Use of computational pathology to generate medical prognosis for cancer patients

Abstract

In some embodiments, the present disclosure relates to a method that includes accessing digitized pathology imaging data from a cancer patient. The digitized pathology imaging data has been segmented to identify a plurality of cancer nuclei within one or more tumor regions. A plurality of nuclei diversity features are extracted using the plurality of cancer nuclei. The plurality of nuclei diversity features include a combination of shape features and second order features generated using the shape features. The plurality of nuclei diversity features are provided to a machine learning model that is trained to generate a medical prognosis regarding the cancer patient.

Description

USE OF COMPUTATIONAL PATHOLOGY TO GENERATE MEDICAL PROGNOSIS FOR CANCER PATIENTSREFERENCE TO RELATED APPLICATION

[0001] This Application claims the benefit of U.S. Provisional Application No. 63 / 574,980, filed on April 5, 2024, the contents of which are incorporated by reference in their entirety.FEDERAL FUNDING INFORMATION

[0002] This invention was made with government support under R01 CA268287 awarded by the National Institutes of Health / National Cancer Institute. The government has certain rights in the invention.BACKGROUND

[0003] Prostate cancer is an uncontrolled growth of cells in the prostate, aa small walnut-shaped gland in males that produces seminal fluid that nourishes and transports sperm. Prostate cancer is one of the most common types of cancer. Prostate cancer typically causes no symptoms in its early stages. While some types of prostate cancer grow slowly and may need minimal or even no treatment, other types are aggressive and can spread quickly. Prostate cancer that's detected when it's still confined to the prostate gland has the best chance for successful treatment.BRIEF DESCRIPTION OF THE DRAWINGS

[0004] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate various example operations, apparatus, methods, and other example embodiments of various aspects discussed herein. It will be appreciated that the illustrated element boundaries (e.g., boxes, groups of boxes, or other shapes) in the figures represent one example of the boundaries. One of ordinary skill in the art will appreciate that, in some examples, one element can be designed as multiple elements or that multiple elements can be designed as one element. In some examples, an element shown as an internal component of another element may be implemented as an external component and vice versa. Furthermore, elements may not be drawn to scale.

[0005] Fig. 1 illustrates some embodiments of a block diagram of an assessment system configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0006] Figs. 2A-2B illustrate some embodiments of block diagrams relating to a feature extraction tool configured to extract nuclei diversity features from digitized pathology imaging data.

[0007] Fig. 2C illustrates some embodiments of a block diagram of an assessment system configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0008] Fig. 3 illustrates some additional embodiments of a block diagram of an assessment system configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0009] Fig. 4 illustrates some additional embodiments of a block diagram of an assessment system configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0010] Fig. 5 illustrates a flow diagram showing some embodiments of a method of using nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0011] Fig. 6 illustrates a block diagram of some additional embodiments of a block diagram of an assessment system configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0012] Figs. 7A-7B illustrate exemplary graphs showing performance metrics of the disclosed assessment system configured to generate a medical prognosis relating to BCR.

[0013] Fig. 8 illustrates some embodiments of a block diagram of an apparatus configured to generate a medical prognosis by operating a machine learning circuit on nuclei diversity features extracted from digitized pathology imaging data.DETAILED DESCRIPTION

[0014] The description herein is made with reference to the drawings, wherein like reference numerals are generally utilized to refer to like elements throughout, and wherein the various structures are not necessarily drawn to scale. In the following description, for purposes of explanation, numerous specific details are setforth in order to facilitate understanding. It may be evident, however, to one of ordinary skill in the art, that one or more aspects described herein may be practiced with a lesser degree of these specific details. In other instances, known structures and devices are shown in block diagram form to facilitate understanding.

[0015] Prostate cancer is often detected using screening tests. For example, blood tests may be used to detect levels of prostate-specific antigen (PSA), which indicate unusual growth of prostate tissue. If high levels of PSA are detected, a diagnosis of prostate cancer typically requires a biopsy of the prostate. The biopsy removes tissue from the prostate. The tissue is subsequently embedded in paraffin and sliced into thin sections that are used to form histology slides. Examination of the histology slides is then performed to diagnose prostate cancer and / or to identify cancer staging.

[0016] Once a patient has been diagnosed with prostate cancer, treatment options for the patient may vary depending on factors such as cancer stage, aggressiveness, overall patient health, and / or the like. One common treatment for early stage prostate cancer is a radical prostatectomy. A radical prostatectomy is a surgical procedure during which an entire prostate gland, some surrounding tissue, and the seminal vesicles may be removed. The goal of a radical prostatectomy is to remove all cancer cells.

[0017] In some cases, after a radical prostatectomy, PSA levels may begin to rise again in a patient’s blood. Biochemical recurrence (BCR) is a condition in which PSA levels in the blood of a prostate cancer patient increase after treatment with surgery and / or radiation (e.g., to above levels exceeding 0.2 ng / mL). Current tools for predicting BCR after a radical prostatectomy often depend on parameters determined by pathologists, such as tumor grade. However, tumor grade is known to vary between reviewers. Genomic risk classifiers may also provide useful information relating to BCR, but require substantial tissue, are costly, and are not commonly accessible in many medical centers. Therefore, there is an urgent, unmet clinical need for predictive biomarkers that can guide therapeutic decision-making, minimize ineffective treatments, and enable more assertive therapy in prostate cancer patients showing a high-risk of BCR.

[0018] In some embodiments, the present disclosure relates to a method and apparatus configured to use artificial intelligence to provide an automated riskassessment of biochemical recurrence for a cancer patient after the cancer patient has undergone a radical prostatectomy. In some embodiments, the method may be performed by accessing digitized pathology imaging data of prostate tissue excised from a cancer patient. The digitized pathology imaging data has been segmented to identify a plurality of cancer nuclei. One or more shape features are generated from the plurality of cancer nuclei, a plurality of second order features are generated using the one or more shape features, and statistical measures of the plurality of second order features are generated to form a plurality of nuclei diversity features. The plurality of nuclei diversity features are provided to a machine learning model that is trained to generate a medical prognosis relating to BCR. The medical prognosis can be utilized by health care professionals to make a more informed decision relating to the treatment of a patient, thereby allowing for the patient to have an improved quality of life (e.g., a lower risk of death, to avoid negative side effects of drugs that are not likely to produce positive outcomes, and / or the like).

[0019] Fig. 1 illustrates some embodiments of a block diagram of an assessment system 100 configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0020] The assessment system 100 comprises a memory 101 configured to store digitized pathology imaging data 102 for one or more cancer patients (e.g., prostate cancer patients, cervical cancer patients, breast cancer patients, and / or the like). In some embodiments, the digitized pathology imaging data 102 comprises one or more digitized pathology images 104 (e.g., one or more digitized biopsy slides) obtained from a pathological tissue sample excised from a cancer patient. In some embodiments, the digitized pathology imaging data 102 includes imaging data for a cancer patient that has been treated for prostate cancer with a radical prostatectomy.

[0021] In some embodiments, a segmentation tool 106 is configured to access the digitized pathology imaging data 102 (e.g., the one or more digitized pathology images 104). The segmentation tool 106 is further configured to identify a plurality of cancer nuclei 111 within the digitized pathology imaging data 102. In some embodiments, the segmentation tool 106 comprises a first segmentation stage 108 and a second segmentation stage 110 downstream of the first segmentation stage 108. The first segmentation stage 108 is configured to perform a segmentation of the digitized pathology imaging data 102 to identify one or more tumor regions 109.The second segmentation stage 110 is configured to perform a segmentation of the one or more tumor regions 109 to identify the plurality of cancer nuclei 111 within the one or more tumor regions 109.

[0022] A feature extraction tool 112 is configured to extract a plurality of nuclei diversity features 114 from the plurality of cancer nuclei 111. The plurality of nuclei diversity features 114 comprise features that describe a shape, a shape diversity, and / or a spatial arrangement of the plurality of cancer nuclei 111. In some embodiments, the plurality of nuclei diversity features 114 may comprise spatial features extracted from the plurality of cancer nuclei 111 and / or second order features generated using spatial features extracted from the plurality of cancer nuclei 111. The second order features describe one or more relationships between the spatial features extracted from the plurality of cancer nuclei 111.

[0023] The plurality of nuclei diversity features 114 are provided to a machine learning model 116 that has been trained to generate a medical prognosis 118 for the cancer patient. In some embodiments, the medical prognosis 118 may relate to a survival e.g., an overall survival, a disease free survival, etc.), a treatment response, and / or the like. In some embodiments, the medical prognosis 118 may relate to BCR. For example, the medical prognosis 118 may correspond to a determination as to whether or not the cancer patient will experience biochemical recurrence (BCR) free survival and / or a timeframe of BCR. In some embodiments, the medical prognosis 118 may categorize the cancer patient as having a low-risk 120 of BCR or a high-risk 122 of BCR.

[0024] The medical prognosis 118 may be used by health care professionals to determine a course of treatment for the cancer patient. For example, it has been appreciated that BCR serves as a surrogate endpoint for prostate cancer and is linked to a hazard ratio (HR) of 4.3212 for disease-specific death. While adjuvant therapy is effective in reducing metastasis and disease-specific death, it is not universally suitable due to the low overall mortality rate of prostate cancer. Therefore, by providing an accurate risk estimate for BCR post-surgery, the medical prognosis 1 18 can assist in identifying cancer patients that have a high likelihood of BCR and that may benefit from adjuvant therapy {e.g., like radiation, docetaxel, and / or the like), while avoiding unnecessary treatment for patients that are at low-risk of BCR.

[0025] Fig. 2A illustrates some embodiments of a block diagram 200 of a feature extraction tool 1 12 configured to extract nuclei diversity features from digitized pathology imaging data.

[0026] The feature extraction tool 112 comprises a shape feature generator 202 configured to generate one or more shape features 204 from a plurality of cancer nuclei within the one or more tumor regions of digitized pathology imaging data. In some embodiments, the shape feature generator 202 is configured to extract the one or more shape features 204 from cell graphs 202a (e.g., cell graphs, cell sub-graphs) generated using the plurality of cancer nuclei and / or from cell clusters 202b generated from the cell graphs 202a. In some embodiments, the one or more shape features 204 may comprise one or more of a minor axis length 206 e.g., a minor axis length of a cancer nuclei), an eccentricity 208 (e.g., an eccentricity of a cancer nuclei), an equivalent diameter 210 (e.g., an equivalent diameter of a cancer nuclei), a solidity 212 (e.g., a solidity of a cancer nuclei), and a circularity 214 (e.g., a circularity of a cancer nuclei). In some embodiments, the one or more shape features 204 may comprise and / or be the minor axis length 206, the eccentricity 208, the equivalent diameter 210, the solidity 212, and the circularity 214.

[0027] The feature extraction tool 112 further comprises a second order feature generator 216 configured to generate a plurality of second order features 218 from the one or more shape features 204. In some embodiments, the second order feature generator 216 may be configured to generate a co-occurrence matrix 216a using the one or more shape features 204 and to generate the plurality of second order features 218 from the co-occurrence matrix 216a. For example, as shown in Fig. 2B, the shape features may be used to form elements in a co-occurrence matrix 216a (e.g., a first matrix element SFn may be generated from one or more of the shape features 204, a second matrix element SF12 may be generated from one or more of the shape features 204, etc.). A processor 224 may be configured to apply one or more formulas upon the co-occurrence matrix 216a to generate the plurality of second order features 218. In some embodiment, the co-occurrence matrix 216a may be operated upon by one or more formulas associated with texture features (e.g., entropy, an energy, a correlation, a homogeneity, an inertia, and / or the like) to generate the plurality of second order features 218. In some embodiments, one or more of the second order features may be imperceptible to the human eye. Forexample, the second order features may comprise Haralick features that are based on a joint probability distribution of spatial features that are imperceptible to the human eye.

[0028] The feature extraction tool 112 further comprises a statistical measurement element 220 configured to generate statistical measures 221 of the plurality of shape features 204 and / or the plurality of second order features 218. In various embodiments, the statistical measures 221 may comprise and / or be one or more of a mean, a median, a skewness, a kurtosis, and / or the like.

[0029] Fig. 2C illustrates some embodiments of a block diagram of a BCR assessment system 226 configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis relating to BCR.

[0030] The BCR assessment system 226 comprises a memory configured to store digitized pathology imaging data 102. The digitized pathology imaging data 102 may include a whole slide image (WSI) comprising neoplastic and non-neo plastic epithelial tissue. The BCR assessment system 226 is configured to segment the WSI to identify tumor tissue and to further identify and classify a plurality of cancer nuclei within the tumor tissue. In some embodiments, the WSI may be of a tissue sample of prostate tissue obtained by a radical prostatectomy procedure.

[0031] The BCR assessment system 226 further comprises a feature extraction tool 112 configured to extract a plurality of nuclei diversity features from the digitized pathology imaging data 102. The feature extraction tool 112 is configured to form nuclei subgraphs and / or cell clusters from the plurality of cancer nuclei. One or more shape features may be extracted from the nuclei subgraphs and / or cell clusters. A co-occurrence matrix may be formed using the one or more shape features. A plurality of second order shape features are extracted from the co-occurrence matrix and / or the nuclei subgraphs and / or the cell clusters. The plurality of nuclei diversity features are provided to an input vector 228 as a plurality of concatenated features. The input vector 228 is further provided to a machine learning model 1 16 comprising one or more regression models. In some embodiments, the one or more regression models may comprise a first regression model and a second regression model. The first regression model is configured to select most prognostic features (e.g., that have a most significant impact in determining a risk of death) from the plurality of concatenated features. In some embodiments, the plurality of concatenated featuresmay include a first number of features (e.g., 3,600 features), while the most prognostic features may include a smaller, second number of features (e.g., 11 features). The most prognostic features are then provided as a plurality of nuclei diversity features to the second regression model, which is configured to generate a medical prognosis.

[0032] Fig. 3 illustrates some additional embodiments of a block diagram of an assessment system 300 configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0033] The assessment system 300 comprises a memory 101 configured to store digitized pathology imaging data 102. In some embodiments, the digitized pathology imaging data 102 comprises one or more digitized pathology images 104 (e.g., one or more digitized biopsy slides) obtained from a pathological tissue sample taken from the prostate of a cancer patient 302 removed during a radical prostatectomy. In some embodiments, the one or more digitized pathology images 104 may comprise one or more whole slide images (WSIs), patches of a WSI, or the like. In some embodiments, the memory 101 may comprise electronic memory (e.g., solid state memory, SRAM (static random-access memory), DRAM (dynamic random-access memory), and / or the like).

[0034] In some embodiments, the one or more digitized pathology images 104 may be generated by an image generation stage 303 that is configured to digitize a stained slide generated from a prostate tissue sample taken from the cancer patient 302. In some embodiments, the image generation stage 303 may comprise a tissue resection tool 304 (e.g., a scalpel, a needle, scissors, a punch biopsy, and / or the like) that is used to surgically excise prostate tissue from the cancer patient 302. The tissue is provided to a tissue sectioning and staining tool 306, which is configured to slice the tissue into thin slices that are placed on one or more transparent slides (e.g., one or more glass slides). The tissue on the one or more transparent slides is then stained to generate one or more tissue slides. The one or more tissue slides are subsequently converted to the one or more digitized pathology images 104 by a slide digitization tool 308 (e.g., comprising a CMOS image sensor, a CCD camera, and / or the like). In some embodiments, the one or more digitized pathology images 104 may comprise a whole slide image of a H&E (Hematoxylin and Eosin) stained slide.

[0035] In some embodiments, a pre-processing stage 310 may be configured to operate upon the one or more digitized pathology images 104. The pre-processing stage 310 may include a normalization tool 312 configured to normalize image characteristics (e.g., color, brightness, contrast, efc.) so as to mitigate batch effects (e.g., differences between images obtained from different sites). The pre-processing stage 310 may further include a patch generator 314 configured to break a WSI into a plurality of non-overlapping patches that cover the WSI. In some embodiments, the plurality of patches may be subsequently stored as part of the digitized pathology imaging data 102.

[0036] In some embodiments, a segmentation tool 106 is configured to access the digitized pathology imaging data 102. The segmentation tool 106 is further configured to identify a plurality of cancer nuclei 111 within the digitized pathology imaging data 102. In some embodiments, the segmentation tool 106 comprises a first segmentation stage 108 configured to segment the one or more digitized pathology images 104 to identify one or more tumor regions 109 and a second segmentation stage 110 configured to segment the one or more tumor regions 109 to identify the plurality of cancer nuclei 11 1.

[0037] In some embodiments, the first segmentation stage 108 and the second segmentation stage 110 may comprise machine learning models. For example, the first segmentation stage 108 may comprise a first machine learning model that has been trained to perform tumor segmentation. In some embodiments, the first machine learning model may comprise a U-net model. The second segmentation stage 110 may comprise a second machine learning model that has been trained to perform both nuclei segmentation and classification (e.g., classification of identified nuclei as cancerous or non-cancerous). In some embodiments, the second machine learning model may comprise a HoVer-Net model. In some embodiments, the first machine learning model and the second segmentation stage may be implemented as computer code run on one or more processors (e.g., a central processing unit including one or more transistor devices configured to operate computer code to achieve a result, a microcontroller, or the like).

[0038] A feature extraction tool 112 is configured to extract a plurality of nuclei diversity features 114 from the plurality of cancer nuclei 111. In some embodiments, the plurality of nuclei diversity features 114 comprise shape features 204, secondorder features 218, and / or statistical measures 221 of the second order features 218. The plurality of nuclei diversity features 114 may be extracted by generating one or more shape features 204 from the plurality of cancer nuclei 111 , generating a plurality of second order features 218 using the one or more shape features 204 {e.g., using second order statistical analysis), and taking statistical measures 221 of the plurality of second order features 218. In some embodiments, the feature extraction tool 1 12 may be implemented as computer code run by a processing unit {e.g., a central processing unit including one or more transistor devices configured to operate computer code to achieve a result, a microcontroller, or the like).

[0039] In some embodiments, the one or more shape features 204 may be extracted from cell graphs / subgraphs and / or cell clusters formed using the plurality of cancer nuclei 111. In some embodiments, a co-occurrence matrix may be formed using the one or more shape features 204. This is in contrast to a gray level cooccurrence matrix, which uses image intensity values to form the gray level cooccurrence matrix. The plurality of nuclei diversity features 114 are generated from the one or more shape features 204 and / or the co-occurrence matrix. It has been appreciated that the formation of the co-occurrence matrix using the one or more shape features 204 allows for the nuclei diversity features 114 to have a high prognostic ability {e.g., in predicting a BCR of the cancer patient 302).

[0040] In some embodiments, the nuclei diversity features 114 may comprise or be nuclei shape features and / or Haralick features that include statistical features extracted from a co-occurrence matrix generated using the nuclei shape features. In some additional embodiments, the nuclei diversity features 114 may include nuclei shape features related to nuclei shape diversity {e.g., Minor Axis Length, Eccentricity, Equivalent Diameter, Solidity, and Circularity) in conjunction with Haralick features {e.g., Joint Average, Inverse Difference, Difference Variance, Cluster Shade, Sum Entropy, Sum Variance, Cluster Tendency, Differentiated Entropy, and Inverse Variance). In some embodiments, the nuclei diversity features 114 may comprise eleven features including five related to nuclei shape diversity (Minor Axis Length, Eccentricity, Equivalent Diameter, Solidity, and Circularity) in conjunction with nine Haralick features (Joint Average, Inverse Difference, Difference Variance, Cluster Shade, Sum Entropy, Sum Variance, Cluster Tendency, Differentiated Entropy, and Inverse Variance.

[0041] The plurality of nuclei diversity features 114 are provided to a machine learning model 116 that has been trained to generate a medical prognosis 118. In some embodiments, the machine learning model 116 may be configured to perform survival analysis to generate a risk score 316 (e.g., corresponding to a risk of BCR for the cancer patient 302). The machine learning model 116 1s further configured to compare the risk score 316 to a threshold 318 to identify the cancer patient 302 as being a low-risk 120 for BCR or as being a high-risk 122 for BCR.

[0042] The medical prognosis 118 can be utilized by health care professionals to make a more informed decision relating to the treatment of the cancer patient 302, thereby allowing for the cancer patient 302 to have an improved quality of life (e.g., a lower risk of death, to avoid negative side effects of drugs that are not likely to produce positive outcomes, and / or the like).

[0043] Fig. 4 illustrates some additional embodiments of a block diagram of an assessment system 400 configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0044] The assessment system 400 comprises a feature extraction tool 112 configured to extract a plurality of nuclei diversity features 114 from digitized pathology imaging data. In some embodiments, the digitized pathology imaging data is generated from prostate tissue excised from a cancer patient that has undergone a radical prostatectomy. A machine learning model 116 is configured to utilize the plurality of nuclei diversity features 114 to generate a medical prognosis 1 18.

[0045] In some embodiments, the machine learning model 116 is configured to use the plurality of nuclei diversity features 114 in conjunction with a genomic risk classification, 404a or 404b, to generate the medical prognosis 118. It has been appreciated that the medical prognosis 118 relating to BCR may be able to accurately stratify risk classes generated using genomic risk classifiers. For example, in some embodiments, genomic risk classifiers 402 may be determined for a cancer patient 302 (e.g., using a Decipher genomic test). The genomic risk classifiers 402 may be used to identify the cancer patient 302 as having a genomic risk classification 404. In various embodiments, the genomic risk classification 404 may include a high-risk genomic group 404a or a low-risk genomic group 404b. The low-risk genomic group 404b may be further stratified using the medical prognosis 118 (e.g., to identify a patient within the low-risk genomic group 404b as high-risk orlow-risk). In some embodiments, the assessment system 400 may be able to stratify the low-risk genomic group 404b into low-risk and high-risk categories over a first data set with a hazard ratio of 4.52 achieved with a 95% confidence interval and a p- value of 0.0085. In some embodiments, the hazard ratios achieved by a first data set may vary between approximately 1.70 and approximately 12.05. Further stratification of the genomic groups (e.g., the low-risk genomic group) allows for a health care professional to determine how aggressive a patient’s prostate cancer is and to tailor treatment to an appropriate level.

[0046] Fig. 5 illustrates a flow diagram showing some embodiments of a method 500 of using nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0047] While the disclosed method 500 is illustrated and described herein as a series of acts or events, it will be appreciated that the illustrated ordering of such acts or events are not to be interpreted in a limiting sense. For example, some acts may occur in different orders and / or concurrently with other acts or events apart from those illustrated and / or described herein. In addition, not all illustrated acts may be required to implement one or more aspects or embodiments of the description herein. Further, one or more of the acts depicted herein may be carried out in one or more separate acts and / or phases.

[0048] At act 502, digitized pathology imaging data comprising one or more digitized pathology images from a cancer patient is accessed. In some embodiments, the cancer patient may have undergone a radical prostatectomy as treatment for prostate cancer.

[0049] At act 504, the digitized pathology imaging data is segmented to identify a plurality of cancer nuclei. In some embodiments, the digitized pathology imaging data may comprise one or more digitized pathology images that are segmented according to acts 506-508.

[0050] At act 506, a first segmentation process is performed to identify one or more tumor regions within the one or more digitized pathology images.

[0051] At act 508, a second segmentation process is performed to generate segmented images that identify the plurality of cancer nuclei within the one or more tumor regions.

[0052] At act 510, the segmented images may be stored in electronic memory as part of the digitized pathology imaging data, in some embodiments.

[0053] At act 512, a plurality of nuclei diversity features are extracted from the digitized pathology imaging data. The plurality of nuclei diversity features may comprise a combination of shape features, second order shape features, and statistical measures of the second order shape features. In some embodiments, the plurality of nuclei diversity features may be extracted according to acts 514-524.

[0054] At act 514, cell graphs and / or subgraphs are generated using the plurality of cancer nuclei.

[0055] At act 516, cell clusters are generated using the cell graphs and / or subgraphs.

[0056] At act 518, a plurality of shape features are generated from the graphs, subgraphs, and / or cell clusters.

[0057] At act 520, a co-occurrence matrix is formed using the plurality of shape features.

[0058] At act 522, a plurality of second order features are extracted from the cooccurrence matrix. The plurality of second order features extracted from the cooccurrence matrix may include texture features (e.g., features that are extracted from the co-occurrence matrix using algorithms and / or formulas used to extract texture features from a gray level co-occurrence matrix).

[0059] At act 524, statistical measures of the plurality of second order features are generated to form the plurality of nuclei diversity features.

[0060] At act 526, a machine learning model is operated on the plurality of nuclei diversity features to generate a medical prognosis relating to BCR.

[0061] At act 528, a treatment may be provided to the patient based upon the medical prognosis, in some embodiments. For example, based upon the medical prognosis it may be determined that postoperative treatment may be beneficial to a patient and post operative treatment may be applied to the patient.

[0062] Therefore, the disclosed method 500 utilizes nuclei diversity features extracted from digitized imaging data of the cancer patient to generate a medical prognosis.

[0063] It will be appreciated that the disclosed methods and / or block diagrams may be implemented as computer-executable instructions, in some embodiments.Thus, in one example, a computer-readable storage device (e.g., a non-transitory computer-readable medium) may store computer executable instructions that if executed by a machine (e.g., computer, processor) cause the machine to perform the disclosed methods and / or block diagrams. While executable instructions associated with the disclosed methods and / or block diagrams are described as being stored on a computer-readable storage device, it is to be appreciated that executable instructions associated with other example disclosed methods and / or block diagrams described or claimed herein may also be stored on a computer-readable storage device.

[0064] Fig. 6 illustrates some additional embodiments of a block diagram of an assessment system 600 configured to utilize nuclei diversity features extracted from digitized pathology imaging data to make a medical prognosis.

[0065] The assessment system 600 comprises a memory 101 configured to store digitized pathology imaging data 102 including a plurality of digitized pathology images 104 from cancer patients. In some embodiments, the cancer patients have undergone radical prostatectomies as treatment for prostate cancer. In various embodiments, the plurality of digitized pathology images 104 may be obtained by an image generation stage 303 and / or from an on-line database 602 and / or archive containing digitized pathology images from cancer patients generated at different sites (e.g., different hospitals, research laboratories, and / or the like). Prior to including digitized pathology images within the digitized pathology imaging data 102, the digitized pathology images may be subjected to a pre-processing stage 310.

[0066] The digitized pathology imaging data 102 may include a training set 104t and a validation set 104v. The training set 104t comprises digitized pathology images from a first plurality of cancer patients. The validation set 104v comprises digitized pathology images from a second plurality of cancer patients. In some embodiments, the memory 101 may also be configured to store ground truth segmentation data (e.g., segmentation results provided by an expert human pathologist).

[0067] The training set 104t may be used to train a downstream segmentation tool 106 to perform segmentations that identify a plurality of cancer nuclei 111. The training set 104t may also be used to train a downstream machine learning model 116 to generate a medical prognosis 118.

[0068] The validation set 104v may be used to validate the results of the segmentation tool 106 to perform segmentations that identify a plurality of cancer nuclei 111. The validation set 104v may also be used to validate the results of the machine learning model 116 to generate the medical prognosis 1 18.

[0069] In some embodiments, machine learning model 116 may include a feature selection element 604 configured to select a set of most prognostic nuclei diversity features to generate the medical prognosis 118. For example, the features extraction tool 1 12 may extract a first number of nuclei diversity features and then select a smaller second number of the nuclei diversity features that are most prognostic (e.g., that have a most significant impact in determining a risk of death from BCR). In some embodiments, the second number of nuclei diversity features may be used to train and validate the machine learning model 116. In some embodiments, the machine learning model 116 may comprise a Cox regression model (e.g., a Cox proportional hazards model). In some embodiments, the Cox regression model may comprise a LASSO (least absolute shrinkage and selection operator) algorithm (e.g., a LASSO Cox regression model) that is configured to operate as the feature selection element 604.

[0070] In some embodiments, the feature selection element 604 may select the most prognostic nuclei diversity features to include or be nuclei shape features and Haralick features. In some additional embodiments, the feature selection element 604 may select the most prognostic nuclei diversity features to include or be nuclei shape features related to nuclei shape diversity (e.g., Minor Axis Length, Eccentricity, Equivalent Diameter, Solidity, and Circularity) in conjunction with Haralick features (e.g., Joint Average, Inverse Difference, Difference Variance, Cluster Shade, Sum Entropy, Sum Variance, Cluster Tendency, Differentiated Entropy, and Inverse Variance). In some embodiments, the feature selection element 604 may select the most prognostic nuclei diversity features to be eleven features including five related to nuclei shape diversity (Minor Axis Length, Eccentricity, Equivalent Diameter, Solidity, and Circularity) in conjunction with nine Haralick features (Joint Average, Inverse Difference, Difference Variance, Cluster Shade, Sum Entropy, Sum Variance, Cluster Tendency, Differentiated Entropy, and Inverse Variance).

[0071] It has been appreciated that the disclosed assessment system 600 may be trained to provide a stand-alone stratification of patient risk {e.g., a stratification of patient risk based solely on the output of the BCR assessment system) as well as a complimentary stratification of patient risk {e.g., a stratification of patient risk within a categorization of patient risk achieved by another metric, such as Genomic risk classifiers). Both the stand-alone stratification and the complementary stratification have been shown through test data to have good prognostic ability, as illustrated in the exemplary performance metrics of Figs. 7A-7B.

[0072] Fig. 7A illustrates an exemplary graph 700 showing performance metrics of the disclosed assessment system configured to generate a medical prognosis relating to BCR.

[0073] Graph 700 shows lines indicative of a survival probability (y-axis) as a function of time (x-axis) for high-risk and low-risk patents generated using nuclei diversity features extracted from images within a first data set. As shown in graph 700, a first line is indicative of low-risk patients 702 and a second line is indicative of high-risk patients 704. The low-risk patients 702 have a higher survival probability over time than the high-risk patients 704. In some embodiments, the nuclei diversity features within the first data set achieved a hazard ratio of 6.82 with a 95% confidence interval and a p-value of 0.0019. In some embodiments, the hazard ratios achieved by the second data set varied between 3.54 and 13.12.

[0074] Fig. 7B illustrates an exemplary graph 706 showing performance metrics of the disclosed BCR assessment system operating upon a low-risk category of patients identified using genomic risk classifiers.

[0075] Graph 706 shows lines indicative of a survival probability (y-axis) as a function of time (x-axis) for high-risk and low-risk patents generated using nuclei diversity features extracted from images within a second data set. As shown in graph 706, a first line is indicative of low-risk patients 708 and a second line is indicative of high-risk patients 710. The low-risk patients 708 have a higher survival probability over time than the high-risk patients 710, thereby showing that the disclosed method and apparatus can be used in conjunction with genomic risk classifiers to provide for an improved classification of patients that are at risk of BCR after radical prostatectomies. In some embodiments, the nuclei diversity features achieved a hazard ratio of 4.52 with a 95% confidence interval and a p-value of0.0085. In some embodiments, the hazard ratios achieved by the second data set varied between 1.7 and 12.05.

[0076] Therefore, the graphs illustrated in Figs. 7A-7B show that the disclosed BCR assessment system is able to generate an accurate medical prognosis of BCR. The ability of the disclosed BCR assessment system to generate an accurate medical prognosis of BCR improves a computer’s ability to analyze medical images in a manner that accurately identifies cancer patients that are likely to experience BCR. The improved ability to accurately identify cancer patients that are likely to experience BCR can improve treatment of the cancer patients.

[0077] Fig. 8 illustrates some embodiments of a block diagram of an apparatus 800 configured to generate a medical prognosis by operating a machine learning circuit on nuclei diversity features extracted from digitized pathology imaging data.

[0078] The apparatus 800 comprises an assessment apparatus 802. The assessment apparatus 802 is coupled to an image generation stage 303, which is configured to generate digitized pathology imaging data (e.g., a digitized pathology image) of tissue samples collected from a cancer patient 302 e.g., that has or that has had prostate cancer).

[0079] The assessment apparatus 802 comprises a processor 806 and a memory 804. The processor 806 can, in various embodiments, comprise circuitry such as, but not limited to, one or more single-core or multi-core processors. The processor 806 can include any combination of general-purpose processors and dedicated processors (e.g., graphics processors, application processors, etc.). The processor(s) 806 can be coupled with and / or can comprise memory (e.g., memory 804) or storage and can be configured to execute instructions stored in the memory 804 or storage to enable various apparatus, applications, or operating systems to perform operations and / or methods discussed herein.

[0080] The memory 804 can be configured to store digitized pathology imaging data 102 comprising digitized pathology images. The digitized pathology images may comprise digitized biopsy images having a plurality of pixels, each pixel having an associated intensity. In some additional embodiments, the digitized pathology images may be stored in the memory 804 as one or more training sets of digitized images for training a classifier and / or one or more test sets (e.g., validation sets) of digitized images.

[0081] The assessment apparatus 802 also comprises an input / output (I / O) interface 808 (e.g., associated with one or more I / O devices), a display 810, and an interface 812 that connects the processor 806, the memory 804, and the I / O interface 808. The I / O interface 812 can be configured to transfer data between the memory 804, the processor 806, and external devices, for example, the image generation stage 303.

[0082] In some embodiments, the assessment apparatus 802 may further comprise one or more circuits 814 that include one or more of a segmentation circuit 816, a nuclei diversity feature extraction circuit 818, and a machine learning circuit 820. In some embodiments, the one or more circuits 814 may operate according to machine learning algorithms stored in the memory 804.

[0083] In some embodiments, the segmentation circuit 816 is configured to segment the plurality of digitized pathology images to generate segmented images that identify cancer nuclei within the digitized pathology images. The nuclei diversity feature extraction circuit 818 is configured to extract a plurality of nuclei diversity features 114 using the cancer nuclei. In some embodiments, the nuclei diversity feature extraction circuit 818 may be configured to generate one or more shape features 204 from the cancer nuclei, to generate plurality of a second order features 218 using the one or more shape features 204, and to take statistical measures 221 of the plurality of second order features 218. The plurality of nuclei diversity features 114 may comprise the one or more shape features 204, the second order features 218, and / or the statistical measures 221 . The machine learning circuit 820 is configured to utilize the plurality of nuclei diversity features 114 to generate a medical prognosis 118 for the cancer patient 302. In some embodiments, the display 810 is configured to output or display the medical prognosis 118 generated by the assessment apparatus 802.Example use case:

[0084] Current tools for predicting prostate cancer biochemical recurrence (BCR) after RP often depend on parameters determined by pathologists, such as tumor grade, which is known to vary between reviewers. Genomic risk classifiers provide useful information but require substantial tissue, are costly, and are not commonly accessible in many medical centers. The study presents an artificialintelligence (Al) model for automated risk assessment of biochemical recurrence (BCR) by analyzing nuclear morphologic patterns from Pea (prostate cancer) archival H&E slides. Nuclear morphometric features are essential for disease diagnosis and gaining insights into cellular functions. Specifically, we sought to evaluate whether the computational pathology approach could provide additional granular risk stratification within the low-risk group identified by Decipher.

[0085] The study employed two cohorts from the University of Pennsylvania D1 and D2. D1 comprised 180 patients, while D2 comprised 175 patients, including 63 individuals from the Decipher low-risk group. None of the patients in either cohort had received any pre-operative treatments. Within the Decipher cohort, a tumor segmentation model was created using the U-net deep learning network. The HoVer- Net model was used to both segment and classify nuclei. The feature extracted included a combination of nuclear texture and quantitative measurements of the spatial arrangement of nuclei. The eleven most significant features, selected via the least absolute shrinkage and selection operator (LASSO), were utilized to build a Cox regression model for predicting the risk of BCR in D1 . In D2, patients were classified as either low-risk or high-risk for BCR based on the median risk score derived from D1 .

[0086] Patients from D1 identified as high-risk by the Al model experienced shorter survival, with a median BCR time of 27 months, in contrast to 55 months for those classified as low-risk. The Al model demonstrated significant prognostic value for BCR in both D1 (HR = 6.82, 95% confidence interval [Cl]: 3.54-13.12, p < 0.0019), and D2 (HR = 2.72, 95% Cl: 1.53-4.82, p < 0.0032). Furthermore, the model effectively stratified the Decipher low-risk group in D2 into distinct low and high-risk categories (HR = 4.52, 95% confidence interval [Cl]: 1 .70-12.05, p < 0.0085). The most prognostic features identified included five related to nuclei shape diversity (Minor Axis Length, Eccentricity, Equivalent Diameter, Solidity, and Circularity) in conjunction with nine nuclear texture features.

[0087] The Al model's findings suggest that a digital image-based prognostic classifier for prostate cancer could serve as an alternative or complementary method to molecular-based companion risk tests. Additional, independent multi-site and prospective validation of these findings are warranted.

[0088] Therefore, the present disclosure relates to an assessment system that is configured to generate a medical prognosis by operating a machine learning model on a plurality of nuclei diversity features that describe a shape, shape diversity, and / or a spatial arrangement of the plurality of cancer nuclei.

[0089] In some embodiments, the present disclosure relates to a method including accessing digitized pathology imaging data from a cancer patient, the digitized pathology imaging data being segmented to identify a plurality of cancer nuclei within one or more tumor regions; extracting a plurality of nuclei diversity features using the plurality of cancer nuclei, the plurality of nuclei diversity features including a combination of shape features and second order features generated using the shape features; and providing the plurality of nuclei diversity features to a machine learning model that is trained to generate a medical prognosis regarding the cancer patient. In some embodiments, the method further includes accessing the digitized pathology imaging data from the cancer patient; performing a first segmentation process on the digitized pathology imaging data to identify the one or more tumor regions; and subsequently performing a second segmentation process on the one or more tumor regions to identify the plurality of cancer nuclei within the one or more tumor regions. In some embodiments, the method further includes generating the shape features using the plurality of cancer nuclei; generating a plurality of second order features using the shape features; and generating one or more statistical measures of the plurality of second order features, the plurality of nuclei diversity features further including the statistical measures of the plurality of second order features. In some embodiments, the shape features include one or more of a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity. In some embodiments, the plurality of nuclei diversity features include the shape features and Haralick features extracted from a co-occurrence matrix generated using the shape features. In some embodiments, the method further includes generating the shape features using the plurality of cancer nuclei; generating a co-occurrence matrix using the shape features; and generating the second order features from the cooccurrence matrix. In some embodiments, the plurality of nuclei diversity features include a minor axis length, an eccentricity, an equivalent diameter, a solidity, a circularity, a joint average, an inverse difference, a difference variance, a clustershade, a sum entropy, a sum variance, a cluster tendency, a differentiated entropy, and an inverse variance. In some embodiments, the method further includes identifying genomic risk classifiers for the cancer patient; using the genomic risk classifiers to identify the cancer patient as being within a low-risk genomic group; and the medical prognosis further stratifies the low-risk genomic group. In some embodiments, the medical prognosis is a risk for biochemical recurrence.

[0090] In other embodiments, the present disclosure relates to a non-transitory computer-readable medium storing computer-executable instructions that, when executed, cause a processor to perform operations including accessing digitized pathology imaging data of prostate tissue excised from a cancer patient, the digitized pathology imaging data being segmented to identify a plurality of cancer nuclei; generating one or more shape features from the plurality of cancer nuclei; generating a plurality of second order features using the one or more shape features; generating statistical measures of the plurality of second order features; and providing a plurality of nuclei diversity features to a machine learning model that is trained to generate a medical prognosis, the plurality of nuclei diversity features including the statistical measures. In some embodiments, the plurality of nuclei diversity features further include the second order features and the one or more shape features. In some embodiments, the operations further include generating cell graphs and cell clusters from the plurality of cancer nuclei; and extracting the one or more shape features from the cell graphs and / or the cell clusters. In some embodiments, the one or more shape features include one or more of a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity. In some embodiments, the operations further include generating a co-occurrence matrix using the one or more shape features; and generating the second order features using the co-occurrence matrix. In some embodiments, the operations further include identifying genomic risk classifiers for the cancer patient; using the genomic risk classifiers to identify the cancer patient as being within a low-risk genomic group; and the medical prognosis further stratifying the low-risk genomic group.

[0091] In yet other embodiments, the present disclosure relates to an apparatus including a memory configured to store digitized pathology imaging data from a prostate cancer patient, the digitized pathology imaging data being segmented to identify a plurality of cancer nuclei; a feature extraction tool configured to extract aplurality of nuclei diversity features from the digitized pathology imaging data by: generating one or more shape features from the plurality of cancer nuclei; generating a plurality of second order features using the one or more shape features; generating statistical measures of the plurality of second order features; and a machine learning model configured to utilize the plurality of nuclei diversity features to generate a medical prognosis. In some embodiments, the one or more shape features include a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity. In some embodiments, the feature extraction tool is further configured to generate a co-occurrence matrix using the one or more shape features; and generate the plurality of second order features from the co-occurrence matrix. In some embodiments, the machine learning model is configured to use the plurality of nuclei diversity features in conjunction with a genomic risk classification to generate the medical prognosis. In some embodiments, the medical prognosis stratifies a low-risk genomic group.

[0092] It will be appreciated that the disclosed methods and / or block diagrams may be implemented as computer-executable instructions, in some embodiments. Thus, in one example, a computer-readable storage device (e.g., a non-transitory computer-readable medium) may store computer executable instructions that if executed by a machine (e.g., computer, processor) cause the machine to perform the disclosed methods and / or block diagrams. While executable instructions associated with the disclosed methods and / or block diagrams are described as being stored on a computer-readable storage device, it is to be appreciated that executable instructions associated with other example disclosed methods and / or block diagrams described or claimed herein may also be stored on a computer-readable storage device.

[0093] Examples herein can include subject matter such as an apparatus, including a digital whole slide scanner, a CT system, an MRI system, a personalized medicine system, a CADx system, a processor, a system, circuitry, a method, means for performing acts, steps, or blocks of the method, at least one machine-readable medium including executable instructions that, when performed by a machine (e.g., a processor with memory, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or the like) cause the machine to perform acts ofthe method or of an apparatus or system, according to embodiments and examples described.

[0094] References to “one embodiment”, “an embodiment”, “one example”, and “an example” indicate that the embodiment(s) or example(s) so described may include a particular feature, structure, characteristic, property, element, or limitation, but that not every embodiment or example necessarily includes that particular feature, structure, characteristic, property, element or limitation. Furthermore, repeated use of the phrase “in one embodiment” does not necessarily refer to the same embodiment, though it may.

[0095] “Computer-readable storage device”, as used herein, refers to a device that stores instructions or data. “Computer-readable storage device” does not refer to propagated signals. A computer-readable storage device may take forms, including, but not limited to, non-volatile media, and volatile media. Non-volatile media may include, for example, optical disks, magnetic disks, tapes, and other media. Volatile media may include, for example, semiconductor memories, dynamic memory, and other media. Common forms of a computer-readable storage device may include, but are not limited to, a floppy disk, a flexible disk, a hard disk, a magnetic tape, other magnetic medium, an application specific integrated circuit (ASIC), a compact disk (CD), other optical medium, a random access memory (RAM), a read only memory (ROM), a memory chip or card, a memory stick, and other media from which a computer, a processor or other electronic device can read.

[0096] “Circuit”, as used herein, includes but is not limited to hardware, firmware, software in execution on a machine, or combinations of each to perform a function(s) or an action(s), or to cause a function or action from another logic, method, or system. A circuit may include a software controlled microprocessor, a discrete logic {e.g., ASIC), an analog circuit, a digital circuit, a programmed logic device, a memory device containing instructions, and other physical devices. A circuit may include one or more gates, combinations of gates, or other circuit components. Where multiple logical circuits are described, it may be possible to incorporate the multiple logical circuits into one physical circuit. Similarly, where a single logical circuit is described, it may be possible to distribute that single logical circuit between multiple physical circuits.

[0097] To the extent that the term “includes” or “including” is employed in the detailed description or the claims, it is intended to be inclusive in a manner similar to the term “comprising” as that term is interpreted when employed as a transitional word in a claim.

[0098] Throughout this specification and the claims that follow, unless the context requires otherwise, the words 'comprise' and 'include' and variations such as 'comprising' and 'including' will be understood to be terms of inclusion and not exclusion. For example, when such terms are used to refer to a stated integer or group of integers, such terms do not imply the exclusion of any other integer or group of integers.

[0099] To the extent that the term “or” is employed in the detailed description or claims (e.g., A or B) it is intended to mean “A or B or both”. When the applicants intend to indicate “only A or B but not both” then the term “only A or B but not both” will be employed. Thus, use of the term “or” herein is the inclusive, and not the exclusive use. See, Bryan A. Garner, A Dictionary of Modern Legal Usage 624 (2d. Ed. 1995).[000100] While example systems, methods, and other embodiments have been illustrated by describing examples, and while the examples have been described in considerable detail, it is not the intention of the applicants to restrict or in any way limit the scope of the appended claims to such detail. It is, of course, not possible to describe every conceivable combination of components or methodologies for purposes of describing the systems, methods, and other embodiments described herein. Therefore, the invention is not limited to the specific details, the representative apparatus, and illustrative examples shown and described. Thus, this application is intended to embrace alterations, modifications, and variations that fall within the scope of the appended claims.

Claims

What is claimed is:1 . A method, comprising: accessing digitized pathology imaging data from a cancer patient, wherein the digitized pathology imaging data has been segmented to identify a plurality of cancer nuclei within one or more tumor regions; extracting a plurality of nuclei diversity features using the plurality of cancer nuclei, wherein the plurality of nuclei diversity features include a combination of shape features and second order features generated using the shape features; and providing the plurality of nuclei diversity features to a machine learning model that is trained to generate a medical prognosis regarding the cancer patient.

2. The method of claim 1 , further comprising: accessing the digitized pathology imaging data from the cancer patient; performing a first segmentation process on the digitized pathology imaging data to identify the one or more tumor regions; and subsequently performing a second segmentation process on the one or more tumor regions to identify the plurality of cancer nuclei within the one or more tumor regions.

3. The method of claim 1 , further comprising: generating the shape features using the plurality of cancer nuclei; generating a plurality of second order features using the shape features; and generating one or more statistical measures of the plurality of second order features, wherein the plurality of nuclei diversity features further include the statistical measures of the plurality of second order features.

4. The method of claim 3, wherein the shape features comprise one or more of a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity.

5. The method of claim 1 , wherein the plurality of nuclei diversity features comprise the shape features and Haralick features extracted from a co-occurrence matrix generated using the shape features.

6. The method of claim 1 , further comprising: generating the shape features using the plurality of cancer nuclei; generating a co-occurrence matrix using the shape features; and generating the second order features from the co-occurrence matrix.

7. The method of claim 1 , wherein the plurality of nuclei diversity features comprise a minor axis length, an eccentricity, an equivalent diameter, a solidity, a circularity, a joint average, an inverse difference, a difference variance, a cluster shade, a sum entropy, a sum variance, a cluster tendency, a differentiated entropy, and an inverse variance.

8. The method of claim 1 , further comprising: identifying genomic risk classifiers for the cancer patient; using the genomic risk classifiers to identify the cancer patient as being within a low-risk genomic group; and wherein the medical prognosis further stratifies the low-risk genomic group.

9. The method of claim 1 , wherein the medical prognosis is a risk for biochemical recurrence.

10. A non-transitory computer-readable medium storing computerexecutable instructions that, when executed, cause a processor to perform operations, comprising: accessing digitized pathology imaging data of prostate tissue excised from a cancer patient, wherein the digitized pathology imaging data has been segmented to identify a plurality of cancer nuclei; generating one or more shape features from the plurality of cancer nuclei; generating a plurality of second order features using the one or more shape features;generating statistical measures of the plurality of second order features; and providing a plurality of nuclei diversity features to a machine learning model that is trained to generate a medical prognosis, wherein the plurality of nuclei diversity features include the statistical measures.

11. The non-transitory computer-readable medium of claim 10, wherein the plurality of nuclei diversity features further include the second order features and the one or more shape features.

12. The non-transitory computer-readable medium of claim 10, wherein the operations further comprise: generating cell graphs and cell clusters from the plurality of cancer nuclei; and extracting the one or more shape features from the cell graphs and / or the cell clusters.

13. The non-transitory computer-readable medium of claim 10, wherein the one or more shape features comprise one or more of a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity.

14. The non-transitory computer-readable medium of claim 10, wherein the operations further comprise: generating a co-occurrence matrix using the one or more shape features; and generating the second order features using the co-occurrence matrix.

15. The non-transitory computer-readable medium of claim 10, wherein the operations further comprise: identifying genomic risk classifiers for the cancer patient; using the genomic risk classifiers to identify the cancer patient as being within a low-risk genomic group; and wherein the medical prognosis further stratifies the low-risk genomic group.

16. An apparatus, comprising:a memory configured to store digitized pathology imaging data from a prostate cancer patient, wherein the digitized pathology imaging data has been segmented to identify a plurality of cancer nuclei; a feature extraction tool configured to extract a plurality of nuclei diversity features from the digitized pathology imaging data by: generating one or more shape features from the plurality of cancer nuclei; generating a plurality of second order features using the one or more shape features; generating statistical measures of the plurality of second order features; and a machine learning model configured to utilize the plurality of nuclei diversity features to generate a medical prognosis.

17. The apparatus of claim 16, wherein the one or more shape features comprise a minor axis length, an eccentricity, an equivalent diameter, a solidity, and a circularity.

18. The apparatus of claim 16, wherein the feature extraction tool is further configured to: generate a co-occurrence matrix using the one or more shape features; and generate the plurality of second order features from the co-occurrence matrix.

19. The apparatus of claim 16, wherein the machine learning model is configured to use the plurality of nuclei diversity features in conjunction with a genomic risk classification to generate the medical prognosis.

20. The apparatus of claim 19, wherein the medical prognosis stratifies a low-risk genomic group.

Images