Hetairos is a histology-based artificial intelligence model for predicting central nervous system tumor methylation subtypes

Jin D, Shmatko A, Patel A, Rutz S, Friedrich L, Banan R, Rahmanzade R, Sievers P, Hamelmann S, Schrimpf D, Göbel K, Bogumil H, Maas SL, Sill M, Hinz FE, Suwala AK, Keller F, Habel A, Rukhovich G, Zettl F, Alhalabi OT, Ille S, Sehring J, Amsel D, Wiestler B, Piovesan Lago P, Suchorska B, Ahmad O, Sturm D, Reuss D, Wesseling P, Wöhrer A, Heppner FL, Blümcke I, Delbridge C, Jakobs M, Herold-Mende C, Krieg SM, Wick W, Jones DT, Pfister SM, Al-Hussaini M, Hou Y, D’Almeida Costa F, Schweizer L, Bertero L, Acker T, Tauziede-Espariat A, Varlet P, Merkler D, Egervari K, Dohmen H, Zoroquiain P, Gejman R, Brandner S, Bai X, von Deimling A, Sahm F, Gerstung M (2026)

Publication Type: Journal article

Publication year: 2026

Journal

DOI: 10.1038/s43018-026-01186-3

Abstract

Molecular testing is essential for classifying central nervous system (CNS) tumors, with methylation profiling providing the highest diagnostic granularity. However, this requires more resources and time than conventional hematoxylin and eosin (H&E) histopathology, which is widely available globally. Here we propose Hetairos, an artificial intelligence algorithm that predicts 102 methylation-based CNS tumor subtypes from digital H&E slides. Built and validated on 9,606 patients and over 11,000 slides from 11 centers across four continents, Hetairos identified 50–70% of cases with high confidence, achieving an accuracy of 0.87 for its highest-rated predictions. Hetairos outperformed five board-certified neuropathologists in a direct histology-only comparison (0.68 versus 0.30). Prospective evaluation in routine diagnostics confirmed its performance, reducing turnaround time from 12 days (molecular testing) to 12 min. Hetairos supports diagnostic decision-making across the full spectrum of pediatric and adult CNS tumors by narrowing differential diagnoses and guiding efficient testing.

Authors with CRIS profile

Involved external institutions

Universitätsklinikum Heidelberg Germany (DE) Leiden University Medical Center Netherlands (NL) Hopp-Kindertumorzentrum Heidelberg - KiTZ Germany (DE) Deutsches Krebsforschungszentrum (DKFZ) Germany (DE) Justus-Liebig-Universität Gießen Germany (DE) Technische Universität München (TUM) Germany (DE) A.C.Camargo Cancer Center Brazil (BR) Medizinische Universität Wien Austria (AT) Princess Máxima Center Netherlands (NL) Pontificia Universidad Católica de Chile Chile (CL) King Hussein Cancer Center (KHCC) / مركز الحسين للسرطان Jordan (JO) Chongqing Medical University China (CN) Goethe-Universität Frankfurt am Main Germany (DE) University of Turin / Università degli Studi di Torino (UNITO) Italy (IT) Centre hospitalier Sainte-Anne France (FR) University of Geneva / Université de Genève (UNIGE) Switzerland (CH) University College London (UCL) United Kingdom (GB) Beihang University, Beijing University of Aeronautics and Astronautics (BUAA) / 北京航空航天大学 China (CN) Berliner Institut für Gesundheitsforschung in der Charité / Berlin Institute of Health at Charité (BIH) Germany (DE)

How to cite

APA:

Jin, D., Shmatko, A., Patel, A., Rutz, S., Friedrich, L., Banan, R.,... Gerstung, M. (2026). Hetairos is a histology-based artificial intelligence model for predicting central nervous system tumor methylation subtypes. Nature Cancer. https://doi.org/10.1038/s43018-026-01186-3

MLA:

Jin, Darui, et al. "Hetairos is a histology-based artificial intelligence model for predicting central nervous system tumor methylation subtypes." Nature Cancer (2026).

BibTeX: Download