A Robust Ensemble Algorithm for Ischemic Stroke Lesion Segmentation: Generalizability and Clinical Utility Beyond the ISLES Challenge

Diffusion-weighted MRI (DWI) is essential for stroke diagnosis, treatment decisions, and prognosis. However, image and disease variability hinder the development of generalizable AI algorithms with clinical value. We address this gap by presenting a novel ensemble algorithm derived from the 2022 Ischemic Stroke Lesion Segmentation (ISLES) challenge. ISLES'22 provided 400 patient scans with ischemic stroke from various medical centers, facilitating the development of a wide range of cutting-edge segmentation algorithms by the research community. Through collaboration with leading teams, we combined top-performing algorithms into an ensemble model that overcomes the limitations of individual solutions. Our ensemble model achieved superior ischemic lesion detection and segmentation accuracy on our internal test set compared to individual algorithms. This accuracy generalized well across diverse image and disease variables. Furthermore, the model excelled in extracting clinical biomarkers. Notably, in a Turing-like test, neuroradiologists consistently preferred the algorithm's segmentations over manual expert efforts, highlighting increased comprehensiveness and precision. Validation using a real-world external dataset (N=1686) confirmed the model's generalizability. The algorithm's outputs also demonstrated strong correlations with clinical scores (admission NIHSS and 90-day mRS) on par with or exceeding expert-derived results, underlining its clinical relevance. This study offers two key findings. First, we present an ensemble algorithm (https://github.com/Tabrisrei/ISLES22_Ensemble) that detects and segments ischemic stroke lesions on DWI across diverse scenarios on par with expert (neuro)radiologists. Second, we show the potential for biomedical challenge outputs to extend beyond the challenge's initial objectives, demonstrating their real-world clinical applicability.

SUNet: a deep learning architecture for acute stroke lesion segmentation and outcome prediction in multimodal MRI

Acute stroke lesion segmentation and prediction tasks are of great clinical interest as they can help doctors make better informed time-critical treatment decisions. Automatic segmentation of these lesions is a complex task due to their heterogeneous appearance, dynamic evolution and inter-patient differences. Typically, acute stroke lesion tasks are approached with methods developed for chronic stroke or other brain lesions. However, the pathophysiology and anatomy of acute stroke establishes an inherently different problem that needs special consideration. In this work, we propose a novel deep learning architecture specially designed for acute stroke tasks that involve approximating complex non-linear functions with reduced data. Within our strategy, class imbalance is tackled using a hybrid strategy based on state-of-the-art train sampling strategies designed for other brain lesion related tasks, which is more suited to the anatomy and pathophysiology of acute stroke lesions. The proposed method is evaluated on three unrelated public international challenge datasets (ISLES) without any dataset specific hyper-parameter tuning. These involve the tasks of sub-acute stroke lesion segmentation, acute stroke penumbra estimation and chronic extent prediction from acute MR images. The performance of the proposed architecture is analysed both against similar deep learning architectures from chronic stroke and related biomedical tasks and also by submitting the segmented test images for blind online evaluation on each of the challenges. When compared with the rest of submitted strategies, our method achieves top-rank performance among the best submitted entries in all the three challenges, showing its capability to deal with different unrelated tasks without hyper-parameter tuning. In order to promote the reproducibility of our results, a public version of the proposed method has been released.