Neural interaction explainable AI predicts drug response across cancers
NAR Cancer · 2025
DOI: 10.1093/narcan/zcaf029 ↗Kurz gefasst
Fachartikel in der Zeitschrift NAR Cancer, 2025. Verfasst von Philipp Keyl und 11 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Pathology, Charité-Universitätsmedizin Berlin, 10117 Berlin, Germany. Hauptthemen laut MEDLINE: Neoplasms, Antineoplastic Agents, Artificial Intelligence. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: German Research Foundation, Bundesministerium für Bildung und Forschung, Institute of Information & Communications Technology Planning & Evaluation.
Abstract (Originalsprache)
Personalized treatment selection is crucial for cancer patients due to the high variability in drug response. While actionable mutations can increasingly inform treatment decisions, most therapies still rely on population-based approaches. Here, we introduce neural interaction explainable AI (NeurixAI), an explainable and highly scalable deep learning framework that models drug–gene interactions and identifies transcriptomic patterns linked with drug response. Trained on data from 546 646 drug perturbation experiments involving 1135 drugs and molecular profiles from 476 tumors, NeurixAI accurately predicted treatment responses for 272 targeted and 30 chemotherapeutic drugs in unseen tumor samples (Spearman’s rho >0.2), maintaining high performance on an external validation set. Additionally, NeurixAI identified the anticancer potential of 160 repurposed non-cancer drugs. Using explainable artificial intelligence (xAI), our framework uncovered key genes influencing drug response at the individual tumor level and revealed both known and novel mechanisms of drug resistance. These findings demonstrate the potential of integrating transcriptomics with xAI to optimize cancer treatment, enable drug repurposing, and identify new therapeutic targets.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Institute of Pathology, Charité-Universitätsmedizin Berlin, 10117 Berlin, Germany.
- Institute of Pathology, Faculty of Medicine, LMU Munich, 80337 Munich, Germany.
- BIFOLD-Berlin Institute for the Foundations of Learning and Data, 10587 Berlin, Germany.
- Institute for Computational Cancer Biology (ICCB), Center for Integrated Oncology (CIO), Cancer Research Center Cologne Essen (CCCE), Faculty of Medicine and University Hospital Cologne, University of Cologne, 50931 Col…
- Institute for Artificial Intelligence in Medicine, University Hospital Essen (AöR), Essen, Germany.
- Institute of Pathology, University Hospital Essen (AöR), 45147 Essen, Germany.
ORCID-Profile: Philipp Keyl, Julius Keyl, Michael Bockmayr, Grégoire Montavon, Klaus-Robert Müller, Frederick Klauschen
Förderung
- German Research Foundation
- Bundesministerium für Bildung und Forschung
- Institute of Information & Communications Technology Planning & Evaluation
- MSIT
- Artificial Intelligence Graduate School Program, Korea University
Zitieren
APA
Keyl, P., Keyl, J., Mock, A., Dernbach, G., Mochmann, L. H., Kiermeyer, N., Jurmeister, P., Bockmayr, M., Schwarz, R. F., Montavon, G., Müller, K. R., & Klauschen, F. (2025). Neural interaction explainable AI predicts drug response across cancers. NAR Cancer. https://doi.org/10.1093/narcan/zcaf029
Vancouver
Keyl P, Keyl J, Mock A, Dernbach G, Mochmann LH, Kiermeyer N, et al. Neural interaction explainable AI predicts drug response across cancers. NAR Cancer. 2025. doi:10.1093/narcan/zcaf029
BibTeX
@article{Keyl2025,
author = {Keyl, Philipp and Keyl, Julius and Mock, Andreas and Dernbach, Gabriel and Mochmann, Liliana H and Kiermeyer, Niklas and Jurmeister, Philipp and Bockmayr, Michael and Schwarz, Roland F and Montavon, Grégoire and Müller, Klaus-Robert and Klauschen, Frederick},
title = {Neural interaction explainable AI predicts drug response across cancers},
journal = {NAR Cancer},
year = {2025},
doi = {10.1093/narcan/zcaf029}
}
EUVIMED