Computational Image Analysis Identifies Histopathological Image Features Associated With Somatic Mutations and Patient Survival in Gastric Adenocarcinoma.

Jun Cheng,Jie Zhang,Wenhui Hong,Kun Huang,Lingling Wang,Yuting Liu,Wei Huang,Zhi Han,Dong Ni,Xiaohui Zhan

doi:10.3389/fonc.2021.623382

Abstract

Computational analysis of histopathological images can identify sub-visual objective image features that may not be visually distinguishable by human eyes, and hence provides better modeling of disease phenotypes. This study aims to investigate whether specific image features are associated with somatic mutations and patient survival in gastric adenocarcinoma (sample size = 310). An automated image analysis pipeline was developed to extract quantitative morphological features from H&E stained whole-slide images. We found that four frequently somatically mutated genes (TP53, ARID1A, OBSCN, and PIK3CA) were significantly associated with tumor morphological changes. A prognostic model built on the image features significantly stratified patients into low-risk and high-risk groups (log-rank test p-value = 2.6e-4). Multivariable Cox regression showed the model predicted risk index was an additional prognostic factor besides tumor grade and stage. Gene ontology enrichment analysis showed that the genes whose expressions mostly correlated with the contributing features in the prognostic model were enriched on biological processes such as cell cycle and muscle contraction. These results demonstrate that histopathological image features can reflect underlying somatic mutations and identify high-risk patients that may benefit from more precise treatment regimens. Both the image features and pipeline are highly interpretable to enable translational applications.

Highlights

Gastric cancer is the world’s third leading cause of cancer mortality following lung and colorectal cancers (1)
With the curation of large-scale genomic data, genes with mutations that are common to gastric adenocarcinoma have been identified
We investigated whether a prognostic model built on selected image features can predict the overall survival of the gastric adenocarcinoma patients

Summary

Introduction

Gastric cancer is the world’s third leading cause of cancer mortality following lung and colorectal cancers (1). Most patients with gastric cancer are in the late stage at the time of diagnosis, and the tumor has often spread to lymph nodes or other organs; the prognosis of gastric cancer is usually poor. Stratification of cancer patients into groups with different tumor genotypes, phenotypes and clinical outcomes is a key process to achieve precision oncological treatment. The most commonly used indicator for assessing the extent of cancer spreading is tumor node metastasis (TNM) staging (8). TNM staging has been widely used on solid tumors for estimating prognosis and planning treatment for patients. Besides TNM staging, there are alternative classification systems for gastric cancers. The effectiveness of these classification systems is limited due to the high heterogeneity of gastric cancers. Extensive efforts have been made to take advantage of high-throughput large-scale molecular profiling data in the hope of discovering better diagnostic (11–13), prognostic (14–16), and predictive (17–20) biomarkers

Objectives

Methods

Results

Conclusion