Background: Tumor infiltrating lymphocytes (TIL) and histological regression in primary melanoma are generally considered indicators of the local immune response but their roles as prognostic factors have been variably reported. We examined the prognostic role of these variables in patients with high risk (T4) primary melanomas in a large series of patients with long-term follow-up.
Methods: From a prospectively maintained cohort of patients diagnosed between 1971 and 2004, 161 patients were retrospectively identified with primary thick melanomas (>4 mm), no clinical evidence of regional nodal disease (RND) at diagnosis and complete histopathologic data. Univariate and multivariate Cox regression models were performed to identify clinical and histopathologic predictors of disease-specific survival (DSS) and to identify subgroups with differential survival.
Results: Factors significantly associated with decreased DSS by univariate analysis included male gender, age ≥ 60 years, axial anatomic location, presence of ulceration, RND, absence of TIL, and presence of regression. In the final multivariate model, TIL and regression, as interacting variables, and RND status remained significantly associated with DSS. In the presence of TIL, concomitant regression was associated with significantly worse survival (p ≤ 0.0001). In the absence of TIL, there was no effect of regression on survival (p = 0.324).
Conclusions: Primary TIL and regression status and RND status are independently associated with melanoma-specific survival in patients with T4 melanomas; presence of TIL in the primary melanoma with concomitant radial growth phase regression is associated with a poor prognosis and may reflect an ineffective local regional immune response.