Combining RNA tape stripping with dermoscopic features for melanoma identification

Abstract

Background

A noninvasive, accurate diagnostic method for melanoma is needed to improve early detection and decrease number of excisions. Gene expression as measured by RNA levels uses tape strips to directly assesses genetic markers from skin lesions. Commercial RNA assays show varying diagnostic accuracy in melanoma detection. Combining dermoscopic features with RNA tape strips could reduce unnecessary removal of benign lesions by increasing specificity.

Objectives

To test an RNA-based rule-out test and integrate dermoscopic features with RNA analysis to improve diagnostic accuracy of malignant melanoma (MM). Furthermore, we explore the association between RNA profiles and dermoscopic features.

Methods

Seventy patients with pigmented skin lesions suspected of being melanoma were imaged with dermoscopy and tape stripped for RNA analysis before surgical excision. The images were evaluated for seven dermoscopic features, and RNA levels of 11 genes were analyzed. A combined test using both gene expression and dermoscopic features was developed. Associations between RNA profiles and dermoscopic features were explored.

Results

Histopathology revealed 19 malignant lesions (17 MM and two basal cell carcinomas) and 51 benign lesions. The combined dermoscopy and RNA test identified all malignant lesions (100% sensitivity) based on PRAME expression, blotch and regression structures and patient age. This combined model increased specificity to 35%, compared to 24% with the original RNA rule-out test, without missing any malignant lesions. Significant differences in RNA profiles were observed for lesions expressing atypical network and regression structures.

Conclusions

Combining RNA tape stripping with dermoscopic features can reduce the removal of benign lesions by over one-third while maintaining 100% sensitivity. We found specific RNA profiles to be strongly associated with dermoscopic features, presenting a promising opportunity to integrate molecular and morphological information and provide valuable guidance for dermatologists managing atypical pigmented lesions.