There is an increased need for predictive and prognostic assays to distinguish more and less aggressive phenotypes of cancer due to A) dramatic increase in cancer incidence and; B) improvements in early diagnosis. Predictive assays in particular will allow for patients with less aggressive disease to be spared more aggressive treatment. Most prognostic tests in the US and Europe are based on gene expression assays (e.g. Oncotype DX (ODx)). Recent studies have shown extensive genetic heterogeneity among cancer cells between tumors and even within the same tumor, suggesting that approaches for recommending therapy for a patient based on the ?average? molecular signal of many cells are overly simplistic. Interestingly, for a number of cancers, tumor grade (morphologic appearance on tissue as assessed qualitatively or semi-quantitatively by a pathologist) has been found to be highly correlated with disease outcome. However pathologic grade tends to suffer from significant inter-observer variability. Digitzation of histological samples, or whole slide imaging, facilitates a quantitative approach towards evaluating disease progression and predicting outcome, while also facilitating the adoption of telepathology. Recently, research groups (including our own) have begun to show that computer extracted measurements of tumor morphology (e.g. capturing nuclear orientation, texture, shape, architecture) from routine H&E stained cancer tissue images can predict disease aggressiveness and treatment outcome. By computationally interrogating the entire tumor landscape and its most invasive elements from a standard H&E slide, these approaches can allow for more accurate capture of tumor heterogeneity, disease risk and hence the most appropriate treatment strategy. The goal of this academic-industrial partnership is to develop and validate a computerized histologic image-based predictor (CHIP) to identify which early-stage, estrogen receptor positive (ER+) breast cancer patients are candidates for hormonal therapy alone and which women are candidates for adjuvant chemotherapy based off analysis of the pathology slides derived from biopsy and surgical specimens. Inspirata Inc., a cancer diagnostics company which has recently licensed a number of histomorphometry based technologies from the Madabhushi group, will bring quality management systems and production software standards to help create a pre-commercial companion diagnostic test of the CHIP assay. Additionally Inspirata Inc. will build a complete regulatory pathway for successful translation of the assay in the US and abroad. Finally, the pre-commercial prototype of the CHIP assay will be independently validated using the same strategy and data cohorts as ODx. Our approach has several advantages over molecular assays such as ODx in that it (1) can interrogate the entire expanse of the pathology image enabling a more accurate capture of tumor heterogeneity and hence disease risk, (2) is non-disruptive of pathology workflow, (3) non-destructive of tissue and would be substantially (4) cheaper (critical in low to middle income countries) and (5) faster.

Public Health Relevance

Of the 1 million women worldwide who in 2015 will be diagnosed with estrogen receptor positive (ER+) breast cancer, most will be treated with chemotherapy, though only a small number (< 20%) will benefit from it. Our goal is to create and validate a pre-commercial prototype of a computerized histologic image based predictor (CHIP) for identifying which early stage ER+ breast cancer patients will benefit from adjuvant chemotherapy. CHIP will employ sophisticated computer vision techniques for comprehensive characterization of disease morphology from digitized images of H&E stained specimens yielding a continuous image-based risk score; low CHIP risk score suggesting hormonal therapy is sufficient while adjuvant chemo is required for high CHIP score patients.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project (R01)
Project #
5R01CA202752-02
Application #
9305968
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Ossandon, Miguel
Project Start
2016-07-01
Project End
2021-06-30
Budget Start
2017-07-01
Budget End
2018-06-30
Support Year
2
Fiscal Year
2017
Total Cost
Indirect Cost
Name
Case Western Reserve University
Department
Biomedical Engineering
Type
Biomed Engr/Col Engr/Engr Sta
DUNS #
077758407
City
Cleveland
State
OH
Country
United States
Zip Code
44106
Bera, Kaustav; Velcheti, Vamsidhar; Madabhushi, Anant (2018) Novel Quantitative Imaging for Predicting Response to Therapy: Techniques and Clinical Applications. Am Soc Clin Oncol Educ Book :1008-1018
Algohary, Ahmad; Viswanath, Satish; Shiradkar, Rakesh et al. (2018) Radiomic features on MRI enable risk categorization of prostate cancer patients on active surveillance: Preliminary findings. J Magn Reson Imaging :
Corredor, Germán; Wang, Xiangxue; Zhou, Yu et al. (2018) Spatial architecture and arrangement of tumor-infiltrating lymphocytes for predicting likelihood of recurrence in early-stage non-small cell lung cancer. Clin Cancer Res :
Beig, Niha; Khorrami, Mohammadhadi; Alilou, Mehdi et al. (2018) Perinodular and Intranodular Radiomic Features on Lung CT Images Distinguish Adenocarcinomas from Granulomas. Radiology :180910
Orooji, Mahdi; Alilou, Mehdi; Rakshit, Sagar et al. (2018) Combination of computer extracted shape and texture features enables discrimination of granulomas from adenocarcinoma on chest computed tomography. J Med Imaging (Bellingham) 5:024501
Cruz-Roa, Angel; Gilmore, Hannah; Basavanhally, Ajay et al. (2018) High-throughput adaptive sampling for whole-slide histopathology image analysis (HASHI) via convolutional neural networks: Application to invasive breast cancer detection. PLoS One 13:e0196828
Shiradkar, Rakesh; Ghose, Soumya; Jambor, Ivan et al. (2018) Radiomic features from pretreatment biparametric MRI predict prostate cancer biochemical recurrence: Preliminary findings. J Magn Reson Imaging 48:1626-1636
Nirschl, Jeffrey J; Janowczyk, Andrew; Peyster, Eliot G et al. (2018) A deep-learning classifier identifies patients with clinical heart failure using whole-slide images of H&E tissue. PLoS One 13:e0192726
Peyster, Eliot G; Madabhushi, Anant; Margulies, Kenneth B (2018) Advanced Morphologic Analysis for Diagnosing Allograft Rejection: The Case of Cardiac Transplant Rejection. Transplantation 102:1230-1239
Penzias, Gregory; Singanamalli, Asha; Elliott, Robin et al. (2018) Identifying the morphologic basis for radiomic features in distinguishing different Gleason grades of prostate cancer on MRI: Preliminary findings. PLoS One 13:e0200730

Showing the most recent 10 out of 54 publications