Distribution pattern of tumor infiltrating lymphocytes and tumor microenvironment composition as prognostic indicators in anorectal malignant melanoma

Abstract

Anorectal malignant melanoma (ARMM) is a rare disease with poor prognosis. Determining ARMM prognosis precisely is difficult due to the lack of proper assessment techniques. Immunotherapy has proven effective against cutaneous malignant melanoma and may show efficacy in ARMM. Herein, we assessed the immune profile of ARMM to identify possible prognostic biomarkers. Twenty-two ARMM formalin-fixed and paraffin-embedded samples were evaluated using an nCounter® PanCancer Immune Profiling Panel. Validation was performed through immunohistochemical staining for CD3, CD8, Foxp3, CD68, CD163, and PD-L1. RNA analysis revealed significantly decreased scores for pathways involved in cell regulation and function, as well as chemokines, in recurrent patients compared to nonrecurrent patients. In cell-type profiling, the recurrent cases displayed significantly low tumor infiltrating lymphocyte (TIL) scores. Recurrence/death prediction models were defined using logistic regression and showed significantly lower scores in recurrent and deceased patients (all, P < 0.001) compared to those in nonrecurrent and surviving patients. The high total TIL and tumor-associated macrophage (TAM) groups had significantly better overall survival outcomes compared to the low total TIL and TAM groups (P = 0.007 and P = 0.035, respectively). In addition, the presence of CD3 + TILs in the invasion front was an independent favorable prognostic indicator (P = 0.003, hazard ratio = 0.21, 95% confidential interval, 0.01–0.41). Patients with inflamed or brisk-infiltration type tumors also had a significantly better overall survival than that of patients with immune-desert/excluded and absent/non-brisk type tumors (P = 0.03 and P = 0.0023, respectively). In conclusion, TILs have a strong prognostic value in ARMM, and the quantification of TILs and an analysis of the TIL phenotype and infiltration pattern during pathological diagnosis are essential to guide treatment strategies and accurate prognosis in ARMM.