Browse by author
Lookup NU author(s): Dr Tom Andrew, Grant Richardson, Professor Philip Sloan, Professor Ruth PlummerORCiD, Professor Penny LovatORCiD
This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).
© The Author(s) 2025.Accurate prognostication is essential to guide clinical management in localised cutaneous melanoma (CM), the form of skin cancer with the highest mortality. While the tumour microenvironment (TME) plays a key role in disease progression, current staging systems rely on limited tumour features and exclude key clinicopathological prognostic features. Here we show that MelanoMAP, a multimodal AI model integrating TME-derived digital biomarkers and clinicopathological features from over 3,500 histology slides, improves prognostication of localised CM. MelanoMAP achieved a C-index of 0.82, a 24% improvement over traditional AJCC staging (0.66) and consistently outperformed clinicopathological-only models across six international patient cohorts. SHAP analysis identified TME-derived digital biomarkers, alongside traditional clinicopathological factors including age, mitotic count, and Breslow depth, were critical determinants of metastatic risk. MelanoMAP establishes a potential foundation for precision oncology in CM, demonstrating how AI-driven digital biomarkers can advance personalised prognostication and inform clinical-decision making.
Author(s): Andrew TW, Combalia M, Hernandez C, Grant S, Paragh G, Puig S, Mc Arthur G, Richardson G, Sloan P, Shalhout SZ, Plummer R, Lovat PE
Publication type: Article
Publication status: Published
Journal: Nature Communications
Year: 2025
Volume: 16
Issue: 1
Online publication date: 18/11/2025
Acceptance date: 01/10/2025
Date deposited: 02/12/2025
ISSN (electronic): 2041-1723
Publisher: Nature Research
URL: https://doi.org/10.1038/s41467-025-65051-0
DOI: 10.1038/s41467-025-65051-0
Data Access Statement: Clinical and histopathological datasets from multiple institutions are under controlled access due to patient privacy, ethics, and institutional agreements. Raw data cannot be deposited publicly. De-identified derived datasets (digital biomarkers, model inputs/outputs, data dictionary) are available upon request to Dr Tom Andrew (tom.an drew@newcastle.ac.uk) for non-commercial academic use, subject to a data use agreement and institutional approval. Requests reviewed within 4–6 weeks; data available for at least 5 years post-publication. Separate Source Data files are provided for each figure/table as individual Excel sheets. Source data are provided with this paper. Code availability Data processing and analyses were conducted using Python programming language. All code used for data preprocessing, feature extraction, model training, and SHAP-based interpretation is publicly available at https://github.com/tomwandrew/MelanoMAP under the MIT License
PubMed id: 41253792
Altmetrics provided by Altmetric
