PDF file - 178KB, Melanoma. (a) Clinical image. (b) Histology image of the lesion showing pagetoid spread in the upper epidermal layers (arrows). (c-e) MPM images of epidermal layers showing advanced melanocytic dendrites (red arrows) and pleomorphic cells/pagetoid spread (yellow arrows) in upper epidermal layers at depths of 40 microm (c), 45 microm, (d) and 50 microm (e). Scale bar is 20 microm.
ARTICLE ABSTRACT
Monitoring of atypical nevi is an important step in early detection of melanoma, a clinical imperative in preventing the disease progression. Current standard diagnosis is based on biopsy and histopathologic examination, a method that is invasive and highly dependent upon physician experience. In this work, we used a clinical multiphoton microscope to image in vivo and noninvasively melanocytic nevi at three different stages: common nevi without dysplastic changes, dysplastic nevi with structural and architectural atypia, and melanoma. We analyzed multiphoton microscopy (MPM) images corresponding to 15 lesions (five in each group) both qualitatively and quantitatively. For the qualitative analysis, we identified the morphologic features characteristic of each group. MPM images corresponding to dysplastic nevi and melanoma were compared with standard histopathology to determine correlations between tissue constituents and morphology and to evaluate whether standard histopathology criteria can be identified in the MPM images. Prominent qualitative correlations included the morphology of epidermal keratinocytes, the appearance of nests of nevus cells surrounded by collagen fibers, and the structure of the epidermal–dermal junction. For the quantitative analysis, we defined a numerical multiphoton melanoma index (MMI) based on three-dimensional in vivo image analysis that scores signals derived from two-photon excited fluorescence, second harmonic generation, and melanocyte morphology features on a continuous 9-point scale. Indices corresponding to common nevi (0–1), dysplastic nevi (1–4), and melanoma (5–8) were significantly different (P < 0.05), suggesting the potential of the method to distinguish between melanocytic nevi in vivo. Cancer Res; 74(10); 2688–97. ©2014 AACR.
