Correlation between clinical atypia and histologic dysplasia in acquired melanocytic nevi

Background: The validity of clinical and histologic criteria in identifying dysplastic nevi is controversial. Recognition of the dysplastic nevus as a distinct clinicopathologic entity requires demonstration of significant agreement between clinical atypia and histologic dysplasia. Objective: We attempted to determine the correlation between clinical atypia and histologic dysplasia in acquired melanocytic nevi and to evaluate the sensitivity and specificity of clinical criteria for dysplastic nevi when compared with histopathologic features. Methods: A total of 940 acquired melanocytic nevi 3 mm in diameter or larger were selected by initially choosing clinically unequivocal dysplastic and nondysplastic nevi and then, from these, histologically unequivocal dysplastic and nondysplastic lesions. The level of concordance between clinical atypia and histologic dysplasia was estimated by κ statistics. Results: Nevi were classified as clinically dysplastic (n = 499) or nondysplastic (n = 441). On the basis of histologic features, 739 were classified as dysplastic and 201 as nondysplastic. Agreement between clinical atypia and histologic dysplasia was found in 432 nevi, that is, a sensitivity of 58.4% (3-5 mm = 27.2%, >5 mm = 69.8%). Agreement between clinical and histologic criteria on the absence of dysplasia was found in 134 nevi, a specificity of 66.6% (3-5 mm = 92.4%, >5 mm = 47.9%). The κ value was 0.17 (3-5 mm = 0.14, >5 mm = 0.10). Conclusion: The limited sensitivity and specificity together with the negligible κ value indicate a poor agreement between clinical and histologic diagnoses of dysplastic nevus. The dysplastic nevus cannot be considered a distinct clinicopathologic entity because histologic dysplasia is found in a range of nevi that may or may not show clinical atypia.