Multi-Dataset Diagnostic Utility of Clinical Visual Concepts in AI Systems for Dermatology

2026-10-06

AI Summary

This study addresses the limited interpretability and lack of clinical trust in deploying AI for dermatology, which stem from heterogeneous annotations. To overcome this, we introduce SkinLex, the first unified dataset integrating morphological attributes across four major datasets. Through multi-source data fusion, supervised classification, and a bootstrap backward elimination algorithm, we systematically analyze the diagnostic utility of 48 visual concepts for nine skin diseases, revealing concept redundancy and establishing a minimal effective concept set. Our findings demonstrate that employing this streamlined set of concept inputs not only improves modeling efficiency but also significantly enhances diagnostic robustness and clinical interpretability. The source code is publicly available.

Abstract

The clinical integration of AI systems in digital dermatology relies heavily on human trust. Clinically interpretable visual concepts can act as intermediate representations enhancing trust and reliability. However, research in this domain is currently limited by scattered, heterogeneous dataset annotations. In this work, we introduce SkinLex, a harmonized dataset of 48 clinical morphological attributes across four public datasets (SkinCon, DermaCon-IN, MM-Skin, and PASSION) for a total of 20,411 records. Supervised nine-partition classification of skin conditions shows that limiting features to specific visual groups, like shapes or colors alone, reduces diagnostic accuracy. Bootstrapped backward elimination reveals that the set of 48 visual concepts has some degree of redundancy for algorithmic nine-partition diagnosis on the examined dataset. This demonstrates that coarse diagnosis on the selected dataset requires a relatively small but varied combination of clinical concepts, and motivates further research to improve concept taxonomy. Results can be translated into clinical benefits by reducing inputs for concept-based models, improving efficiency for annotation and modeling, and further enhancing interpretability. Code and prompt templates are available at https://github.com/Digital-Dermatology/SkinLex.

Problem

• dermatology AI
• clinical visual concepts
• dataset harmonization
• diagnostic utility
• interpretability

Innovation

• harmonized dataset
• clinical visual concepts
• concept redundancy
• backward elimination
• interpretability