ABSTRACT
ABSTRACT: Mycosis fungoides (MF) has become one of the most difficult diagnostic challenges for both dermatologists and dermatopathologists because its clinical presentation and microscopic findings may mimic benign reactive processes, specifically those displaying histopathological features of interface dermatitis. The goal of our study was to prove with digital scanning and automated sample methodology through algorithmic analysis, combined with the utility of TOX marker a more precise, faster, and objective evaluation of each sample. Moreover, this would offer high levels of reproducibility with the possibility of establishing cut-off points, allowing us to distinguish between inflammatory dermatoses (ID) and MF. A retrospective longitudinal-descriptive and observational study was conducted to compare the diagnostic criteria (immunohistochemical studies of anti-TOX stain) in patients with clinical suspicion of MF by dividing them into 2 groups: samples with a positive biopsy for MF (MF group) and those with a negative biopsy, therefore diagnosed as an ID (control group). The algorithm assessed 5 selected areas with lymphocytic representative cellularity, and based on the intensity, nuclear staining was classified as 0 (negative), 1+ (weak/yellow), 2+ (moderate/orange), and 3+ (strong/scarlet red) nuclei. The results showed statistically significant differences ( P = 0.040) between the mean number of (2+) nuclei in the positive final diagnosis group (MF group) and the negative final diagnosis group (ID group).