Efficacy of adjuvant radiotherapy in recurrent melanoma after adjuvant immunotherapy.

Abstract

9578 Background: Adjuvant (adj) radiotherapy (RT) halves the risk of locoregional (LR) recurrence in patients (pts) with high risk stage III melanoma after lymphadenectomy (CLND), however its role in the adj immunotherapy (IO) era without CLND is unknown. Methods: Pts with resected stage III melanoma who received adj IO and recurred with resectable LR only disease were studied. After resection of this 1st recurrence, adj RT may or may not have been administered. Disease characteristics, treatment at relapse and outcomes were examined. Results: 71 pts from 9 centres were included. Prior to adj IO, median age was 60y, 59% male, 56% BRAF mutant, 61% stage IIIC (AJCC V8), 52% underwent CLND and 17% had in-transit (IT) only disease. Adj IO included: 90% single agent anti-PD1, 8% ipilimumab-nivolumab (IN) and 1% nivolumab or IN (blinded on trial). Median duration of adj IO was 5 months. 21(30%) pts had high risk stage III disease at diagnosis, per previously established TROG criteria; 3 (4%) received upfront adj RT prior to recurrence. Median time to 1st recurrence was 7 months. 49 (69%) pts recurred during and 22 (31%) after cessation of adj IO. At 1st recurrence, 9 (13%) pts had stage IIIB disease, 55 (77%) IIIC, 7 (10%) IIID and 8 (11%) continued prior adj IO, 31 (44%) commenced therapy and 32 (45%) had no systemic therapy. 24 (34%) pts received adj RT after resection of 1st recurrence and 47 (66%) did not (Table). Adj RT was associated with a reduced risk of any 2nd recurrence (7/24, 29% vs 26/47, 55%, p=0.03) and LR 2nd recurrence (2/24, 8% vs 17/47, 36%, p=0.012). Whilst pts who received adj RT at 1st recurrence were more likely to have LN only disease, extra nodal extension and involved surgical margins, these factors did not significantly affect overall risk of 2nd recurrence on multivariate analysis. Of note, 70% of pts who did not receive adj RT at 1st recurrence had IT only disease, and though this did not significantly affect rate of 2nd recurrence (p=0.19), this likely reflects an inherent selection bias in this study. RT toxicity occurred in 16 (67%) pts, 10 with dermatitis only, and all grade 1 or 2. Median follow up was 22 months. Median recurrence free survival to 2nd recurrence was 23 months for all pts, not reached for those who had adj RT at 1st recurrence and 19 months for those who did not have adj RT (p=0.047). Median overall survival was not reached. Conclusions: Whilst adj RT appears to reduce 2nd recurrences, this may have been influenced by an unavoidable selection bias in the data, particularly an imbalance in the percentage of pts with IT disease. Prospective data with larger cohorts is needed to validate our results.[Table: see text]