pixel coordinates) within the CT image 215. In some embodiments, the segment detector 140 can determine multiple ROIs 225 corresponding to respective organs (or tumors or lesion in organs) in the scanned volume 210. The organs can include, for instance, one organ corresponding to a primary anatomical site from which the tumororiginated and another organ corresponding to a metastatic anatomical site to which the tumor has spread within the subject 205.

[0044] In some embodiments, the segment detector 140 can retrieve, identify, or receive a selection of the organ to be identified from the CT image 215 via the user interface 160. The segment detector 140 can provide the user interface 160 with a set of interface elements for selecting one or more of the organs to detect from the CT image 215. The user interface 160 can be presented via a display (e.g., coupled with the image processing system 105) to the user. The user can interact with the interface elements of the user interface 160 to enter the selection of the one or more organs. Upon detection of the entry, the segment detector 140 can identify the organ corresponding to the selected interface elements on the user interface 160. With the identification, the segment detector 140 can commence determination of the ROI 225 corresponding to the organ within the CT image 215.

[0045] The segment detector 140 can apply a machine learning (ML) model to the CT image 215 to detect, identify, or otherwise determine the ROI 225 from the CT image 215. The ML model can be any artificial intelligence model or algorithm to segment images to detect ROIs from the images. In general, the ML model can have an input corresponding to an image (e.g., a CT scan image) and an output corresponding to a segment identifying the ROI within the image. The ML model can include a set of weights (e.g., kernel parameters) relating the input to the output. The ML model may have been initialized, trained, and established using a training dataset in accordance with learning techniques (e.g., supervised or semi-supervised). The training dataset can include or identify a set of examples. Each example can include a respective image (e.g., CT scan image) of an organ with a tumor, as well as an annotation defining a ROI (e.g., the organ or the tumor or lesion within the organ) within the image. Upon training, the ML model can be used to recognize or detect the ROI within a respective input image.

[0046] In applying, the segment detector 140 can provide, input, or otherwise feed the CT image 215 into the ML model. Upon input, the segment detector 140 can process the CT image 215 in accordance with the set of weights in the ML model. From the processing, the segment detector 140 can determine the ROI 225 corresponding to the organ (or the tumor or lesion in the organ) within the volume 210 of the subject 205. In some embodiments, the segment detector 140 can produce, create, or otherwise generate at least one segmentation mask identifying the ROI 225 within the CT image 215. Thesegmentation mask can identify a set of coordinates (e.g., pixel coordinates) corresponding to the ROI 225 in the CT image 215. With the generation, the segment detector 140 can store and maintain an association of the CT image 215 and an identification of the ROI 225 using one or more data structures (e.g., an array, a matrix, a list, a table, a heap, or a tree) on the database 125.

[0047] Referring now to FIG. 2B, depicted is a block diagram of a process 230 to extract features in the system for determining quantitative measures. The process 230 can correspond to or include one or more operations performed within the system 100 to derive quantitative measures from biomedical images of organs in a subject. Under the process 230, the contour transferrer 145 executing on the image processing system 105 can determine or identify at least one portion 235 within the PET image 220 corresponding to the ROI 225 determined from the CT image 215. The portion 235 can correspond to the organ (or the tumor or lesion on the organ) associated with the tumor within the volume 210 scanned by the PET scanner 115. The portion 235 can correspond to a region within the PET image 220 corresponding to the ROI 225 within the CT image 215. Similar to the ROI 225, the portion 235 can be defined in terms of coordinates (e.g., pixel coordinates) within the PET image 220.

[0048] In some embodiments, the contour transferrer 145 can alter, change, or otherwise resize dimensions (e.g., width and height) of the PET image 220 to dimensions of the CT image 215. The resizing of the dimensions can be prior to the identification of the portion 235. To resize, the contour transferrer 145 can identify the dimensions of the CT image 215. With the identification, the contour transferer 145 can resample the PET image 220 to change the dimensions of the PET image 220 to correspond to match the dimensions of the CT image 215. When the dimensions of the CT image 215 are larger than the dimensions of the PET image 220, the contour transferrer 145 can up-sample the PET image 220 to resize. Conversely, when the dimensions of the CT image 215 are smaller than the dimensions of the PET image 220, the contour transferrer 145 can down-sample the PET image 220 to resize.

[0049] To identify, the contour transferrer 145 can move, port over, or otherwise copy the ROI 225 determined from the CT image 215 to the corresponding portion 235 within the PET image 220. For instance, the contour transferrer 145 can use the set of coordinates defining the ROI 225 within the CT image 215 to determine corresponding setof coordinates (e.g., pixel coordinates) in the PET image 220. Porting the coordinates over, the contour transferrer 145 can determine the corresponding portion 235 within the PET image 220. With the determination, the contour transferrer 145 can store and maintain an association of the PET image 220 and an identification of the portion 235 using one or more data structures (e.g., an array, a matrix, a list, a table, a heap, or a tree) on the database 125.

[0050] In some embodiments, the contour transferrer 145 can match or register the PET image 220 with the CT image 215 in identifying the portion 235. The scanned volume 210 for the PET image 220 can differ and partially overlap with the scanned volume 210 for the CT image 215. The registration can be in accordance with any number of image registration techniques, such as feature-based, intensity-based, or similarity-based (e.g., cross-correlation, mutual information, or intensity differences) techniques, among others. For example, the contour transferrer 145 can calculate a similarity measure (e.g., normalized mutual information) between the CT image 215 and the PET image 220. With the calculation, the contour transferer 145 can determine transformation parameters for the PET image 220 to correspond to the CT image 215 using the similarity measure. Using the transformation parameters, the contour transferrer 145 can conform the PET image 220 to the CT image 215 and can identify the portion 235 within the PET image 220 corresponding to the ROI 225 within the CT image 215.

[0051] The feature analyzer 150 executing on the image processing system 105 can retrieve, identify, or otherwise receive a set of parameters 240A-N (hereinafter referred to as parameters 240). The set of parameters 240 can be used to alter the PET image 220 to facilitate extraction of features from the PET image 220 and in particular the portion 235 therein. In some embodiments, the set of parameters 240 can identify or include one or more organs within the subject 205 to be evaluated using the PET image 220. In some embodiments, the set of parameters 240 can identify or includer a threshold defining an intensity value or a size at which to further differentiate the portion 235 from a remainder of the PET image 220. The threshold can define a value for an intensity of pixels or a size of a set of pixels forming a feature, below which to attenuate or suppress from the PET image 220 when evaluating. In some embodiments, the set of parameters 240 can be used to define modifications to visual properties (e.g., brightness, contrast, texture, depth, orientation, focus, transparency, and saturation) of the PET image 220 to facilitate in extraction of features. The set of parameters 240 can be received via the user interface 160.

[0052] The feature analyzer 150 can present or provide the user interface 160 with a set of user interface elements for entry of the set of parameters 240. The user interface 160 including the set of interface elements can be provided by the feature analyzer 150 for display (e.g., coupled with the image processing system 105). The user interface elements for the user interface 160 can include check boxes, radio buttons, scroll bars, buttons, and text boxes, among others. The user can enter or input the set of parameters 240 via the interface 160, such as the organs to be evaluated, the threshold values, or the modifications to be made to the visual properties, among others. The feature analyzer 150 can monitor for interactions with the user interface elements of the user interface 160 by the user. Upon entry, the feature analyzer 150 can retrieve, identify, or otherwise receive the set of parameters 240 via the user interface 160. With receipt, the feature analyzer 150 can identify the set of parameters 240 corresponding to the selection on the user interface 160.

[0053] In accordance with the set of parameters 240, the feature analyzer 150 can alter, change, or otherwise modify at least a part of the PET image 220, such as the portion 235 corresponding to the organ of the subject 205. When the set of parameters 240 identify the selected organs, the feature analyzer 150 can apply one or more modifications corresponding to the organ to accentuate or differentiate the portion 235 within the PET image 220. The modifications can identify an alteration or change to the visual properties (e.g., brightness, contrast, texture, depth, orientation, focus, transparency, and saturation) or a threshold (e.g., for intensity or size) to be applied within the portion 235.[0054[ To identify the modifications to apply, the feature analyzer 150 can use a rule set defining which modifications are to be applied for each respective organ to accentuate or differentiate a part of the image (e.g., the PET image 220) corresponding to the organ from the remainder of the image. With the identification, the feature analyzer 150 can change the visual properties in the portion 235 of the PET image 220, in accordance with the modifications specified for the organ. In some embodiments, when the set of parameters 240 identify the modifications to the visual properties, the feature analyzer 150 can apply the modifications to the PET image 220. The feature analyzer 150 can change the visual properties in the portion 235 of the PET image 220, in accordance with the modifications specified within the set of parameters 240 as inputted via the user interface 160.

[0055] In some embodiments, when the set of parameters 240 identify the threshold to be applied, the feature analyzer 150 can apply the threshold to at least the portion 235 within the PET image 220. In applying, if the threshold is for intensity value, the feature analyzer 150 can identify a value of each pixel within the portion 235 to compare against the threshold. When the intensity satisfies (e.g., greater than or equal to) the threshold, the feature analyzer 150 can maintain or boost (e.g., increase) the value of the corresponding pixel. Otherwise, when the intensity does not satisfy (e.g., less than) the threshold, the feature analyzer 150 can null or suppress (e.g., reduce) the value of the corresponding pixel.

[0056] In addition, if the threshold is for size, the feature analyzer 150 can detect or identify one or more areas within the portion 235. Each area can correspond to a feature (e.g., a non-negative space) within the portion 235. With the identification, the feature analyzer 150 can compare a size of the area with threshold size. When the size of the area satisfies (e.g., greater than or equal to) the threshold, the feature analyzer 150 can maintain or boost (e.g., increase) the values of the corresponding pixels within the area. Conversely, when the size of the area does not satisfy (e.g., less than) the threshold, the feature analyzer 150 can null or suppress (e.g., reduce) the values of the corresponding pixels within the area.

[0057] With the application of the set of parameters 240, the feature analyzer 150 can calculate, generate, or otherwise determine one or more contours 245A-N (hereinafter referred to as contours 245) from the portion 235 of the PET image 220. Each contour 245 can correspond to at least one outline or boundary within the portion 235 of the PET image 220 correlated with or associated with an uptake of the radiopharmaceutical tracer or a concentration of the tracers within the volume 210 in the subject 205. The contours 245 can be correlated or associated with the presence of the tumor (or lesion) within the organ depicted within the portion 235 of the PET image 220. The contours 245 can be used to quantify the tracer uptake within the organ corresponding the portion 235 in the PET image 220. In determining, the feature analyzer 150 can apply a computer vision algorithm to the portion 235 in the PET image 220. The computer vision algorithm can include, for example, an edge detection algorithm (e.g., Canny, Deriche, differential, Sobel operator, and Prewitt, and Roberts cross algorithms), an active contour model, a gradient vector flow, a Hough transform, a ridge detection algorithm, and thresholding, among others. Uponapplying, the feature analyzer 150 can detect the contours 245 within the portion 235 of the PET image 220.

[0058] Based on the determined contours 245, the feature analyzer 150 can output, determine, or otherwise generate a set of quantitative measures 250A-N (hereinafter referred to as quantitative measures 250 or more generally as measures). The quantitative measures 250 can define or identify various information regarding the tumor on the organ of subject 205, as depicted in the portion 235 of the PET image 220. The quantitative measures 250 can identify or include, for example, a radiomic feature associated with the PET scan of the organ, a number of contours identified from the portion, and an uptake volume of a radioactive tracer for the PET scan by the subject, among others. In some embodiments, the quantitative measures 250 can identify or include a distance metric between two portions 235 or a distance metric between two contours 245, among others. For instance, the distance metric can be a maximum distance between the center of two portions 235 corresponding to the furthest apart tumors or lesions associated with the cancer in the organ of the subject 205 To identify, the feature analyzer 150 can identify a centroid of each portion 235, calculate the distance between each pair of portions 235 (e.g., between a pair of tumors or lesions), and can select the maximum distance for a pair of portions 235 in the PET image 220.

[0059] In the quantitative measures 250, the radiomic features can define or identify a descriptor of features of the contours 245 or the PET image 220, and can include intensitybased features (e.g., a mean intensity, median intensity, standard deviation, a skewness, or kurtosis), histogram-based features (e.g., entropy, homogeneity, and contrast), shape-based features (e.g., volume, area, or compactness), and texture features (e.g., Gray-Level Cooccurrence Matrix (GLCM), Gray -Level Run-Length Matrix (GLRLM), Gray -Level Size Zone Matrix (GLSZM), and Gray -Level Dependence Matrix (GLDM)) among others. The number of contours can identify a number of different contours 245, with each associated with a respective concentration of tracers within the portion 235 of the PET image 220. The uptake volume can identify or correspond to a volume of tissue with the radiopharmaceutical tracer, in the organ associated with the portion 235 in the PET image 220. For the quantitative measures 250, the uptake volume related information can identify a standardized uptake volume (SUV) (e.g., concentration of radiotracer within a region ofinterest (RO I) of the body, normalized by injected dose of the radiotracer and weight of the subject 205), SUV mean, SUV intensity maximum, or SUV minimum, among others.

[0060] To generate, the feature analyzer 150 can determine the quantitative measures 250 as a function of the contours 245 determined from the portion 235 of the PET image 220. The feature analyzer 150 can use various visual characteristics (e.g., intensity, shape, area, or volume) of the contours 245 to identify, calculate, or otherwise determine corresponding values (e.g., radiomic features, number of contours, and the uptake volume) for the quantitative measures 250. For instance, the feature analyzer 150 can use the intensity values of the pixels associated the contours 245 to calculate radiomic features, such as a gray -level co-occurrence matrix (GCLM) and other texture based features. In some embodiments, the feature analyzer 150 can determine the quantitative measures 250 for each organ analyzed in the subject 205. For example, the feature analyzer 150 can generate the quantitative measures 250 for each of a bone, lung, liver, and heart, among others. With the generation, the feature analyzer 150 can store and maintain an association of the subject 205 with the set of quantitative measures 250 on the database 125. The association can be among the subject 205 (e.g., using an anonymized identifier), the organs, the metadata for the acquisition of the images, the CT image 215, the PET image 220, and the set of quantitative measures 250, among others.

[0061] Referring now to FIG. 2C, depicted is s a block diagram of a process 260 to generate reports in the system 100 for determining quantitative measures. The process 260 can correspond to or include operations performed in the system 100 to generate and provide reports. Under the process 260, the report generator 155 executing on the image processing system 105 can output, create, or otherwise generate at least one report 265. The generation of the report 265 can be based on the association between the subject 205 and the set of quantitative measures 250. In some embodiments, the report generator 155 can generate the report 265 using the association among the metadata, the subject 205 (e.g., using an anonymized identifier), the organs, the CT image 215, the PET image 220, and the set of quantitative measures 250.

[0062] The report 265 can include, for example, information on the metadata associated with the acquisition of the images (e.g., time and date of acquisition of the CT image 215 and the PET image 220), the CT image 215 and the PET image 220 themselves, an identification defining the portion 235 within the PET image 220, and the quantitativemeasures 250, among others. In some embodiments, the creation of the report 265 by the report generator 155 can be in accordance with a standardized template. The template may define, identify, or specify a format or structure in which the information is to be included in the report 265. The structure may be, for example, in the form of a field-value pair, defining which types of information are to be included in the report 265. The report generator 155 may store and maintain the report 265 on the database 125. The report 265 may be stored and maintained as one or more data structure (e.g., linked list, table, tree, array, matrix, or heap) or files (e.g., digital imaging and communication in medicine (DICOM) files, extensible markup language (XML) file, or comma-separated values (CSV) file), among others.

[0063] With the generation of the report 265, the report generator 155 can send, transmit, or otherwise provide the report 265 for presentation via the display 120 (e.g., to a computing device communicatively coupled with the image processing system 105). Upon receipt, the display 120 in turn can render, display, or otherwise present the information included in the report 265. For example, an application (e.g., native or web application) executing on a computing device can present the quantitative measures 250 in the report 265, along with the CT image 215 or the PET image 215. The information (e.g., the quantitative measures 250) presented in the report 265 can be used by the clinician examining the subject 205 to make further diagnosis and treatment planning to address the cancer presented within one or more organs within the subject 205. An example of the report 265 can be found in FIGs. 10A and 10B.

[0064] In this manner, the image processing system 105 can generate and provide information on the tumors associated with the cancer in the subject 205, from the CT image 215 and the PET image 220 in an automated or semi-automated manner. The input of parameters 240 can allow the image processing system 105 to better detect contours 245 from the portions 235 of the PET image 220 and by extension derive quantitative measures 250 using the contours 245. With semi-automation, the image processing system 105 can detect these contours 245 with higher reliability and accuracy, compared to approaches that can miss contours 245 and by extension quantitative measures 250 from such contours 245. Furthermore, the image processing system 105 can reduce the amount of user interactions for determining the quantitative measures 250 from the PET image 220 using data fromprocessing the CT image 215, relative to more manual approaches in which users are to repeatedly specify how the machine is to analyze images.

[0065] In addition, the reduction of user interactions can lower the amount of consumption in computing resources (e.g., processing and memory) and network bandwidth that would have otherwise spent on handling repeated user interactions (e.g., to modify threshold or focus the analysis on particular regions). The image processing system 105 can also significantly shorten the time between retrieval of the CT image 215 and the PET image 220 and extraction of quantitative features 250, from the scale of hours to minutes. This can allow for the rapid creation of reports 265 identifying information about the CT image 215 and the PET image 220. The standard format for the report 265 may permit the loading and usage of the report 265 across a wide variety of platforms, allowing a greater number of authorized users to access the information contained in the report 265. With less human involvement, the quality of human computer interaction (HCI) with the user can be improved. Furthermore, the report 265 can be used by a clinician to more accurately pinpoint and remedy the presence of tumors associated with cancer within the subject 205.

[0066] Referring now to FIG. 3, depicted is a flow diagram of a method 300 of determining quantitative measures of tumors from biomedical images. The method 300 may be performed by or implemented using the system 100 described herein in conjunction with FIGs. 1-2C or the system 1100 detailed herein in Section C. Under the method 300, a computing system (e.g., the image processing system 105) can receive a dataset including a computer tomography (CT) scan image (e.g., the CT image 215) and a positron emission tomography (PET) scan image (e.g., the PET image 220) (305). The computing system can determine a region of interest (ROI) (e.g., the ROI 225) from the CT scan image (310). The computing system can identify a portion (e.g., the portion 235) in the PET scan image corresponding to the ROI (315). The computing system can receive a set of selection parameters (e.g., the parameters 240) (320). The computing system can determine contours (e.g., the contours 245) within portion using the selection parameters (325). The computing system can generate quantitative measures (e.g., the contours 245) for tumor from the contours (330). The computing system can provide report (e.g., the report 265) with quantitative measures (335).

[0067] Referring now to FIG. 4, depicted is diagram of a process 400 of segmentation using a screenshot of a computed tomography (CT) scan image 405 togenerate a screenshot of a segmented CT scan image 410. The CT scan image 405 can be an instance of the CT image 215 discussed above. The segmented CT scan image 405 can correspond to the CT scan image 410, with one or more segments. In the segmented CT image 405, each segment can identify a region within the CT scan image 405 corresponding to a respective organ within a subject. The segmentation may have been performed using a machine learning (ML) model applied to the acquired CT scan image 405.

[0068] Referring now to FIG. 5, depicted is a process 500 of transferring segments between images of different modalities. Under the process 500, a three-dimension (3D) computed tomography (CT) scan image 505A and a corresponding two-dimensional (2D) CT scan image 505B may have been segmented. The 3D CT scan image 505A can include volume segments each identifying a respective organ, and the corresponding 2D CT scan image 505B can include sectional segments each also identifying a respective organ. With the segmentation performed, the segments can be transferred or copied over to image of positron emission tomography (PET) scan modalities. The segments of the 3D CT scan image 505A can be transferred over to a 3D PET scan image 510A. Likewise, the segments of the 2D CT scan image 5105B can be moved over to the 2D PET scan image 510B. The segments can be transferred to facilitate further processing and analysis of the organs using the PET scan images. Certain feature related to the tumor and cancer may be more easily ascertainable in the PET scan images but may be difficult to ascertain from the CT scan images, and vice-versa.[0069 J Referring now to FIG. 6, depicted a screenshot of a user interface 600 in the system for determining quantitative measures. The user interface 600 can correspond to an instance of the user interface 160 and can be used to enter at least one or more of the values for the set of parameters 240. As depicted, the user interface 600 can include a set of user interface elements corresponding to organs to be excluded from analysis, another set of user interface elements corresponding to exclusions of parts of a gastrointestinal organ from analysis, as well as smallest contour and pixel edge, among others. The user can make selections on the user interface 600 to enter values for the parameters 240 to be used for analysis.

[0070] Referring now to FIGs. 7A and 7B, depicted are screenshots 700 and 750 of positron emission tomography (PET) scan image with definition of a region of interest (ROI) therein. In context, the user can use the user interface 600 to focus in on certainregions within the PET scan image, as indicated in the arrow in screenshots 700 and 750. The screenshot 700 can be a PET scan image of a body (including torso and head) of a subject. The screenshot 750 can be a PET scan image zoomed in on a particular organ (e.g., the liver).

[0071] Referring now to FIG. 8, depicted is a diagram a process 800 to resample images. Under the process, a positron emission tomography (PET) scan image can be resampled (e.g., down-sampled or up-sampled), such that the dimensions (e.g., height and width) of the PET scan image is to match the dimensions of the corresponding CT scan image of the same volume within the subject. In addition, zoom constraint (an example of a parameter) can be set to a particular value for different organs to extend the definition about the organ (e.g., corresponding to the portion 235 in the PET image 220). For example, the zoom constraint for a liver can be extended by 3 pixels or 4.1016 mm and the zoom constraint for another organ can be also extended by 3 pixels or 4.1016 mm as depicted.

[0072] Referring now to FIG. 9A, depicted is a screenshot 900 of a positron emission tomography (PET) scan image fused with computed tomography (CT) scan image. In context, the user can use a user interface (e.g., the user interface 160) to overlay a CT scan image with a PET scan image (on left and right-side of depiction). The user can identify a region within both the CT scan image and the PET scan image to be further analyzed, for example, as identified in the cross-mark. Referring now to FIG. 9B, depicted is a screenshot 905 of a positron emission tomography (PET) scan image with a region of interest (ROI) identified from a corresponding computed tomography (CT) scan image. In context, the user can user the user interface (e.g., the user interface 160) to focus in an on a particular region within the PET image, as identified in the cross-mark in the depicted example. Referring now to FIG. 9C, depicted is a screenshot 910 of a zoomed in view of the positron emission tomography (PET) scan image with into a region of interest (ROI). In context, the user can user the user interface (e.g., the user interface 160) to focus in an on a particular region within the CT image, corresponding to the focused-on region in the screenshot 905 of the PET scan image.

[0073] Referring now to FIGs. 10A and 10B, depicted is screenshots of a report 1000 identifying various quantitative measures of tumors determined from a computed tomography (CT) scan image and a positron emission tomography (PET) scan image. The report 1000 can be an instance of the report 625 discussed above. As depicted, starting fromFIG. 10 A, the report 1000 can identify or include patient information, such as name, identifier, gender, age, birth date, weight, and weight. The report 1000 can also identify or include a rendering, such as a three-dimensional rendering of the CT or PET image with the segments defining the organs within the scanning volume of the subject. The report 1000 can identify or include information associated with acquisition of the CT or PET image of the subject, such as a series data time, acquisition date time, injection date time, a radiopharmaceutical tracer injected into the subject, a radionuclide, an amount of half-life, a standardized uptake volume (SUV) additional decay correction, administered activity, and total calculated activity, among others. Moving onto FIG. 10B, the report 1000 can identify or include a metabolic tumor volume (MTV) (e.g., entire tumor burden over volume) and a total lesion glycolysis (TLG) defined as a product of SUV and MTV. The report 1000 can identify or include contour information, such as number of contours, mean, maximum, and volume of contours, among others, for each anatomical location, such as bone, soft tissue, lung, liver, parotid, and blood pool as depicted. The report 1000 can also include an uptake volume histogram, as a function of uptake and total volume fraction.B . Computing and Network Environment

[0074] Various operations described herein can be implemented on computer systems. FIG. 11 shows a simplified block diagram of a representative server system 1100, client computing system 1114, and network 1126 usable to implement certain embodiments of the present disclosure. In various embodiments, server system 1100 or similar systems can implement services or servers described herein or portions thereof. Client computing system 1114 or similar systems can implement clients described herein. The system 100 described herein can be similar to the server system 1100. Server system 1100 can have a modular design that incorporates a number of modules 1102 (e.g., blades in a blade server embodiment); while two modules 1102 are shown, any number can be provided. Each module 1102 can include processing unit(s) 1104 and local storage 1106.

[0075] Processing unit(s) 1104 can include a single processor, which can have one or more cores, or multiple processors. In some embodiments, processing unit(s) 1104 can include a general-purpose primary processor as well as one or more special-purpose coprocessors such as graphics processors, digital signal processors, or the like. In some embodiments, some or all processing units 1104 can be implemented using customized circuits, such as application specific integrated circuits (ASICs) or field programmable gatearrays (FPGAs). In some embodiments, such integrated circuits execute instructions that are stored on the circuit itself. In other embodiments, processing unit(s) 1104 can execute instructions stored in local storage 1106. Any type of processors in any combination can be included in processing unit(s) 1104.

[0076] Local storage 1106 can include volatile storage media (e.g., DRAM, SRAM, SDRAM, or the like) and / or non-volatile storage media (e.g., magnetic or optical disk, flash memory, or the like). Storage media incorporated in local storage 1106 can be fixed, removable or upgradeable as desired. Local storage 1106 can be physically or logically divided into various subunits such as a system memory, a read-only memory (ROM), and a permanent storage device. The system memory can be a read-and-write memory device or a volatile read-and-write memory, such as dynamic random-access memory. The system memory can store some or all of the instructions and data that processing unit(s) 1104 need at runtime. The ROM can store static data and instructions that are needed by processing unit(s) 1104. The permanent storage device can be a non-volatile read-and-write memory device that can store instructions and data even when module 1102 is powered down. The term “storage medium” as used herein includes any medium in which data can be stored indefinitely (subject to overwriting, electrical disturbance, power loss, or the like) and does not include carrier waves and transitory electronic signals propagating wirelessly or over wired connections.

[0077] In some embodiments, local storage 1106 can store one or more software programs to be executed by processing unit(s) 1104, such as an operating system and / or programs implementing various server functions such as functions of the system 100 of FIG. 1 or any other system described herein, or any other server(s) associated with system 100 or any other system described herein.

[0078] Software” refers generally to sequences of instructions that, when executed by processing unit(s) 1104 cause server system 1100 (or portions thereof) to perform various operations, thus defining one or more specific machine embodiments that execute and perform the operations of the software programs. The instructions can be stored as firmware residing in read-only memory and / or program code stored in non-volatile storage media that can be read into volatile working memory for execution by processing unit(s) 1104. Software can be implemented as a single program or a collection of separate programs or program modules that interact as desired. From local storage 1106 (or non-local storage described below), processing unit(s) 1104 can retrieve program instructions to execute and data to process in order to execute various operations described above.

[0079] In some server systems 1100, multiple modules 1102 can be interconnected via a bus or other interconnect 1108, forming a local area network that supports communication between modules 1102 and other components of server system 1100. Interconnect 1108 can be implemented using various technologies including server racks, hubs, routers, etc.

[0080] A wide area network (WAN) interface 1110 can provide data communication capability between the local area network (interconnect 1108) and the network 1126, such as the Internet. Technologies can be used, including wired (e.g., Ethernet, IEEE 1102.3 standards) and / or wireless technologies (e.g., Wi-Fi, IEEE 1102.11 standards).

[0081] In some embodiments, local storage 1106 is intended to provide working memory for processing unit(s) 1104, providing fast access to programs and / or data to be processed while reducing traffic on interconnect 1108. Storage for larger quantities of data can be provided on the local area network by one or more mass storage subsystems 1112 that can be connected to interconnect 1108. Mass storage subsystem 1112 can be based on magnetic, optical, semiconductor, or other data storage media. Direct attached storage, storage area networks, network-attached storage, and the like can be used. Any data stores or other collections of data described herein as being produced, consumed, or maintained by a service or server can be stored in mass storage subsystem 1112. In some embodiments, additional data storage resources may be accessible via WAN interface 1110 (potentially with increased latency).

[0082] Server system 1100 can operate in response to requests received via WAN interface 1110. For example, one of modules 1102 can implement a supervisory function and assign discrete tasks to other modules 1102 in response to received requests. Work allocation techniques can be used. As requests are processed, results can be returned to the requester via WAN interface 1110. Such operation can generally be automated. Further, in some embodiments, WAN interface 1110 can connect multiple server systems 1100 to each other, providing scalable systems capable of managing high volumes of activity. Other techniques for managing server systems and server farms (collections of server systems that cooperate) can be used, including dynamic resource allocation and reallocation.

[0083] Server system 1100 can interact with various user-owned or user-operated devices via a wide-area network such as the Internet. An example of a user-operated device is shown in FIG. 11 as client computing system 1114. Client computing system 1114 can be implemented, for example, as a consumer device such as a smartphone, other mobile phone, tablet computer, wearable computing device (e.g., smart watch, eyeglasses), desktop computer, laptop computer, and so on.

[0084] For example, client computing system 1114 can communicate via WAN interface 1110. Client computing system 1114 can include computer components such as processing unit(s) 1116, storage device 1118, network interface 1120, user input device 1122, and user output device 1124. Client computing system 1114 can be a computing device implemented in a variety of form factors, such as a desktop computer, laptop computer, tablet computer, smartphone, other mobile computing device, wearable computing device, or the like.

[0085] Processing unit(s) 1116 and storage device 1118 can be similar to processing unit(s) 1104 and local storage 1106 described above. Suitable devices can be selected based on the demands to be placed on client computing system 1114; for example, client computing system 1114 can be implemented as a “thin” client with limited processing capability or as a high-powered computing device. Client computing system 1114 can be provisioned with program code executable by processing unit(s) 1116 to enable various interactions with server system 1100.

[0086] Network interface 1120 can provide a connection to the network 1126, such as a wide area network (e.g., the Internet) to which WAN interface 1110 of server system 1100 is also connected. In various embodiments, network interface 1120 can include a wired interface (e.g., Ethernet) and / or a wireless interface implementing various RF data communication standards such as Wi-Fi, Bluetooth, or cellular data network standards (e.g., 3G, 4G, LTE, etc ).

[0087] User input device 1122 can include any device (or devices) via which a user can provide signals to client computing system 1114; client computing system 1114 can interpret the signals as indicative of particular user requests or information. In various embodiments, user input device 1122 can include any or all of a keyboard, touch pad, touchscreen, mouse or other pointing device, scroll wheel, click wheel, dial, button, switch, keypad, microphone, and so on.

[0088] User output device 1124 can include any device via which client computing system 1114 can provide information to a user. For example, user output device 1124 can include a display to display images generated by or delivered to client computing system 1114. The display can incorporate various image generation technologies, e.g., a liquid crystal display (LCD), light-emitting diode (LED) including organic light-emitting diodes (OLED), projection system, cathode ray tube (CRT), or the like, together with supporting electronics (e.g., digital -to-analog or analog-to-digital converters, signal processors, or the like). Some embodiments can include a device such as a touchscreen that function as both input and output device. In some embodiments, other user output devices 1124 can be provided in addition to or instead of a display. Examples include indicator lights, speakers, tactile “display” devices, printers, and so on.

[0089] Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a computer-readable storage medium. Many of the features described in this specification can be implemented as processes that are specified as a set of program instructions encoded on a computer-readable storage medium. When these program instructions are executed by one or more processing units, they cause the processing unit(s) to perform various operation indicated in the program instructions. Examples of program instructions or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter. Through suitable programming, processing unit(s) 1104 and 1116 can provide various functionality for server system 1100 and client computing system 1114, including any of the functionality described herein as being performed by a server or client, or other functionality.

[0090] It will be appreciated that server system 1100 and client computing system 1114 are illustrative and that variations and modifications are possible. Computer systems used in connection with embodiments of the present disclosure can have other capabilities not specifically described here. Further, while server system 1100 and client computing system 1114 are described with reference to particular blocks, it is to be understood that these blocks are defined for convenience of description and are not intended to imply aparticular physical arrangement of component parts. For instance, different blocks can be but need not be located in the same facility, in the same server rack, or on the same motherboard. Further, the blocks need not correspond to physically distinct components. Blocks can be configured to perform various operations, e.g., by programming a processor or providing appropriate control circuitry, and various blocks might or might not be reconfigurable depending on how the initial configuration is obtained. Embodiments of the present disclosure can be realized in a variety of apparatus including electronic devices implemented using any combination of circuitry and software.

[0091] While the disclosure has been described with respect to specific embodiments, one skilled in the art will recognize that numerous modifications are possible. Embodiments of the disclosure can be realized using a variety of computer systems and communication technologies including but not limited to the specific examples described herein. Embodiments of the present disclosure can be realized using any combination of dedicated components and / or programmable processors and / or other programmable devices. The various processes described herein can be implemented on the same processor or different processors in any combination. Where components are described as being configured to perform certain operations, such configuration can be accomplished, e.g., by designing electronic circuits to perform the operation, by programming programmable electronic circuits (such as microprocessors) to perform the operation, or any combination thereof. Further, while the embodiments described above may make reference to specific hardware and software components, those skilled in the art will appreciate that different combinations of hardware and / or software components may also be used and that particular operations described as being implemented in hardware might also be implemented in software or vice versa.

[0092] Computer programs incorporating various features of the present disclosure may be encoded and stored on various computer-readable storage media; suitable media include magnetic disk or tape, optical storage media such as compact disk (CD) or DVD (digital versatile disk), flash memory, and other non-transitory media. Computer-readable media encoded with the program code may be packaged with a compatible electronic device, or the program code may be provided separately from electronic devices (e.g., via Internet download or as a separately packaged computer-readable storage medium).

[0093] Thus, although the disclosure has been described with respect to specific embodiments, it will be appreciated that the disclosure is intended to cover all modifications and equivalents within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A method of determining measures of tumors from biomedical images, comprising: receiving, by a computing system, (i) a first biomedical image derived via a computed tomography (CT) scan of an organ associated with a tumor in a subject and (ii) a second biomedical image derived via a positron emission tomography (PET) scan of the organ; determining, by the computing system, from the first biomedical image, a region of interest (RO I) corresponding to the organ associated with the tumor in the subject; identifying, by the computing system, a portion in the second biomedical image corresponding to the ROI determined from the first biomedical image; generating, by the computing system, a plurality of measures of the tumor in the subject based on one or more contours of the portion; and storing, by the computing system, using one or more data structures, an association between the subject and the plurality of measures of the tumor.

2. The method of claim 1, further comprising: receiving, by the computing system via a user interface, a threshold defining at least one of an intensity or a size, at which to differentiate the portion corresponding to the tumor from a remainder of the second biomedical image; and determining, by the computing system, the one or more contours of the portion within the second biomedical image using the threshold.

3. The method of claim 1, further comprising: receiving, by the computing system via a user interface, a selection of one or more organs from a plurality of organs; and modifying, by the computing system, the portion within the second biomedical image based on the one or more organs selected from the plurality of organs.

4. The method of claim 1, further comprising resizing, by the computing system, a dimension of the second biomedical image to correspond a dimension of first biomedical image, prior to identification of the portion in the biomedical image corresponding to the ROI from the first biomedical image.

5. The method of claim 1, wherein determining further comprises determining, from the first biomedical image, the ROI corresponding to a subset of organs selected from a plurality of organs, the subset of organs including (i) a first organ corresponding to a primary anatomical site for the tumor and (ii) a second organ corresponding to a metastatic anatomical site to which the tumor has spread.

6. The method of claim 1, wherein determining further comprises applying a machine learning (ML) model to the first biomedical image to determine the ROI, wherein the ML model is established using a training dataset comprising a plurality of examples, each of the plurality of examples identifying (i) a respective biomedical image derived from a corresponding CT scan of an organ associated with a tumor and (ii) an annotation identifying a respective ROI defining the organ.

7. The method of claim 1, wherein receiving further comprises receiving (i) the first biomedical image derived via the CT scan of a first volume corresponding to a torso region of the subject and including one or more organs at least one of which has the tumor and (ii) the second biomedical image derived via the PET scan of a second volume at least partially overlapping with the first volume in the subject.

8. The method of claim 1, wherein identifying further comprises registering the second biomedical image with the first biomedical image to identify the portion of the second biomedical image corresponding to the ROI determined from the first biomedical image.

9. The method of claim 1, wherein the plurality of measures further comprises at least one of (i) a radiomic feature associated with the PET scan of the organ, (ii) a number of contours identified from the portion, and (iii) an uptake volume of a radioactive tracer for the PET scan by the subj ect.

10. The method of claim 1, further comprising generating, by the computing system, a report to provide for the subject based on the association between the subject and the plurality of measures of the tumor.

11. A system for determining measures of tumors from biomedical images, comprising: a computing system having one or more processors coupled with memory, configured to: receive (i) a first biomedical image derived via a computed tomography (CT) scan of an organ associated with a tumor in a subject and (ii) a second biomedical image derived via a positron emission tomography (PET) scan of the organ; determine, from the first biomedical image, a region of interest (ROI) corresponding to the organ associated with the tumor in the subject; identify a portion in the second biomedical image corresponding to the ROI determined from the first biomedical image; generate a plurality of measures of the tumor in the subject based on one or more contours of the portion; and store, using one or more data structures, an association between the subject and the plurality of measures of the tumor.

12. The system of claim 11, wherein the computing system is further configured to receive, via a user interface, a threshold defining at least one of an intensity or a size, at which to differentiate the portion corresponding to the tumor from a remainder of the second biomedical image; and determine the one or more contours of the portion within the second biomedical image using the threshold.

13. The system of claim 11, wherein the computing system is further configured to receive, via a user interface, a selection of one or more organs from a plurality of organs; and modify the portion within the second biomedical image based on the one or more organs selected from the plurality of organs.

14. The system of claim 11, wherein the computing system is further configured to resize a dimension of the second biomedical image to correspond a dimension of first biomedical image, prior to identification of the portion in the biomedical image corresponding to the ROI from the first biomedical image.

15. The system of claim 11, wherein the computing system is further configured to determine, from the first biomedical image, the ROI corresponding to a subset of organs selected from a plurality of organs, the subset of organs including (i) a first organ corresponding to a primary anatomical site for the tumor and (ii) a second organ corresponding to a metastatic anatomical site to which the tumor has spread.

16. The system of claim 11, wherein the computing system is further configured to apply a machine learning (ML) model to the first biomedical image to determine the ROI, wherein the ML model is established using a training dataset comprising a plurality of examples, each of the plurality of examples identifying (i) a respective biomedical image derived from a corresponding CT scan of an organ associated with a tumor and (ii) an annotation identifying a respective ROI defining the organ.

17. The system of claim 11, wherein the computing system is further configured to receive (i) the first biomedical image derived via the CT scan of a first volume corresponding to a torso region of the subject and including one or more organs at least one of which has the tumor and (ii) the second biomedical image derived via the PET scan of a second volume at least partially overlapping with the first volume in the subject.

18. The system of claim 11, wherein the computing system is further configured to register the second biomedical image with the first biomedical image to identify the portion of the second biomedical image corresponding to the ROI determined from the first biomedical image.

19. The system of claim 11, wherein the plurality of measures further comprises at least one of (i) a radiomic feature associated with the PET scan of the organ, (ii) a number of contours identified from the portion, and (iii) an uptake volume of a radioactive tracer for the PET scan by the subj ect.

20. The system of claim 11, wherein the computing system is further configured to generate a report to provide for the subject based on the association between the subject and the plurality of measures of the tumor.

Images