Deep Biological Pathway Informed Pathology-Genomic Multimodal Survival Prediction

TL;DR

This paper introduces PONET, a deep learning model that integrates pathology images and genomic data to improve cancer survival prediction and identify key biological pathways, demonstrating superior performance on TCGA datasets.

Contribution

The paper presents a novel biologically informed deep model that enhances survival prediction and uncovers relevant genes and pathways from multimodal data.

Findings

Achieves superior survival prediction accuracy on TCGA datasets.

Identifies meaningful genes and pathways related to patient survival.

Provides biological insights into disease mechanisms.

Abstract

The integration of multi-modal data, such as pathological images and genomic data, is essential for understanding cancer heterogeneity and complexity for personalized treatments, as well as for enhancing survival predictions. Despite the progress made in integrating pathology and genomic data, most existing methods cannot mine the complex inter-modality relations thoroughly. Additionally, identifying explainable features from these models that govern preclinical discovery and clinical prediction is crucial for cancer diagnosis, prognosis, and therapeutic response studies. We propose PONET- a novel biological pathway-informed pathology-genomic deep model that integrates pathological images and genomic data not only to improve survival prediction but also to identify genes and pathways that cause different survival rates in patients. Empirical results on six of The Cancer Genome Atlas (TCGA) datasets show that our proposed method achieves superior predictive performance and reveals meaningful biological interpretations. The proposed method establishes insight into how to train biologically informed deep networks on multimodal biomedical data which will have general applicability for understanding diseases and predicting response and resistance to treatment.