LesionTABE: Equitable AI for Skin Lesion Detection

TL;DR

LesionTABE is a fairness-focused AI framework for skin lesion detection that improves diagnostic equity across skin tones while maintaining high accuracy, using adversarial debiasing and foundation model embeddings.

Contribution

This work introduces LesionTABE, a novel framework combining adversarial debiasing with dermatology-specific foundation models to address bias in skin lesion AI diagnostics.

Findings

Over 25% improvement in fairness metrics.

Outperforms existing debiasing methods.

Enhances overall diagnostic accuracy.

Abstract

Bias remains a major barrier to the clinical adoption of AI in dermatology, as diagnostic models underperform on darker skin tones. We present LesionTABE, a fairness-centric framework that couples adversarial debiasing with dermatology-specific foundation model embeddings. Evaluated across multiple datasets covering both malignant and inflammatory conditions, LesionTABE achieves over a 25\% improvement in fairness metrics compared to a ResNet-152 baseline, outperforming existing debiasing methods while simultaneously enhancing overall diagnostic accuracy. These results highlight the potential of foundation model debiasing as a step towards equitable clinical AI adoption.