Presence of histological regression as a prognostic factor in cutaneous melanoma patients

Regression is caused by a host immunological response primarily characterized by lymphocytic infiltration directed against melanoma cells. The prognostic significance of regression remains controversial in cutaneous melanoma patients. The aim of this study was to determine the clinical significance of the histological regression status in patients with cutaneous melanoma. A total of 664 patients with a pathologically confirmed cutaneous melanoma were enrolled into this study and were investigated retrospectively. The median age of the patients was 51 years, ranging in age from 16 to 104 years. The majority of them had lesions without regression (n=495; 74.5%) and others had lesions with regression (n=169; 25.5%). Melanoma patients with regression were more frequently males (60.1 vs 51.7%; P=0.038) and had axial localized lesions (67.5 vs 53.7%; P=0.002), superficial spreading histologic subtype (73.2 vs 49.1%; P=0.000), thin Breslow depth (<2mm) (44.6 vs 33.5%; P=0.01), and the presence of tumor-infiltrating lymphocytes (74.4 vs 60.0%; P=0.001) than those without regression. However, regression was not significantly associated with age, Clark level, mitotic rate, ulceration, vertical growth phase, neurotropism, lymphovascular invasion, nor association with a pre-existing melanocytic nevus. Similarly, no significant correlations were found between regression and lymph node involvement, recurrence, nor metastasis of disease. Patients with, nodular pathology, advanced Clark invasion level (IV-V), thick Breslow depth (2mm), high mitotic rate (>3/mm(2)), ulceration, vertical growth phase, neurotropism, lymphovascular invasion, lymph node involvement, metastasis, and recurrence of disease, and male patients had poor prognostic variables for both relapse-free survival and overall survival. However, the presence of regression was not associated with relapse-free survival (P=0.093) nor overall survival (P=0.113) similar to other factors such as age, tumor localization, tumor-infiltrating lymphocytes, and association with a pre-existing melanocytic nevus. Similar insignificant P values were also observed in multivariate analyses (P=0.115 and 0.816, respectively). In conclusion, the presence of histological regression plays no prognostic role in nodal involvement nor survival in patients with cutaneous melanoma.