A MULTIOBSERVER, POPULATION-BASED ANALYSIS OF HISTOLOGIC DYSPLASIA INMELANOCYTIC NEVI

Background: Nevi that are clinically atypical and histologically dyspl astic have been associated with increased melanoma risk. There are few reproducibility studies or population-based studies of nevus histolog y. Objective: Our purpose was to quantify concordance in histologic di agnosis of melanocytic lesions among a diverse group of pathologists, to assess intraobserver concordance by comparing readings of the same slide as well as of adjacent recuts from the same block, to correlate histology with nevus appearance and melanoma risk, and to estimate the range of prevalence of histologic dysplasia. Methods: Histologic slid es were prepared from 149 tissue blocks of pigmented lesions from mela noma cases, relatives, and controls. Six dermatopathologists independe ntly evaluated the lesions for histologic dysplasia, without prior agr eement on criteria. Results: According to h statistics, intraobserver reproducibility was substantial, and interobserver concordance was fai r, despite differences in criteria. The estimated prevalences of histo logic dysplasia for the six pathologists ranged from 7% to 32%. Histol ogic dysplasia was correlated with nevus size for most observers, conf ounding the observed correlation between nevus appearance and histolog y. Conclusion: Although experienced dermatopathologists use different diagnostic criteria for histologic dysplasia, their usage is consisten t. Histologic changes ascribed to melanocytic dysplasia are prevalent in the white population for all pathologists. The term nevus with hist ologic dysplasia should be used in preference to dysplastic nevus.