Abstract 5432: Validation of spatial gene expression patterns predicted by deep convolutional neural networks from breast cancer histopathology images

Abstract

Abstract Molecular profiling is central in cancer precision medicine but remains costly and is only based on tumor average profiles. Morphologic patterns observable in histopathology sections from tumors are determined by the underlying molecular phenotype and therefore have the potential to be exploited for the prediction of molecular phenotypes. Transcriptome-wide expression morphology (EMO) analysis with deep convolutional neural networks (CNN) enables the prediction of mRNA expression and proliferation markers from routine histopathology whole slide images in breast cancer. The NanoString GeoMx® Digital Spatial Profiler (DSP) platform has capabilities of associating crucial spatial information with intratumor variabilities of gene expression, serving as an ideal tool to characterize and validate the prediction of intratumor heterogeneity. The GeoMx platform also provides functionalities of overlaying and aligning hematoxylin and eosin (H&E) whole slide scans on serial tissue sections with morphology marker staining to facilitate region of interest (ROI) selection to match relevant image tiles from the model input, which is a key step to ensure accuracy for the evaluation of the model predictions. The GeoMx RNA Immune Pathways Panel contains 84 gene targets including key genes involved in immune pathways and tumorigenesis and was used to validate the prediction of mRNA expression output from deep CNN models. Citation Format: Kathy Ton, Yinxi Wang, Liuliu Pan, Kimmo Kartasalo, Balazs Acs, Philippe Weitz, Liang Zhang, Yan Liang, Johan Hartman, Masi Volkonen, Christer Larsson, Pekka Ruusuvuori, Joseph Beechem, Mattias Rantalainen. Validation of spatial gene expression patterns predicted by deep convolutional neural networks from breast cancer histopathology images. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5432.