Abstract
Background and Objectives: In the realm of the rising incidence of cutaneous and mucous melanoma, CDKN2A mutations characterize familial and multiple primary melanoma cases. The involvement of tumor-infiltrating lymphocytes (TILs) is interconnected with survival rates, but may extend even further. The aim of this study is to verify the accuracy of the classical "naked eye" count of CD8-positive T cells comprised within the tumoral population and peritumoral infiltrate versus that obtained via a special software run by the aid of artificial intelligence (AI), used to determine the percentage of CD8-positive TILs. Materials and Methods: The present retrospective cross-sectional study conducted over a period of 5 years (2018-2022) focused on patients diagnosed with mucous and/or cutaneous melanoma, with a positive family history for melanoma, or personal antecedents of primary malignant melanocytic lesions. The 23 selected cases were diagnosed histopathologically, tested for CDKN2A mutations through fluorescent hybridization in situ, and CD8 immunohistochemistry was performed. The included slides were evaluated both manually (naked-eye examination) and automatically (via QuPath platform) for quantifying the CD8-positive TILs. Results: The number of CD8-positive TILs in melanoma samples has been more accurately identified through the use of an AI-mediated software as compared to the human-eye evaluation performed by experimental pathologists. A higher percentage of CD8-positive intratumoral lymphocytes versus stromal lymphocytes was positively associated with more numerous metastatic sites. Conclusions: The CD8 lymphocytic phenotype harbors major significance in the context of familial and multiple primary melanoma and may comprise a cost-effective investigation meant to help in the establishment of melanoma prognosis and response to immunotherapy.