PuSH - Publikationsserver des Helmholtz Zentrums München

Export: TXT Text RIS Endnote (RIS) BIB BibTeX
de la Rosa, E.* ; Reyes, M.* ; Liew, S.L.* ; Hutton, A.* ; Wiest, R.* ; Kaesmacher, J.* ; Hanning, U.* ; Hakim, A.J.* ; Zubal, R.* ; Valenzuela, W.* ; Robben, D.* ; Sima, D.M.* ; Anania, V.* ; Brys, A.* ; Meakin, J.A.* ; Mickan, A.* ; Broocks, G.* ; Heitkamp, C.* ; Gao, S.* ; Liang, K.W.* ; Zhang, Z.* ; Rahman Siddiquee, M.M.* ; Myronenko, A.* ; Ashtari, P.* ; Van Huffel, S.* ; Jeong, H.* ; Yoon, C.* ; Kim, C.* ; Huo, J.* ; Ourselin, S.* ; Sparks, R.* ; Clèrigues, A.* ; Oliver, A.J.* ; Lladó, X.* ; Chalcroft, L.* ; Pappas, I.* ; Bertels, J.* ; Heylen, E.* ; Moreau, J.* ; Hatami, N.* ; Frindel, C.* ; Qayyum, A.* ; Mazher, M.* ; Puig, D.* ; Lin, S.C.* ; Juan, C.J.* ; Hu, T.* ; Boone, L.* ; Goubran, M.* ; Liu, Y.J.* ; Wegener, S.* ; Kofler, F. ; Ezhov, I.* ; Shit, S.* ; Hernandez Petzsche, M.R.* ; Müller, M.* ; Menze, B.* ; Kirschke, J.S.* ; Wiestler, B.*

DeepISLES: A clinically validated ischemic stroke segmentation model from the ISLES'22 challenge.

Nat. Commun. 16:7357 (2025)
Verlagsversion Forschungsdaten DOI PMC
Open Access Gold
Creative Commons Lizenzvertrag
Diffusion-weighted MRI is critical for diagnosing and managing ischemic stroke, but variability in images and disease presentation limits the generalizability of AI algorithms. We present DeepISLES, a robust ensemble algorithm developed from top submissions to the 2022 Ischemic Stroke Lesion Segmentation challenge we organized. By combining the strengths of best-performing methods from leading research groups, DeepISLES achieves superior accuracy in detecting and segmenting ischemic lesions, generalizing well across diverse axes. Validation on a large external dataset (N = 1685) confirms its robustness, outperforming previous state-of-the-art models by 7.4% in Dice score and 12.6% in F1 score. It also excels at extracting clinical biomarkers and correlates strongly with clinical stroke scores, closely matching expert performance. Neuroradiologists prefer DeepISLES' segmentations over manual annotations in a Turing-like test. Our work demonstrates DeepISLES' clinical relevance and highlights the value of biomedical challenges in developing real-world, generalizable AI tools. DeepISLES is freely available at https://github.com/ezequieldlrosa/DeepIsles .
Impact Factor
Scopus SNIP
Altmetric
15.700
0.000
Tags
Anmerkungen
Besondere Publikation
Auf Hompepage verbergern

Zusatzinfos bearbeiten
Eigene Tags bearbeiten
Privat
Eigene Anmerkung bearbeiten
Privat
Auf Publikationslisten für
Homepage nicht anzeigen
Als besondere Publikation
markieren
Publikationstyp Artikel: Journalartikel
Dokumenttyp Wissenschaftlicher Artikel
Schlagwörter Lesion Segmentation; Computed-tomography; Benchmark; Association; Images; Core
Sprache englisch
Veröffentlichungsjahr 2025
HGF-Berichtsjahr 2025
ISSN (print) / ISBN 2041-1723
e-ISSN 2041-1723
Zeitschrift Nature Communications
Quellenangaben Band: 16, Heft: 1, Seiten: , Artikelnummer: 7357 Supplement: ,
Verlag Nature Publishing Group
Verlagsort London
Begutachtungsstatus Peer reviewed
Institut(e) Helmholtz Artifical Intelligence Cooperation Unit (HAICU)
POF Topic(s) 30205 - Bioengineering and Digital Health
Forschungsfeld(er) Enabling and Novel Technologies
PSP-Element(e) G-530001-001
Förderungen Korea Evaluation Institute of Industrial Technology (KEIT) - Korea government (MOTIE)
Artificial Intelligence Graduate School Program (POSTECH)
Institute of Information & communications Technology Planning & Evaluation (IITP) - Korea government (MSIT)
Basic Science Research Program through the National Research Foundation of Korea (NRF) - Ministry of Education
Swiss Heart Foundation
Swiss National Science Foundation
National Institutes of Health, National Institutes of Neurological Disorders and Stroke (NIH NINDS)
Flemish Government (AI Research Program)
Helmut Horten Foundation
DOI 10.1038/s41467-025-62373-x
WOS ID 001547173700026
Scopus ID 105012936890
PubMed ID 40783484
Erfassungsdatum 2025-10-10
[➜Korrektur melden]