# Automated Multi-Modal MRI Segmentation of Stroke Lesions and Corticospinal Tract Integrity for Functional Outcome Prediction

## TL;DR

This paper introduces a new MRI-based framework to predict stroke recovery outcomes using standard imaging techniques and automated analysis.

## Contribution

The novel contribution is a practical, automated multimodal MRI pipeline for stroke lesion and corticospinal tract analysis using routine clinical imaging.

## Key findings

- Automated lesion segmentation achieved a Dice score of 0.82 on the ISLES 2022 dataset.
- Key predictors of stroke outcomes included lesion–CST overlap, lesion volume, and texture features.

## Abstract

Stroke recovery varies widely between patients, making it difficult to estimate functional outcomes at hospital discharge. Reliable prediction methods are often based on advanced imaging techniques that are not routinely used in clinical care. This study presents a practical MRI-based framework that relies only on standard imaging commonly acquired in stroke patients. By automatically identifying stroke lesions and assessing their relationship to key motor pathways, we derived interpretable imaging markers linked to functional outcome. Our findings suggest that routinely available MRI data can support clinically meaningful and transparent prediction of short-term stroke outcomes, highlighting the potential for broader clinical adoption.

Background/Objectives: Stroke is a leading cause of long-term disability, and predicting functional outcome at discharge, such as the modified Rankin Scale (mRS), is important for guiding treatment and rehabilitation. Many existing approaches depend on advanced imaging or complex corticospinal tract (CST) segmentation from multi-shell diffusion MRI, limiting clinical feasibility. Automated lesion segmentation is also challenging due to lesion heterogeneity and MRI variability. This study proposes a clinically feasible multimodal MRI pipeline based on routine imaging. Methods: Lesion segmentation models were trained and evaluated on the ISLES 2022 dataset (250 training, 150 test cases). Zero-shot external evaluation was performed on 149 cases from ISLES 2024 using standard MRI sequences only. An ensemble of deep learning models (SEALS, NVAUTO, FACTORIZER) was evaluated on ISLES 2022, while SEALS alone was used for external testing. CST segmentation was performed using TractSeg on single-shell diffusion-weighted imaging. Imaging biomarkers included lesion volume, shape, ADC-based texture features, CST integrity, and lesion–CST overlap. These features were used to train machine learning models for binary mRS prediction at discharge. Results: The ensemble achieved a Dice score of 0.82 on ISLES 2022, while zero-shot evaluation on ISLES 2024 achieved 0.57. In exploratory analysis, CatBoost achieved the highest point estimates (accuracy 0.88, F1-score 0.87, ROC-AUC 0.83). Key predictors included lesion–CST overlap, lesion volume, surface area, dissimilarity, and contrast. Conclusions: This exploratory study demonstrates the feasibility of combining automated lesion segmentation with anatomically informed biomarkers using routine clinical MRI, supporting interpretable stroke outcome modelling and motivating future large-scale validation.