Attention-based Multiple Instance Learning for Survival Prediction on Lung Cancer Tissue Microarrays

TL;DR

This paper introduces an attention-based multiple instance learning model for survival prediction on lung cancer tissue microarrays, leveraging gigapixel images and Cox partial likelihood to achieve competitive performance with clinical factors.

Contribution

The study extends AMIL to survival analysis using Cox loss, demonstrating its effectiveness on TMA slides and matching traditional clinical factor-based methods.

Findings

AMIL handles small tissue samples effectively.

Model achieves comparable C-index to clinical factor-based methods.

Approach is applicable to gigapixel whole-slide images.

Abstract

Attention-based multiple instance learning (AMIL) algorithms have proven to be successful in utilizing gigapixel whole-slide images (WSIs) for a variety of different computational pathology tasks such as outcome prediction and cancer subtyping problems. We extended an AMIL approach to the task of survival prediction by utilizing the classical Cox partial likelihood as a loss function, converting the AMIL model into a nonlinear proportional hazards model. We applied the model to tissue microarray (TMA) slides of 330 lung cancer patients. The results show that AMIL approaches can handle very small amounts of tissue from a TMA and reach similar C-index performance compared to established survival prediction methods trained with highly discriminative clinical factors such as age, cancer grade, and cancer stage