Published January 1, 2026
| Version v1
Journal article
Open
Deep Learning Predicts Mutations and Outcomes in Gastrointestinal Stromal Tumors
Authors/Creators
Authors/Creators
- Bonetti A.1
- Le V.-L.2
- Carrero Z.I.1
- Wolf F.1
- Gustav M.1
- Lam S.W.3
- Vanhersecke L.4
- Sobczuk P.5
- Le Loarer F.6
- Lenarcik M.7
- Rutkowski P.5
- Sabben J.M.V.8
- Steeghs N.9
- Van Boven H.H.10
- Machado I.11
- Bagué S.12
- Navarro S.13
- Medina-Ceballos E.14
- Agra-Pujol C.15
-
Giner F.16
- Tapia G.17
- Hernàndez-Gallego A.17
- Civantos-Jubera G.18
- Cuatrecasas M.19
- Lopez-Prades S.20
- Perret R.6
- Soubeyran I.4
- Khalifa E.4
- Blouin L.4
- Wardelmann E.21
- Meurgey A.22
- Collini P.23
- Voloshin A.21
- Yatabe Y.24
- Hirano H.25
- Gronchi A.26
- Nishida T.27
- Bouche O.28
- Emile J.-F.29
- Ngo C.30
- Hohenberger P.31
- Cotarelo C.32
- Jakob J.33
- Bovee J.V.M.G.3
- Gelderblom H.34
- Szumera-Cieckiewicz A.35
- Jean-Denis M.36
- Bollard J.36
- Lassau N.37
- Cesne A.L.38
- Blay J.-Y.39
- Italiano A.40
- Crombe A.41
- Coindre J.-M.6
- Kather J.N.42
- 1. Else Kroener Fresenius Center for Digital Health, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany
- 2. Department of Data and Digital Health, Institut Bergonié, Bordeaux, France; Bordeaux Mathematics Institute (IMB), CNRS UMR 5251, University of Bordeaux, Talence, France; INRIA center at University of Bordeaux, Talence, France
- 3. Department of Pathology, Leiden University Medical Center, Leiden, Netherlands
- 4. Department of Biopathology, Institut Bergonié, Bordeaux, France
- 5. Department of Soft Tissue, Bone Sarcoma and Melanoma, Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland
- 6. Department of Biopathology, Institut Bergonié, Bordeaux, France; Institute of Oncology, BRIC U1312, INSERM, Université de Bordeaux, Institut Bergonié, Bordeaux, France
- 7. Department of Gastroenterology, Hepatology and Clinical Oncology, Centre of Postgraduate Medical Education, Warsaw, Poland; Department of Cancer Pathology, Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland
- 8. Department of Medical Oncology, The Netherlands Cancer Institute, Amsterdam, Netherlands
- 9. Department of Medical Oncology, The Netherlands Cancer Institute, Amsterdam, Netherlands; Department of Medical Oncology, University Medical Centre Utrecht, Utrecht University, Utrecht, Netherlands
- 10. Department of Pathology, The Netherlands Cancer Institute, Amsterdam, Netherlands
- 11. Department of Pathology, Fundaci´on Instituto Valenciano de Oncología, Valencia, Spain; Patologika Laboratory Hospital Quir´on-Salud, Valencia, Spain; Department of Pathology, University of Valencia, Valencia, Spain; CIBERONC(ISCIII), Madrid, Spain
- 12. Department of Pathology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain
- 13. Department of Pathology, University of Valencia, Valencia, Spain; CIBERONC(ISCIII), Madrid, Spain; INCLIVA, Valencia, Spain
- 14. Department of Pathology, University of Valencia, Valencia, Spain
- 15. Pathology Department, HGU Gregorio Marañon, Madrid, Spain
- 16. Department of Pathology, University of Valencia, Valencia, Spain; Department of Pathology, Hospital Universitari i Politècnic la Fe, Valencia, Spain
- 17. Department of Pathology, Hospital Universitari Germans Trias i Pujol, IGTP, Universitat Aut`onoma de Barcelona, Barcelona, Spain
- 18. Department of Pathology, Hospital Universitario Virgen del Rocío, Sevilla, Spain
- 19. Pathology Department, Hospital Clínic, Barcelona, Spain; Liver and Digestive Diseases Biomedical Research Center Network (CIBERehd), Madrid, Spain; Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain; University of Barcelona, Faculty of Medicine, Barcelona, Spain
- 20. Pathology Department, Hospital Clínic, Barcelona, Spain; Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain
- 21. Gerhard-Domagk-Institute of Pathology, University Hospital Münster, Münster, Germany
- 22. Biopathology Department, Centre Léon Bérard, Unicancer, Lyon, France
- 23. Department of Advanced Diagnostics, Soft Tissue Tumor Pathology Unit, Fondazione IRCCS Istituto Nazionale dei Tumori di Milano, Milan, Italy
- 24. Department of Diagnostic Pathology, National Cancer Center Hospital, Tokyo, Japan
- 25. Department of Gastrointestinal Medical Oncology, National Cancer Center Hospital, Tokyo, Japan
- 26. Department of Surgery, Fondazione IRCCS Istituto Nazionale dei Tumori, University of Milan, Milan, Italy
- 27. Department of Surgery, Japan Community Healthcare Organization Osaka Hospital, Osaka, Japan; Department of Surgery, National Cancer Center Hospital, Tokyo, Japan
- 28. Gastrointestinal Oncology Department, CHU Reims, Reims, France
- 29. Paris-Saclay University, Versailles SQY University, INSERM U981, Assistance Publique–Hôpitaux de Paris (AP-HP), Ambroise-Paré Hospital, Smart Imaging, Service de Pathologie, Boulogne, France
- 30. Department of Pathology, Institut Gustave Roussy, Villejuif, France
- 31. Division of Surgical Oncology, Mannheim University Medical Center, Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
- 32. Institute of Pathology, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany
- 33. Department of Surgery, Sarcoma Unit, University Medical Center and Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany
- 34. Department of Medical Oncology, Leiden, Netherlands
- 35. Biobank, Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland
- 36. Research and Innovation Department, Centre Léon Bérard, Unicancer, Lyon, France
- 37. Imaging Department, Institut Gustave Roussy, BIOMAPS, UMR1281, University Paris-Saclay, Villejuif, France
- 38. Department of Medical Oncology, Institut Gustave Roussy, Villejuif, France
- 39. Department of Medicine, Centre Leon Berard, University Claude Bernard Lyon, Villeurbanne, France
- 40. Department of Medicine, Institut Bergonié, Bordeaux, France; Institute of Oncology, BRIC U1312, INSERM, Université de Bordeaux, Institut Bergonié, Bordeaux, France
- 41. Department of Oncologic Imaging, Institut Gustave Roussy, Villejuif, France
- 42. Else Kroener Fresenius Center for Digital Health, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany; Department of Medicine I, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany; Medical Oncology, National Center for Tumor Diseases (NCT), University Hospital Heidelberg, Heidelberg, Germany; Pathology and Data Analytics, Leeds Institute of Medical Research, St James's University of Leeds, Leeds, United Kingdom
Description
Gastrointestinal stromal tumor (GIST) is the most common gastrointestinal mesenchymal tumor, driven by tyrosine– protein kinase (KIT) and platelet-derived growth factor receptor A (PDGFRA) mutations. Specific variants, such as KIT exon 11 deletions, carry prognostic and therapeutic implications, whereas wild-type variants derive limited benefit from tyrosine kinase inhibitors. Given the limited reproducibility of established clinicopathologic risk models, deep learning (DL) applied to whole-slide images (WSI) emerged as a promising tool for molecular classification and prognostic assessment. We analyzed 8398 GIST cases from 21 centers in seven countries, including 7,238 with molecular data and 2,638 with clinical follow-up. DL models were trained on WSIs to predict mutations, treatment sensitivity, and recurrence-free survival (RFS). DL predicted mutational status in GIST from WSIs, with area under the curve of 0.87 for KIT and 0.96 for PDGFRA, and high performance was observed for subtypes, including KIT exon 11 del–inss 557 to 558 (0.67) and PDGFRA exon 18 D842V (0.93). For therapeutic categories, performance reached 0.84 for avapritinib sensitivity and 0.81 for imatinib sensitivity. DL models predicted RFS, with hazard ratios of 8.44 in the overall cohort and 4.74 in patients receiving adjuvant therapy. Prognostic performance was comparable with pathology-based scores, with highest discrimination in the overall cohort and in patients without adjuvant therapy. DL applied to WSIs enables prediction of molecular alterations, treatment sensitivity, and RFS in GIST, performing comparably with established risk scores across international cohorts, providing a baseline for future multimodal predictors. © (2026), (American Association for Cancer Research Inc.). All rights reserved.
Notes
The authors gratefully acknowledge the GWK support for funding this project by providing computing time through the Center for Information Services and HPC (ZIH) at TU Dresden. The authors gratefully acknowledge the Gauss Centre for Supercomputing e.V. (www.gauss-centre.eu) for funding this project by providing computing time through the John von Neumann Institute for Computing on the GCS Supercomputer JUWELS at J\u00A8ulich Supercomputing Centre. During the preparation of this work, the author(s) used Google Gemini 2.5, ChatGPT 4, and ChatGPT 5 in order to improve grammar, sentence structure, and overall clarity. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. J.N. Kather is supported by the German Cancer Aid DKH (DECADE, 70,115,166), the German Federal Ministry of Research, Technology, and Space BMFTR (PEARL, 01KD2104C; CAMINO, 01EO2101; TRANSFORM LIVER, 031L0312A; TANGERINE, 01KT2302 through ERA-NET Transcan; Come2Data, 16DKZ2044A; DEEP-HCC, 031L0315A; DECIPHER-M, 01KD2420A; NextBIG, 01ZU2402A), the German Research Foundation DFG (CRC/TR 412, 535081457; SFB 1709/1 2025, 533056198), the German Academic Exchange Service DAAD (SECAI, 57616814), the German Federal Joint Committee G-BA (TransplantKI, 01VSF21048), the European Union\u2019s (EU) Horizon Europe Research and Innovation Program (ODELIA, 101057091; GENIAL, 101096312), the European Research Council ERC (NADIR, 101114631), the National Institutes of Health (NIH, EPICO, R01 CA263318), and the National Institute for Health and Care Research (NIHR, Leeds Biomedical Research Centre, NIHR203331). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care. This work was funded by the EU. The views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the EU. Neither the EU nor the granting authority can be held responsible for them. J.-M. Coindre is supported by Info Sarcomes and Intersarc. V.-L. Le is supported by La Fondation Bergoni\u00B4e. J. Jakob is supported by the German Cancer Aid DKH (70116152), the German Federal Ministry of Research, Technology, and Space BMFTR (01KD2511A), and the DKFZ-Hector Cancer Institute at the University Medical Center Mannheim (TETRIS). J.-Y. Blay received grants from NETSARC+ (INCAZ FRance) & LYRICAN+ (INCA France), and EURACAN EU commission).