TranSOP: Transformer-based Multimodal Classification for Stroke Treatment Outcome Prediction

TL;DR

This paper introduces TranSOP, a transformer-based multimodal network that combines clinical data and imaging to predict stroke treatment outcomes, achieving state-of-the-art accuracy on the MRCLEAN dataset.

Contribution

The paper presents a novel multimodal transformer model that effectively fuses clinical and imaging data for stroke outcome prediction.

Findings

Achieved a state-of-the-art AUC score of 0.85 on the MRCLEAN dataset.

Demonstrated superior performance of multimodal over unimodal approaches.

Validated the effectiveness of the fusion module in combining clinical and imaging features.

Abstract

Acute ischaemic stroke, caused by an interruption in blood flow to brain tissue, is a leading cause of disability and mortality worldwide. The selection of patients for the most optimal ischaemic stroke treatment is a crucial step for a successful outcome, as the effect of treatment highly depends on the time to treatment. We propose a transformer-based multimodal network (TranSOP) for a classification approach that employs clinical metadata and imaging information, acquired on hospital admission, to predict the functional outcome of stroke treatment based on the modified Rankin Scale (mRS). This includes a fusion module to efficiently combine 3D non-contrast computed tomography (NCCT) features and clinical information. In comparative experiments using unimodal and multimodal data on the MRCLEAN dataset, we achieve a state-of-the-art AUC score of 0.85.