Diagnostic Algorithm for Conjunctival Melanocytic Lesions.

Abstract

This study aimed to develop an algorithm allowing the differentiation between conjunctival melanoma and other melanocytic infiltrations of the conjunctiva, on the basis of a dermatoscopic examination. A total of 160 conjunctival pigmented lesions were studied (40 melanomas and 120 non-melanoma conjunctival infiltrations). The clinical characteristics of the tumours were assessed with the use of dermatoscopic characteristics as described by Kittler, and with taking into consideration the typical characteristics of conjunctival lesions. On the basis of modified dermatoscopic criteria, an algorithm was generated consisting of an assessment of the presence of 9 suspicious characteristics, e.g. more than two colours, colour asymmetry, pattern asymmetry, vascular polymorphism, presence of short vessels, linear vascular pattern, the presence of a peripheral structureless area, the presence of a grey structureless area and black dots in any part of the lesion. The presence of any of these characteristics scores 1 point. If a melanocytic lesion scores ≥3 points, the probability of diagnosing melanoma is on the level of p>0.001. The use of the proposed algorithm, based on modified dermatoscopic characteristics, may be a valuable tool for the diagnosis of conjunctival melanoma.