The TopCoW Challenge — Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

The Circle of Willis (CoW) is an important network of arteries connecting major circulations of the brain. Its vascular architecture is believed to affect the risk, severity, and clinical outcome of serious neurovascular diseases. However, characterizing the highly variable CoW anatomy is still a manual and time-consuming expert task. The CoW is usually imaged by two non-invasive angiographic imaging modalities, magnetic resonance angiography (MRA) and computed tomography angiography (CTA), but there exist limited datasets with annotations on CoW anatomy, especially for CTA. Therefore, we organized the TopCoW challenge with the release of an annotated CoW dataset. The TopCoW dataset is the first public dataset with voxel-level annotations for 13 CoW vessel components, enabled by virtual reality technology. It is also the first large dataset using 200 pairs of MRA and CTA from the same patients. As part of the benchmark, we invited submissions worldwide and attracted over 250 registered participants from six continents. The submissions were evaluated on both internal and external test datasets of 226 scans from over five centers. The top performing teams achieved over 90% Dice scores at segmenting the CoW components, over 80% F1 scores at detecting key CoW components, and over 70% balanced accuracy at classifying CoW variants for nearly all test sets. The best algorithms also showed clinical potential in classifying fetal-type posterior cerebral artery and locating aneurysms with CoW anatomy. TopCoW demonstrated the utility and versatility of CoW segmentation algorithms for a wide range of downstream clinical applications with explainability. The annotated datasets and best performing algorithms have been released as public Zenodo records to foster further methodological development and clinical tool building.

Authors

Institutions

Publication Details

Journal
NEJM AI
Published
2026-07-23
DOI
https://doi.org/10.1056/aidbp2500994
Citations
10
Primary Topic
Cerebrovascular and Carotid Artery Diseases
Type
article
Field-Weighted Citation Impact
14.84
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

The TopCoW Challenge — Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

Rami Al-Maskari, Maximilian Rokuss, Jan S. Kirschke, Iris N. Vos et al.
10 citations
NEJM AI
Cerebrovascular and Carotid Artery Diseases
14.84
article

The TopCoW Challenge — Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

Rami Al-Maskari, Maximilian Rokuss, Jan S. Kirschke, Iris N. Vos, Martin J. Menten, Daniel Rueckert, Junqiang Chen, Nico Disch, Emmanuel Montagnon, Marek Wodziński, Hakim Baazaoui, Valeriia Abramova, Suprosanna Shit, Florian Kofler, Elisa Colombo, Jeroen Bisschop, Luca Giancardo, Klaus Maier‐Hein, Andrew Makmur, Philippe Bijlenga, Uma Maria Lal-Trehan Estrada, Ivan Ezhov, Benedikt Wiestler, Jialu Liu, Francesco Galati, Ynte M. Ruigrok, Roland Wiest, Laura P. Westphal, Xavier Lladó, İbrahim Ethem Hamamcı, Chinmay Prabhakar, Birgitta K. Velthuis, Fabian Isensee, Norman Juchler, Houjing Huang, Ezequiel de la Rosa, Fernando Navarro, Hugo J. Kuijf, Diana Waldmannstetter, James Thomas Patrick Decourcy Hallinan, Johannes C. Paetzold, Anjany Sekuboyina, Yannick Kirchhoff, Luciano Höher, Daniele Falcetta, Hongwei Li, Hyunjin Park, Wei Liu, Chengcheng Zhu, Pengcheng Shi, Jongyun Hwang, Yihui Ma, Kaiyuan Yang, Ting Ma, Haoran Zhao, Fabio Musio, Cansu Yalçın, Julien Hämmerli, Sinyoung Ra, Catherine Wurster
article en
10 citations

Abstract

The Circle of Willis (CoW) is an important network of arteries connecting major circulations of the brain. Its vascular architecture is believed to affect the risk, severity, and clinical outcome of serious neurovascular diseases. However, characterizing the highly variable CoW anatomy is still a manual and time-consuming expert task. The CoW is usually imaged by two non-invasive angiographic imaging modalities, magnetic resonance angiography (MRA) and computed tomography angiography (CTA), but there exist limited datasets with annotations on CoW anatomy, especially for CTA. Therefore, we organized the TopCoW challenge with the release of an annotated CoW dataset. The TopCoW dataset is the first public dataset with voxel-level annotations for 13 CoW vessel components, enabled by virtual reality technology. It is also the first large dataset using 200 pairs of MRA and CTA from the same patients. As part of the benchmark, we invited submissions worldwide and attracted over 250 registered participants from six continents. The submissions were evaluated on both internal and external test datasets of 226 scans from over five centers. The top performing teams achieved over 90% Dice scores at segmenting the CoW components, over 80% F1 scores at detecting key CoW components, and over 70% balanced accuracy at classifying CoW variants for nearly all test sets. The best algorithms also showed clinical potential in classifying fetal-type posterior cerebral artery and locating aneurysms with CoW anatomy. TopCoW demonstrated the utility and versatility of CoW segmentation algorithms for a wide range of downstream clinical applications with explainability. The annotated datasets and best performing algorithms have been released as public Zenodo records to foster further methodological development and clinical tool building.

NEJM AIVol. 3(8)
University of Bern (CH), Harvard University (US), Shenzhen Institute of Information Technology (CN), Universitat Pompeu Fabra (ES), German Cancer Research Center (DE), Shanghai Jiao Tong University (CN), University of Toronto (CA), EURECOM (FR), Utrecht University (NL), Korea University (KR), Chinese Academy of Sciences (CN), ZHAW Zurich University of Applied Sciences (CH), University of Washington (US), Peking University (CN), University of Zurich (CH), Harbin Institute of Technology (CN), Cornell University (US), Heidelberg University (DE), Universitat de Girona (ES), Renji Hospital (CN), University Hospital of Bern (CH), University Hospital Heidelberg (DE), Helmholtz Zentrum München (DE), University Medical Center Utrecht (NL), Wuhan University (CN), University of Illinois Chicago (US), University of Chicago (US), Center for Excellence in Brain Science and Intelligence Technology (CN), Beijing Academy of Artificial Intelligence (CN), University Hospital of Zurich (CH), Research Institute of Precision Instruments (Russia) (RU), University Hospital of Geneva (CH), Laboratoire de Thermique et Energie de Nantes (FR), Institute of Automation (CN), Canon (United States) (US), Zhongnan Hospital of Wuhan University (CN), Duke Medical Center (US), Athinoula A. Martinos Center for Biomedical Imaging (US), HES-SO Valais-Wallis (CH), Peng Cheng Laboratory (CN), Centre Hospitalier de l’Université de Montréal (CA), Institut du Thorax (FR), National University Hospital (SG), Venus Medtech (China) (CN), University of Chinese Academy of Sciences (CN), Munich Center for Machine Learning, University College London (GB), Imperial College London (GB), Seattle University (US), Technical University of Munich (DE), University at Buffalo, State University of New York (US), AGH University of Krakow (PL), Université de Montréal (CA), Xiamen University of Technology (CN), Charité - Universitätsmedizin Berlin (DE), Sungkyunkwan University (KR), ZHAW Zurich University of Applied Sciences (CH), The University of Texas Health Science Center at Houston (US), Nantes Université (FR)
Openalex Percentile: Top 2%
Cerebrovascular and Carotid Artery Diseases
14.84
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.