risk of tumor necrosis and toxicity resulting from virus reactivation.7 Second, autologous cytotoxic T cells expanded against EBV antigens have been tested, but the very frequent expression of programmed death ligand 1 (PDL1) and immunosuppressive cytokines by NK/T lymphoma cells could inhibit T-cell cytotoxicity. 8,9 Indeed, in most cases, an immune deficiency is not observed in patients with NK/T-cell lymphoma, strongly suggesting that tumor cells have developed strategies to escape the strong immune response usually developed against EBV antigen-expressing cells. This expression of PDL1 by NK/T lymphoma cells is likely to be involved in this resistance, as seen in other EBV-associated neoplasia such as nasopharyngeal lesions, gastric carcinoma, and Hodgkin disease. This combination of frequent PDL1 expression and presence of foreign antigen expression on tumor cells makes the use of checkpoint inhibitors very attractive. Kwong and colleagues, on behalf of the Asia Lymphoma Study Group, demonstrated in this small series of patients the effectiveness of pembrolizumab, an anti-PD1 antibody, in patients with NK/T-cell lymphoma. The 7 patients, previously treated using asparaginasebased regimens, all had advanced disease, with 5 patients having systemic hemophagocytic syndrome. All patients experienced a rapid response to pembrolizumab, with 5 complete responses (median follow-up, 6 months). Three of 4 patients with strong PDL1 expression achieved complete remission. Interestingly, residual lesions were biopsied after pembrolizumab treatment in 3 patients, and immunohistochemical staining showed that the infiltrating lymphoid cells were predominantly CD31 T cells, with a mix of CD4 1 and CD8 1 and only a minority of CD56 1 EBER 1 cells. This finding is consistent with the hypothesis that pembrolizumab treatment allows T cells to recognize and kill EBV-infected NK/T lymphoma cells. The outcome of NK/T-cell lymphoma patients with relapsing disseminated diseases after asparaginase-based regimens is dismal. The results of anti-inhibitory receptor programmed death 1 treatment in this very small series with short follow-up supports the use of this treatment in compassionate use programs. Clinical trials should be performed to extend the indications of checkpoint inhibitors in NK/T-cell lymphoma. However, it is still not known if these responses will be observed in all patients and what durations will be. As shown in this small group of patients, high expression of PDL1 might be a very strong predictor of response. This report, however, raises several questions. First, will we use immune checkpoint inhibitors in first-line treatment of NK/T-cell lymphoma? Second, will we be able in the near future to avoid radiotherapy and its deleterious side effects? Third, will we use combination therapy (see figure) to prevent relapses and selection of resistant cells? L-asparaginase, especially pegylated forms that are less immunogenic, will probably remain an important component to reduce tumor bulk without induction of a significa...