Discontinuation of anti-PD1 in advanced melanoma: an observational retrospective study from the Italian Melanoma Intergroup

Background: Immunotherapy has improved the survival of patients with stage IV melanoma. In responders, clinical benefits may be long-lasting and persist even after treatment discontinuation. The optimal duration of anti-PD1 (anti-Programmed cell death-1) therapy in metastatic melanoma patients remains to be elucidated. Moreover, limited data are available on clinical outcomes of patients that discontinued anti-PD1 immunotherapy in a real-life setting. The aim of this study was to evaluate the progression-free survival (PFS) in patients with metastatic melanoma who interrupted anti-PD-1 treatment in the in the absence of disease progression. Methods: We retrospectively reviewed patients with advanced/metastatic melanoma treated with anti-PD1 immunotherapy at 23 Italian Melanoma Intergroup (IMI) centres. The study investigated the risk of relapse in patients who stopped anti-PD1 therapy due to CR (Complete response), treatment-related toxicity, or by their own choice after a long period of treatment. Clinical and biological factors associated with or without recurrence were evaluated. Results: The study population included 237 patients. The median age of patients was 68.9 years (standard deviation: 13; range 33-95). The median time on treatment was 33 months (standard deviation: 18, 7; range 1-98). Among the 237 patients, 128 (54%) interrupted the anti-PD1 for CR, 74 patients (31.2%) for adverse events (37 patients in CR, 27 patients in partial response (PR), ten patients in stable disease (SD), and 35 patients (14.8%) by their own choice (12 patients in CR, 17 patients in PR, and 6 patients in SD). After a mean follow-up of 21 months (range 1-81), PFS after anti-PD1 discontinuation was 85.7%. Thirty-four patients (14.3%) developed disease progression after a median of 12 months (range 1-35): ten patients (29.4%) after discontinuation in CR, 17 patients (50%) after discontinuation for treatment-related toxicity (seven in CR, five in PR, five in SD), and seven (20.6%) after discontinuation due to the patient's decision (two in CR, four in PR, one in SD). Only 7.8% of patients who interrupted in CR (10/128), along with 23% of patients who interrupted for limiting toxicity (17/74) and 20% of patients who interrupted by their own choice (7/35), developed recurrence. Regarding patients who discontinued therapy because of CR, we observed a negative association between recurrence and site of primary melanoma, especially mucosal sites (p = <0.05, HR (Hazard ratio) 15.57 IC (confidence interval) 95% 2.64-91.73). Moreover, M1b patients who achieved a CR showed a lower number of relapses (p = <0.05, HR 3.84 IC 95% 1.40-8.48). Conclusions: This study shows in a real-life setting that, with anti-PD-1 therapy, long-lasting responses, can be maintained after anti-PD1 interruption. In 70.6% of cases, recurrences were observed among patients who did not obtain a CR at treatment discontinuation