Case Western Reserve University

Cheng Lu

Anant Madabhushi

Satish E Viswanath

Nicholas P Reder

Jonathan Teng-Chieh Liu

Can Koyuncu

Khoi Le

Embodiments include accessing an image of a region of tissue demonstrating cancerous pathology; detecting a plurality of cells represented in the image; segmenting a cellular nucleus of a first member of the plurality of cells and a cellular nucleus of at least one second, different member of the plurality of cells; extracting a set of nuclear morphology features from the plurality of cells; constructing a feature driven local cell graph (FeDeG) based on the set of nuclear morphology features and a spatial relationship between the cellular nuclei using a mean-shift clustering approach; computing a set of FeDeG features based on the FeDeG; providing the FeDeG features to a machine learning classifier; receiving, from the machine learning classifier, a classification of the region of tissue as a long-term or a short-term survivor, based, at least in part, on the set of FeDeG features; and displaying the classification.

Predicting overall survival in early stage lung cancer with feature driven local cell graphs (FeDeG)

Predicting overall survival in early stage lung cancer with feature driven local cell graphs (FEDEG)

Embodiments discussed herein facilitate determining a diagnosis and/or prognosis for prostate cancer based at least in part on three-dimensional (3D) pathomic feature(s). One example embodiment comprises a computer-readable medium storing computer-executable instructions that, when executed, cause a processor to perform operations, comprising: accessing a three-dimensional (3D) optical image volume comprising a prostate gland of a patient; segmenting the prostate gland of the 3D optical image volume; extracting one or more features from the segmented prostate gland, wherein the one or more features comprise at least one 3D pathomic feature; and generating, via a model based at least on the one or more features, one or more of the following based at least on the extracted one or more features: a classification of the prostate gland as one of benign or malignant, a Gleason score associated with the prostate gland, or a prognosis for the patient.

Analysis of prostate glands using three-dimensional (3D) morphology features of prostate from 3D pathology images

Embodiments discussed herein facilitate training deep learning models to generate synthetic versions of histological texture features and employing such deep learning models. One example embodiment is an apparatus configured to convert a stained histological image to grayscale; extract patches from the grayscale image; for each patch of the plurality of patches: provide that patch to a deep learning model trained to generate a synthetic version of a texture feature; and obtain an associated patch from the deep learning model that indicates an associated value of the synthetic version of the histology texture feature for each pixel of that patch; and merge the associated patches for each patch of the plurality of patches to generate an associated feature map for the stained histological image, wherein the associated feature map indicates the associated value of the synthetic version of the histology texture feature for each pixel of the plurality of pixels.

Fast synthetic Haralick texture generation for histology images

Embodiments include apparatus for predicting cancer recurrence based on local co-occurrence of cell morphology (LoCoM). The apparatus includes image acquisition circuitry that identifies and segments at least one cellular nucleus represented in an image of a region of tissue demonstrating cancerous pathology; local nuclei graph (LNG) circuitry that constructs an LNG based on the at least one cellular nucleus, and computes a set of nuclear morphology features for a nucleus represented in the LNG; LoCoM circuitry that constructs a co-occurrence matrix based on the nuclear morphology features, computes a set of LoCoM features for the co-occurrence matrix, and computes a LoCoM signature for the image based on the set of LoCoM features; progression circuitry that generates a probability that the region of tissue will experience cancer progression based on the LoCoM signature, and classifies the region of tissue as a progressor or non-progressor based on the probability.

Predicting cancer recurrence using local co-occurrence of cell morphology (LoCoM)

Embodiments include an image acquisition circuit configured to access an image of a region of tissue demonstrating cancerous pathology, a nuclei detection and graphing circuit configured to detect cellular nuclei represented in the image; and construct a nuclear sub-graph based on the detected cellular nuclei, where a node of the sub-graph is a nuclear centroid of a cellular nucleus; a cell run length (CRF) circuit configured to compute a CRF vector based on the sub-graph; compute a set of CRF features based on the CRF vector and the sub-graph; and generate a CRF signature based, at least in part, on the set of CRF features; and a classification circuit configured to compute a probability that the region of tissue will experience cancer progression, based, at least in part, on the CRF signature; and generate a classification of the region of tissue as a progressor or non-progressor.

Predicting cancer progression using cell run length features

Methods, apparatus, and other embodiments detect nuclei in histopathological images. One example method includes accessing a histopathology image that includes a plurality of pixels, generating a gradient field map based on the histopathology image, generating a refined gradient field map based on the gradient field map, calculating a voting map for a member of the plurality of pixels based, at least in part, on the refined gradient field map and a voting kernel, generating an aggregated voting map based on the voting map, computing a global threshold, and identifying a nuclear centroid based on the global threshold and the aggregated voting map.

Multi-pass adaptive voting for nuclei detection in histopathological images

Methods, apparatus, and other embodiments associated with predicting Crohn's Disease (CD) patient response to immunosuppressive (IS) therapy using radiomic features extracted from diagnostic magnetic resonance enterography (MRE). One example apparatus includes an image acquisition circuit that acquires an MRE image of a region of tissue demonstrating CD pathology, a segmentation circuit that segments a region of interest (ROI) from the diagnostic radiological image, a classification circuit that extracts a set of discriminative features from the ROI and that distinguishes the ROI as a responder or non-responder to IS therapy, and a CD prediction circuit that generates a radiomic enterographic (RET) score based on the diagnostic radiological image or the set of discriminative features. A prognosis or treatment plan may be provided based on the RET score.

Growing community of inventors

Cleveland Heights, OH, United States of America

Cheng Lu