Abstract
BACKGROUND: Plasma Epstein-Barr virus (EBV) DNA has clinical utility for prognosis, recurrence, surveillance, and treatment response in nasopharyngeal carcinoma (NPC). This exploratory analysis evaluated associations between plasma EBV DNA load and clinical outcomes in participants treated with pembrolizumab or chemotherapy in the phase 3 KEYNOTE-122 trial (NCT02611960). METHODS: Participants with platinum-pretreated, histologically confirmed, EBV-positive, recurrent/metastatic NPC were randomly assigned (1:1) to pembrolizumab 200 mg intravenously every 3 weeks (≤ 35 cycles) or standard of care (SOC; investigator's choice of capecitabine, gemcitabine, or docetaxel). Associations between baseline plasma EBV DNA load as a continuous variable and plasma EBV DNA fold change at cycle 2 day 1 (C2D1), with clinical outcomes (progression-free survival [PFS], overall survival [OS], and objective response rate [ORR]) were evaluated within each treatment arm. Nominal significance was prespecified at 0.05 for 1-sided p values. RESULTS: Of 228 treated participants, 215 (94.3%) had evaluable baseline plasma EBV DNA load data (pembrolizumab, 111; SOC, 104). Baseline plasma EBV DNA load was negatively associated with PFS and OS for pembrolizumab and SOC (both p < 0.005) but not ORR (p = 0.105, pembrolizumab; p = 0.473, SOC). Larger decreases in plasma EBV DNA load at C2D1 relative to baseline were associated with improved PFS, OS, and ORR for pembrolizumab and SOC (p ≤ 0.005). CONCLUSIONS: Higher baseline plasma EBV DNA load was negatively associated with outcomes in participants with NPC treated with pembrolizumab or SOC. These findings provide additional support for plasma EBV DNA as a prognostic biomarker for NPC. TRIAL REGISTRATION: ClinicalTrials.gov, NCT02611960.