Prospective Lesion-Level Analysis of the Inverse Approach for Lentigo Maligna in Real-World Clinical Practice.

INTRODUCTION

Lentigo maligna (LM) is a melanoma subtype that develops predominantly on photodamaged skin. Classic criteria may miss early lesions due to subtle or partially present malignancy-associated features or their overlap with benign or premalignant conditions related to photodamage. The recently introduced inverse approach, which focuses on the absence of typical non-melanoma dermoscopic features, has shown promise in identifying these challenging cases.

OBJECTIVES

This study prospectively evaluated the real-life performance of the inverse approach for differentiating LM from its mimickers and assessed its ability to detect lesions lacking classic criteria.

METHODS

This prospective study enrolled consecutive patients presenting with pigmented flat lesion(s) on the head and neck region. Each lesion was photographed and independently assessed by two dermatologists using classic and inverse approach criteria. Lesions showing at least one classic LM criterion or that did not predominantly (≥50% of the surface area) display any inverse approach feature were considered suspicious for LM and underwent biopsy for histopathological evaluation.

RESULTS

A total of 956 lesions from 68 patients were initially evaluated. Of these, 795 (83.2%) were diagnosed as solar lentigo or flat seborrheic keratosis and 145 (15.2%) as actinic keratosis. Classic criteria for lentigo maligna identified four lesions as suspicious, all of which were histopathologically confirmed; three additional lesions were suspicious only by the inverse approach and were likewise confirmed as LM.

CONCLUSIONS

In this real-life cohort, the inverse approach identified all observed LM cases, including those lacking classic criteria or showing only subtle classic criteria; this approach may serve as a practical adjunct for evaluating pigmented facial lesions. Larger multicenter studies with systematic verification are needed for broader validation.