Reinterpreting dermoscopic pigment network with reflectance confocal microscopy for identification of melanoma-specific features

Background: Pigment network is an important dermoscopic feature for melanocytic lesions, but alterations in grid line thickness are also observed in melanomas. Objective: To investigate features of thick, thin and mixed pigment networks at dermoscopy and their respective features at reflectance confocal microscopy (RCM) for differential diagnosis, correlated with histology. Methods: All melanocytic lesions with histological diagnosis, evaluated between January 2010 and May 2014, were enrolled and classified according to dermoscopy evaluation of the pigment networks: thin, thick and mixed. Results: Thin network in melanoma was characterized by a honeycombed pattern (P < 0.001), dendritic cells (P < 0.001), atypical ringed pattern (P = 0.035) and structureless area (P = 0.012), whereas round cells (P < 0.001), dendritic cells (P < 0.001) and atypical meshwork pattern (<0.001) characterized thick network in melanoma. Mixed network type in melanoma shared honeycombed (P = 0.049) and typical ringed patterns (P = 0.045) in the thin area and round cells (P < 0.001) and atypical meshwork pattern (P < 0.001) in the thick area. Thin network in nevi was characterized by cobblestone (P < 0.001) and typical ringed patterns (P = 0.035), whereas thick network in nevi showed a typical meshwork pattern (P < 0.001). Mixed nevi shared the same features and patterns, but more frequently with inflammatory infiltrate (P = 0.047). Conclusion: Differential diagnosis between melanocytic lesions (nevi or melanoma) in thin, thick and mixed pigment networks observed at dermoscopy can be assisted by RCM to improve diagnostic accuracy.