Tumor-infiltrating lymphocytes (TILs) and regression are manifestations of the host immune response to tumor, but their influence on outcome remains undefined. There is a paucity of data on the elderly who represent a growing proportion of melanoma patients. The aim of this study was to evaluate the influence of TILs and regression as an indirect measure of immunity on outcome in elderly patients with melanoma. From a prospective database, we identified 250 consecutive cutaneous melanoma patients aged at least 65 years at the time of diagnosis. Data were verified by record review. Within the primary melanoma, a brisk TIL response was present in 66 (31%), nonbrisk TILs in 36 (17%), and absent in 111 (52%). The presence of a brisk infiltrate conferred a three-fold increased risk of sentinel lymph node (SLN) metastasis (P=0.02). Despite this, nonbrisk or absent TILs were associated with a five-fold increased risk of recurrence (P=0.0001). In multivariate analysis, nonbrisk or absent TILs were independently associated with recurrence (P<0.0001), diminished 5-year disease-free survival (76 vs. 91%, P=0.0006), and 5-year melanoma-specific survival (82 vs. 95%, P=0.0008). Regression was not an independent predictor of SLN metastasis, disease-free survival, or melanoma-specific survival. Our study demonstrates that an active antitumor immune response exists in elderly melanoma patients that, paradoxically, predicts both SLN metastasis and improved melanoma-specific outcomes. Further investigation to characterize this lymphocytic infiltrate and to confirm its clinical significance as a predictor of nodal status, patient outcome, and response to immunotherapy in elderly melanoma patients appears warranted.