HISTOPATHOLOGICAL ANALYSIS OF METASTATIC MELANOMA DEPOSITS IN PATIENTS RECEIVING ADOPTIVE IMMUNOTHERAPY WITH TUMOR-INFILTRATING LYMPHOCYTES

Tumor-infiltrating lymphocytes (TIL) from a wide range of human and mu rine tumors can be expanded in vitro using interleukin-2 (IL-2). These TIL are cytolytic T lymphocytes with in vivo and in vitro antitumor a ctivity in mice and in humans. TIL from human melanoma can recognize a utologous tumor in an MHC-restricted fashion, localize in vivo after I n-111 labeling, and mediate regression of large metastatic deposits. A lthough studied extensively in vitro, less is known in vivo about TIL activity associated with tumor regression. This study was undertaken, in association with a study of TIL localization, to investigate mechan isms of TIL action by evaluating histopathological changes that occur at the tumor site during TIL administration. A total of 106 pre- and p ost-treatment pathological specimens from 25 patients enrolled in phas e II TIL treatment and In-111-TIL. imaging protocols were examined bli ndly by a single pathologist. Histological subtype, lymphocytic infilt ration, melanin content, vascularity, and necrosis were documented for each tumor specimen. Average baseline and post-treatment parameters w ere compared. Any significant changes were evaluated for correlation w ith clinical response and In-111-TIL localization to tumor. Melanin co ntent and vascularity of the tumor did not change as a result of thera py or correlate with either response or TIL localization. However, bot h increased lymphocytic infiltration and tumor necrosis were present a fter TIL administration (P = 0.044 and 0.032 respectively). Furthermor e, increases in lymphocytic infiltration correlated with tumor imaging using In-111-TIL, and with the percentage of In-111-labeled injectate present per gram of tumor specimen (P = 0.036 and 0.0041 respectively ). This suggests that TIL either account for the increased lymphocytes directly, or localize to tumor and recruit endogenous lymphocytes. We were unable to demonstrate any pretreatment histopathological predict ors of response or variables that significantly correlated with subseq uent clinical response, although peak and average values of necrosis w ere higher in responding patients compared to non-responding patients.