ISLES’24: Final stroke infarct prediction with multimodal imaging and clinical data. Where do we stand?
Abstract
Accurate estimation of brain infarction (i.e., irreversibly damaged tissue) is critical for guiding treatment decisions in acute ischemic stroke. Reliable infarct prediction informs key clinical interventions, including the need for patient transfer to comprehensive stroke centers, the potential benefit of additional reperfusion attempts during mechanical thrombectomy, decisions regarding secondary neuroprotective treatments, and, ultimately, prognosis of clinical outcomes. This work introduces the Ischemic Stroke Lesion Segmentation (ISLES) 2024 challenge, which focuses on the prediction of final infarct volumes from pre-interventional acute stroke imaging and clinical data. ISLES’24 provides a comprehensive, multimodal setting where participants can leverage all clinically and practically available data —including full acute CT imaging, sub-acute follow-up MRI, and structured clinical information— across a train set of N = 150 cases. On the hidden test set ( N = 98 ), the top-performing model, a multimodal nnU-Net-based architecture, achieved a Dice score of 0.285 ± 0.213 and an absolute volume difference of 21.2 ± 37.2 mL, underlining the significant challenges posed by this task and the need for further advances in multimodal learning. This overview of the ISLES’24 challenge offers two primary contributions: first, it provides a standardized, clinically realistic benchmark for post-treatment infarct prediction, built on a comprehensive, open, and multimodal longitudinal stroke imaging dataset; second, it provides insights into the strengths and limitations of current algorithms and delineates key research directions to guide the development of next-generation infarct prediction models.
Related
- Endovascular thrombectomy for acute ischemic stroke with large infarct core: translating results from a randomized trial into clinical practice
- Multimodal preoperative risk stratification of pancreatic mass lesions using clinical variables, imaging radiomics and plasma cfDNA methylation: protocol for an ambispective multicentre diagnostic prediction model study
- Choriocapillaris Dysfunction in Retinal Diseases: Multimodal Imaging Biomarkers, Pathophysiological Insights, and Emerging Clinical Applications
- Performance and incremental value of preoperative multimodal imaging and perioperative clinical characteristics in predicting postoperative major adverse events in type A aortic dissection
- Early prediction of 3-month functional outcome after acute ischemic stroke: an explainable model based on routine clinical data
BibTeX
@article{rosa2026isles24,
title = {{ISLES’24: Final stroke infarct prediction with multimodal imaging and clinical data. Where do we stand?}},
author = {Ezequiel de la Rosa and Ruisheng Su and Mauricio Reyes and Evamaria Olga Riedel and Hakim Baazaoui and The Anh Baran and R Wiest and Florian Kofler and Kaiyuan Yang and Murong Xu and Julian Deseoe and Lisa Herzog and David Robben and Mahsa Mojtahedi and Laura Van. Poppel and Lucas de Vries and Anthony Winder and Kimberly Amador and Nils D. Forkert and Gyeongyeon Hwang and Jiwoo Song and Dohyun Kim and Eneko Uruñuela and Annabella Bregazzi and Matthias Wilms and Hyun Yang and Jin Tae Kwak and Sumin Jung and Luan Matheus Trindade Dalmazo and Kumaradevan Punithakumar and Moona Mazher and Abdul Qayyum and Steven Niederer and Jacob Idoko and Mariana Bento and Gouri Ginde and Tianyi Ren and Juampablo Heras Rivera and Mehmet Kurt and Cansu Yalçın and Adrià Casamitjana and Arnau Oliver and Xavier Lladó and Carole Frindel and Susanne Wegener and Jan Stefan Kirschke and Benedikt Wiestler and Bjoern H. Menze},
journal = {Medical Image Analysis},
year = {2026},
month = nov,
doi = {10.1016/j.media.2026.104362},
url = {https://doi.org/10.1016/j.media.2026.104362},
note = {CC BY},
}
RIS
TY - JOUR
TI - ISLES’24: Final stroke infarct prediction with multimodal imaging and clinical data. Where do we stand?
AU - Ezequiel de la Rosa
AU - Ruisheng Su
AU - Mauricio Reyes
AU - Evamaria Olga Riedel
AU - Hakim Baazaoui
AU - The Anh Baran
AU - R Wiest
AU - Florian Kofler
AU - Kaiyuan Yang
AU - Murong Xu
AU - Julian Deseoe
AU - Lisa Herzog
AU - David Robben
AU - Mahsa Mojtahedi
AU - Laura Van. Poppel
AU - Lucas de Vries
AU - Anthony Winder
AU - Kimberly Amador
AU - Nils D. Forkert
AU - Gyeongyeon Hwang
AU - Jiwoo Song
AU - Dohyun Kim
AU - Eneko Uruñuela
AU - Annabella Bregazzi
AU - Matthias Wilms
AU - Hyun Yang
AU - Jin Tae Kwak
AU - Sumin Jung
AU - Luan Matheus Trindade Dalmazo
AU - Kumaradevan Punithakumar
AU - Moona Mazher
AU - Abdul Qayyum
AU - Steven Niederer
AU - Jacob Idoko
AU - Mariana Bento
AU - Gouri Ginde
AU - Tianyi Ren
AU - Juampablo Heras Rivera
AU - Mehmet Kurt
AU - Cansu Yalçın
AU - Adrià Casamitjana
AU - Arnau Oliver
AU - Xavier Lladó
AU - Carole Frindel
AU - Susanne Wegener
AU - Jan Stefan Kirschke
AU - Benedikt Wiestler
AU - Bjoern H. Menze
PY - 2026
DA - 2026/11/01
JO - Medical Image Analysis
DO - 10.1016/j.media.2026.104362
UR - https://doi.org/10.1016/j.media.2026.104362
AB - Accurate estimation of brain infarction (i.e., irreversibly damaged tissue) is critical for guiding treatment decisions in acute ischemic stroke. Reliable infarct prediction informs key clinical interventions, including the need for patient transfer to comprehensive stroke centers, the potential benefit of additional reperfusion attempts during mechanical thrombectomy, decisions regarding secondary neuroprotective treatments, and, ultimately, prognosis of clinical outcomes. This work introduces the Ischemic Stroke Lesion Segmentation (ISLES) 2024 challenge, which focuses on the prediction of final infarct volumes from pre-interventional acute stroke imaging and clinical data. ISLES’24 provides a comprehensive, multimodal setting where participants can leverage all clinically and practically available data —including full acute CT imaging, sub-acute follow-up MRI, and structured clinical information— across a train set of N = 150 cases. On the hidden test set ( N = 98 ), the top-performing model, a multimodal nnU-Net-based architecture, achieved a Dice score of 0.285 ± 0.213 and an absolute volume difference of 21.2 ± 37.2 mL, underlining the significant challenges posed by this task and the need for further advances in multimodal learning. This overview of the ISLES’24 challenge offers two primary contributions: first, it provides a standardized, clinically realistic benchmark for post-treatment infarct prediction, built on a comprehensive, open, and multimodal longitudinal stroke imaging dataset; second, it provides insights into the strengths and limitations of current algorithms and delineates key research directions to guide the development of next-generation infarct prediction models.
SN - 1361-8415
SN - 1361-8423
SN - 1361-8431
ER -