Colorectal cancer survival prediction using deep distribution based multiple-instance learning

TL;DR

This paper introduces DeepDisMISL, a novel deep learning method that leverages distributional information of patch scores in whole slide images to improve colorectal cancer survival prediction, outperforming existing algorithms.

Contribution

The study proposes a new distribution-based multiple-instance learning algorithm for survival prediction, emphasizing holistic distribution information over top instances.

Findings

Distribution of patch scores improves prediction accuracy.

Including neighborhood instances enhances model performance.

DeepDisMISL outperforms state-of-the-art algorithms.

Abstract

Several deep learning algorithms have been developed to predict survival of cancer patients using whole slide images (WSIs).However, identification of image phenotypes within the WSIs that are relevant to patient survival and disease progression is difficult for both clinicians, and deep learning algorithms. Most deep learning based Multiple Instance Learning (MIL) algorithms for survival prediction use either top instances (e.g., maxpooling) or top/bottom instances (e.g., MesoNet) to identify image phenotypes. In this study, we hypothesize that wholistic information of the distribution of the patch scores within a WSI can predict the cancer survival better. We developed a distribution based multiple-instance survival learning algorithm (DeepDisMISL) to validate this hypothesis. We designed and executed experiments using two large international colorectal cancer WSIs datasets - MCO CRC and TCGA COAD-READ. Our results suggest that the more information about the distribution of the patch scores for a WSI, the better is the prediction performance. Including multiple neighborhood instances around each selected distribution location (e.g., percentiles) could further improve the prediction. DeepDisMISL demonstrated superior predictive ability compared to other recently published, state-of-the-art algorithms. Furthermore, our algorithm is interpretable and could assist in understanding the relationship between cancer morphological phenotypes and patients cancer survival risk.