---
title: "Deep learning on routine histopathology adds prognostic value beyond AJCC stage in early-stage mel"
id: "medrxiv-15-deep-learning-reveals-prognostic-information-beyond-ajcc-stage-in-routine"
canonical_url: "https://medichelpline.com/clinical-feed/medrxiv-15-deep-learning-reveals-prognostic-information-beyond-ajcc-stage-in-routine"
content_type: "clinical_feed_article"
specialty: "Oncology"
source_name: "medRxiv (Clinical Preprints)"
source_url: "https://www.medrxiv.org/content/10.64898/2026.10.29.26366599v1?rss=1"
published_at: "2026-10-30T12:00:00.000Z"
evidence_level: "Verified Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Deep learning on routine histopathology adds prognostic value beyond AJCC stage in early-stage mel
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/medrxiv-15-deep-learning-reveals-prognostic-information-beyond-ajcc-stage-in-routine
- **Specialty:** [Oncology](https://medichelpline.com/clinical-feed/oncology.md)
- **Primary Source:** medRxiv (Clinical Preprints)
- **Source URL:** [Original Journal Publication](https://www.medrxiv.org/content/10.64898/2026.10.29.26366599v1?rss=1)
- **Published At:** 2026-10-30T12:00:00.000Z
- **Evidence Rating:** Verified Feed
## Executive GIST (TL;DR)
- The population-based **D-ESMEL** study evaluated whether histopathological images from primary early-stage melanoma contain prognostic signals beyond clinical stage. - Investigators used a discovery set of n=328 case-control sets matched on follow-up time and established prognostic features (Breslow thickness, ulceration, age, sex). - A larger validation cohort followed a nested case-control (NCC) design drawn from a >5,500-patient population; the validation analysis included n=266 stage-matched controls. - The team developed an end-to-end **deep learning** workflow to extract patches automatically from **whole slide images** (WSIs) and produce patient-level metastatic risk predictions based solely on histopathology. - Training used 10-fold cross-validation on the discovery set; the best validation fold produced a balanced accuracy of 0.88 (mean 0.81; 95% CI 0.77–0.85). - In the validation cohort, model predictions were independently associated with future metastasis (adjusted odds ratio 1.97; p<0.01), indicating prognostic information beyond **AJCC** stage. - Discrimination metrics: the model alone reached a C-index of 0.86; **AJCC** stage alone had a C-index of 0.80; combining predictions with **AJCC** stage increased the C-index to 0.87, a statistically significant improvement versus stage alone (p<0.001). - Authors reported that histopathology-based predictions may help refine selection for (neo)adjuvant therapy, but they cautioned that the work is a preprint and requires further biological investigation and external validation. - Ethical approvals were obtained from Erasmus MC committees and national registries; image and clinical data are available on reasonable request with permissions from the Netherlands Comprehensive Cancer Organisation. - The source did not report detailed model architecture, hyperparameters, per-class confusion matrices, or individual-level case examples within the abstract text; those details may be in supplementary material or the full preprint.
## Clinical Analysis & Structured Key Points
Deep learning reveals prognostic information beyond AJCC stage in routine histopathology of early-stage melanoma: the population-based D-ESMEL study | medRxiv Skip to main content Deep learning reveals prognostic information beyond AJCC stage in routine histopathology of early-stage melanoma: the population-based D-ESMEL study View ORCID Profile Jasper Ouwerkerk , View ORCID Profile Farhan Akram , View ORCID Profile Loes Hollestein , View ORCID Profile Thamila Kerkour , View ORCID Profile Catherine Zhou , View ORCID Profile Andrew Stubbs , View ORCID Profile Antien Mooyaart , View ORCID Profile Yunlei Li doi: https://doi.org/10.64898/2026.10.29.26366599 Jasper Ouwerkerk 1 Department of Pathology and Clinical Bioinformatics, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jasper Ouwerkerk Farhan Akram 1 Department of Pathology and Clinical Bioinformatics, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Farhan Akram Loes Hollestein 2 Department of Dermatology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Loes Hollestein Thamila Kerkour 2 Department of Dermatology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Thamila Kerkour Catherine Zhou 2 Department of Dermatology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Catherine Zhou Andrew Stubbs 1 Department of Pathology and Clinical Bioinformatics, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Andrew Stubbs Antien Mooyaart 1 Department of Pathology and Clinical Bioinformatics, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Antien Mooyaart Yunlei Li 1 Department of Pathology and Clinical Bioinformatics, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands; Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Yunlei Li For correspondence: y.li.1{at}erasmusmc.nl Abstract Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Of all melanoma patients who develop distant metastasis over time, 41% were initially diagnosed with stage I/II. More accurate prognostic models are needed to select patients for potential (neo)adjuvant therapy. We hypothesise that histopathological features in whole slide images (WSIs) from primary early-stage melanomas provide additional information beyond American Joint Committee on Cancer (AJCC) stage. We utilised data from the population-based Dutch Early-Stage Melanoma (D-ESMEL) study. This included a discovery set (n=328) of case-control sets matched on follow-up time and known prognostic features (i.e. Breslow thickness, ulceration, age, and sex) and a validation cohort based on >5,500 patients following a nested case-control (NCC) design with stage-matched controls (n=266). We developed an end-to-end deep learning workflow to automatically extract relevant patches from WSIs and predict patient-level metastatic risk based on histopathological features independent of AJCC stage. The discovery set served as a training set for the workflow using 10-fold cross-validation. The prognostic performance of this model and its independence from clinical risk factors was evaluated in the validation cohort using odds ratios (OR) and the weighted C-index. We achieved a balanced accuracy of 0.88 (mean: 0.81; 95% CI: 0.77-0.85) in the best validation fold of the discovery set. In the validation cohort, metastatic risk predictions from the deep learning model were an independent prognostic factor (adjusted OR: 1.97; p<0.01). Combining the deep learning predictions (C-index: 0.86) with AJCC stage (C-index: 0.80) yielded the highest C-index of 0.87, significantly improving the C-index compared to AJCC stage alone (p<0.001). Our histopathology-based deep learning model was rigorously evaluated in a population-based cohort and demonstrated added prognostic value beyond AJCC stage. Further investigations of the underlying biological features could improve clinical decision making and patient outcomes. Competing Interest Statement The authors have declared no competing interest. Clinical Protocols https://link.springer.com/article/10.1007/s10654-024-01188-4 https://link.springer.com/article/10.1007/s10654-025-01337-3 Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The scientific committees of Erasmus MC (MEC-2018-1738), the Dutch Nationwide Pathology Databank, and the Netherlands Cancer Registry gave ethical approval for this work. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data Availability The clinical data from these patients was provided to the authors by the IKNL. These data are not publicly available and restrictions apply to the availability of the data used for the current study. However, the histopathological images and clinical data are available upon reasonable request to the authors and with permission of the Netherlands Comprehensive Cancer Organisation. Funder Information Declared KWF Kankerbestrijding , 2021-1/13470 Copyright The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . Back to top Previous Next Posted October 30, 2026. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Deep learning reveals prognostic information beyond AJCC stage in routine histopathology of early-stage melanoma: the population-based D-ESMEL study Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Deep learning reveals prognostic information beyond AJCC stage in routine histopathology of early-stage melanoma: the population-based D-ESMEL study Jasper Ouwerkerk , Farhan Akram , Loes Hollestein , Thamila Kerkour , Catherine Zhou , Andrew Stubbs , Antien Mooyaart , Yunlei Li medRxiv 2026.10.29.26366599; doi: https://doi.org/10.64898/2026.10.29.26366599 Share This Article: Copy Citation Tools Deep learning reveals prognostic information beyond AJCC stage in routine histopathology of early-stage melanoma: the population-based D-ESMEL study Jasper Ouwerkerk , Farhan Akram , Loes Hollestein , Thamila Kerkour , Catherine Zhou , Andrew Stubbs , Antien Mooyaart , Yunlei Li medRxiv 2026.10.29.26366599; doi: https://doi.org/10.64898/2026.10.29.26366599 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Areas All Articles Addiction Medicine (637) Allergy and Immunology (916) Anesthesia (348) Cardiovascular Medicine (4974) Dentistry and Oral Medicine (493) Dermatology (426) Emergency Medicine (674) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1681) Epidemiology (16202) Forensic Medicine (36) Gastroenterology (1238) Genetic and Genomic Medicine (7249) Geriatric Medicine (751) Health Economics (1092) Health Informatics (5261) Health Policy (1463) Health Systems and Quality Improvement (1825) Hematology (599) HIV/AIDS (1375) Infectious Diseases (except HIV/AIDS) (16463) Intensive Care and Critical Care Medicine (1199) Medical Education (696) Medical Ethics (156) Nephrology (747) Neurology (7551) Nursing (377) Nutrition (1120) Obstetrics and Gynecology (1293) Occupational and Environmental Health (1038) Oncology (3729) Ophthalmology (1087) Orthopedics (405) Otolaryngology (463) Pain Medicine (499) Palliative Medicine (143) Pathology (720) Pediatrics (1851) Pharmacology and Therapeutics (752) Primary Care Research (784) Psychiatry and Clinical Psychology (6099) Public and Global Health (9956) Radiology and Imaging (2515) Rehabilitation Medicine and Physical Therapy (1506) Respiratory Medicine (1264) Rheumatology (657) Sexual and Reproductive Health (798) Sports Medicine (595) Surgery (799) Toxicology (109) Transplantation (317) Urology (305)
## Related Clinical Research

- [UV-related Tumor Mutational Burden Predicts Melanoma Mortality in Stage II/III InterMEL Cohort](https://medichelpline.com/clinical-feed/medrxiv-11-uv-induced-mutations-are-associated-with-melanoma-specific-mortality-in-stage.md)
- [Foundation-Model Whole-Slide AI plus PRAME IHC for Melanoma Diagnosis](https://medichelpline.com/clinical-feed/medrxiv-19-integrating-histopathology-foundation-models-and-prame-immunohistochemistry-in.md)
- [ESMO recap: Reactions to Merck and Moderna’s mRNA cancer vaccine data](https://medichelpline.com/clinical-feed/stat-news-1-assessing-merck-and-moderna-s-mrna-cancer-vaccine.md)
- [Single‑cell atlas of 189 colorectal tumors links tumor biology to worse outcomes in young patients](https://medichelpline.com/clinical-feed/stat-news-0-stat-new-atlas-of-colorectal-tumors-sheds-light-on-why-cancers-in-young-people.md)
- [Cellular Metabolism, Immunity, and Oncogenesis: Drug Discovery Opportunities](https://medichelpline.com/clinical-feed/frontiers-in-immunology-5-editorial-cellular-metabolism-the-immune-system-and-oncogenesis-opportunities.md)

## Navigation
- [← Back to Oncology Feed](https://medichelpline.com/clinical-feed/oncology.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.