Multi‐task AI models in dermatology: Overcoming critical clinical translation challenges for enhanced skin lesion diagnosis
Journal of the European Academy of Dermatology and Venereology · 2025-12 · 39 (12) · 2121-2133
DOI: 10.1111/jdv.20551 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the European Academy of Dermatology and Venereology, online erschienen am 07.02.2025. Verfasst von Deval Mehta und 11 weiteren Autor:innen. Beteiligte Einrichtung: AIM for Health Lab, Faculty of IT, Monash University, Melbourne, Victoria, Australia. Hauptthemen laut MEDLINE: Artificial Intelligence, Dermatology, Skin Neoplasms, Skin Diseases. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: National Health and Medical Research Council (NHMRC) Synergy Grant: Roadmap Options for Melanoma Screening in Australia, National Health and Medical Research Council (NHMRC) Grant: Skin Imaging and Precision Diagnosis, National Health and Medical Research Council.
Abstract (Originalsprache)
Background
The surge in AI models for diagnosing skin lesions through image analysis is notable, yet their clinical implementation faces challenges. Common limitations include an over reliance on dermoscopy, lack of real‐world applicability when only binary output (e.g. benign/malignant) is offered and low accuracy when faced with rare skin conditions.
Objectives
To address these common constraints associated with limited diagnostic output, and applicability to real‐world settings.
Methods
We developed an All‐In‐One H ierarchical‐ O ut of Distribution‐Clinical T riage ( HOT ) AI model for skin lesion analysis. Trained on a large dataset of ~208,000 lesion images, our HOT AI model generates three outputs: a hierarchical three‐level prediction, an alert for out‐of‐distribution (OOD) images and a recommendation for dermoscopy to improve diagnostic prediction.
Results
Our hierarchical prediction output provides a binary level 1 prediction (benign/malignant), Level 2 prediction of eight possible categories (e.g. melanocytic and keratinocytic) and a more definitive Level 3 prediction from 44 lesion categories. The model produced high sensitivity for Level 1 prediction (88.14% CI: 87.42–88.51); however, significantly lower for Level 3 prediction (63.90%, CI: 62.27–65.61). By relying on all three prediction levels for consensus, Level 1 false‐positives were reduced by 20–25%, and false‐negatives were decreased by 11–13% of cases. OOD detection was benchmarked against previous landmark models and outperformed comparative models. Lastly, 44% of images were recommended for dermoscopy, and with additional image input, Level 3 sensitivity increased from 48.13% (CI:45.08–49.57) to 52.54% (CI:50.25–55.04).
Conclusions
Our HOT‐AI model attempts to address common challenges in existing models by combining three tasks in one model to increase accuracy and clinical utility. By providing a more nuanced prediction, and alert for OOD, the model output provides greater explainability of the AI decision process. Prospective clinical testing is required to measure how this additional output impacts user trust, and how the model performs in a real‐world setting.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- AIM for Health Lab, Faculty of IT, Monash University, Melbourne, Victoria, Australia.
- Dermatology Research Centre, Frazer Institute, the University of Queensland, Brisbane, Queensland, Australia.
- E-Research Center, Monash University, Melbourne, Victoria, Australia.
- Molemap Pty. Ltd., Melbourne, Victoria, Australia.
- Australian Centre for Blood Diseases, Monash University, Melbourne, Victoria, Australia.
- The University of Queensland, Brisbane, Queensland, Australia.
ORCID-Profile: Deval Mehta, Clare Primiero, Brigid Betz-Stablein
Förderung
- National Health and Medical Research Council (NHMRC) Synergy Grant: Roadmap Options for Melanoma Screening in Australia
- National Health and Medical Research Council (NHMRC) Grant: Skin Imaging and Precision Diagnosis
- National Health and Medical Research Council
Zitieren
APA
Mehta, D., Primiero, C., Betz-Stablein, B., Nguyen, T. D., Gal, Y., Bowling, A., Haskett, M., Sashindranath, M., Bonnington, P., Mar, V., Soyer, H. P., & Ge, Z. (2025). Multi‐task AI models in dermatology: Overcoming critical clinical translation challenges for enhanced skin lesion diagnosis. Journal of the European Academy of Dermatology and Venereology, 39(12), 2121-2133. https://doi.org/10.1111/jdv.20551
Vancouver
Mehta D, Primiero C, Betz-Stablein B, Nguyen TD, Gal Y, Bowling A, et al. Multi‐task AI models in dermatology: Overcoming critical clinical translation challenges for enhanced skin lesion diagnosis. Journal of the European Academy of Dermatology and Venereology. 2025;39(12):2121-2133. doi:10.1111/jdv.20551
BibTeX
@article{Mehta2025,
author = {Mehta, Deval and Primiero, Clare and Betz-Stablein, Brigid and Nguyen, Toan D and Gal, Yaniv and Bowling, Adrian and Haskett, Martin and Sashindranath, Maithili and Bonnington, Paul and Mar, Victoria and Soyer, H Peter and Ge, Zongyuan},
title = {Multi‐task AI models in dermatology: Overcoming critical clinical translation challenges for enhanced skin lesion diagnosis},
journal = {Journal of the European Academy of Dermatology and Venereology},
year = {2025},
volume = {39},
number = {12},
pages = {2121--2133},
doi = {10.1111/jdv.20551}
}
EUVIMED