Development of a High-Throughput Three-Dimensional H&E Platform for the Characterization of Breast Cancer Biopsies

Visikol, Inc., Hampton NJ

Investigators

Abstract

PROJECT SUMMARY The purpose of this project is to develop a prognostic imaging tool for breast cancer that can be used to predict specific clinical outcomes such as tumor reoccurrence and drug resistance based upon a tumor?s histomorphology features. The current histopathology paradigm for evaluating tissues is focused on diagnosing disease wherein these same tissues if evaluated in their entirety and digitized, provide an opportunity to correlate clinical outcomes to specific tissue features. Through this project, we are developing a high- throughput breast tumor imaging approach that is capable of imaging breast tumor tissue in its entirety and generating virtual H&E optical Z sections in 3D of equivalent quality to traditional H&E sections so that all of the cells within a biopsy are characterized. To achieve this objective, we are combining our patented tissue clearing approach with fluorescent labeling, high-content confocal microscopy and an unbiased machine learning approach. This approach allows for biopsies to be digitized in their entirety and for all of the features and heterogeneity of tumors to be assessed instead of just looking at a few ultra-thin 2D slides. Through the combination of this unique imaging approach with hierarchical agglomerative clustering, specific histomorphological features can be correlated to clinical outcomes using a detailed sample library with corresponding clinical outcome data. The main objectives of this project are to 1) develop a robust 3D H&E labeling approach, 2) demonstrate that tissues can be imaged in 3D using a fluorescent approach to generate ?H&E-like? images of equivalent quality to traditional H&E and 3) show that this tissue analysis approach can be transferred to an automated high- content confocal microscope. Additionally, we will show the ability to cluster tissues based on their histomorphological features and will with a small data set of 34 breast tumor biopsies show how these features are correlated to clinical outcomes. If successful, we will develop this proof-of-concept into a robust CLIA 21 CFR Part 11 compliant assay that complies with the ICH guidelines for analytical assays. This assay would ultimately allow clinicians to better predict how a tumor will respond to certain treatments and best tailor a treatment for a specific patient. This type of precision medicine approach will lead to improved patient outcomes and a more efficacious treatment regimen.