ISLES 2024: The first longitudinal multimodal multi-center real-world dataset in (sub-)acute stroke

2024-08-20

arXiv.org

AI Summary

A publicly available longitudinal, multimodal, multicenter dataset with clinical outcome annotations is lacking for AI modeling in stroke, hindering advances in lesion identification and prognostic prediction. Method: We constructed the first real-world longitudinal multimodal dataset for (sub)acute ischemic stroke, integrating acute-phase non-contrast CT, CTA, and CTP with follow-up MRI acquired 2–9 days later, and linking imaging to 3-month clinical outcomes. We established a cross-center standardized acquisition, annotation, and data curation pipeline, and released a high-quality public benchmark—comprising 150 training and 100 test cases—via the ISLES 2024 Challenge. Contribution/Results: This dataset is the first to bridge acute CT and follow-up MRI modalities longitudinally, filling a critical gap in stroke benchmark resources. It enables robust lesion segmentation, brain health quantification, and interpretable prognostic modeling, establishing a new standard for developing and evaluating AI algorithms in stroke.

Technology Category

• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Stroke remains a leading cause of global morbidity and mortality, placing a heavy socioeconomic burden. Over the past decade, advances in endovascular reperfusion therapy and the use of CT and MRI imaging for treatment guidance have significantly improved patient outcomes and are now standard in clinical practice. To develop machine learning algorithms that can extract meaningful and reproducible models of brain function for both clinical and research purposes from stroke images - particularly for lesion identification, brain health quantification, and prognosis - large, diverse, and well-annotated public datasets are essential. While only a few datasets with (sub-)acute stroke data were previously available, several large, high-quality datasets have recently been made publicly accessible. However, these existing datasets include only MRI data. In contrast, our dataset is the first to offer comprehensive longitudinal stroke data, including acute CT imaging with angiography and perfusion, follow-up MRI at 2-9 days, as well as acute and longitudinal clinical data up to a three-month outcome. The dataset includes a training dataset of n = 150 and a test dataset of n = 100 scans. Training data is publicly available, while test data will be used exclusively for model validation. We are making this dataset available as part of the 2024 edition of the Ischemic Stroke Lesion Segmentation (ISLES) challenge (https://www.isles-challenge.org/), which continuously aims to establish benchmark methods for acute and sub-acute ischemic stroke lesion segmentation, aiding in creating open stroke imaging datasets and evaluating cutting-edge image processing algorithms.

Problem

• Lack of longitudinal stroke data for accurate lesion identification
• Need for diverse annotated datasets to predict tissue survival
• Absence of multimodal time-point data for prognosis prediction

Innovation

• Longitudinal multimodal stroke dataset
• Includes acute CT and follow-up MRI
• Vessel occlusion and infarction masks