Analysis of tumour hyperprogression (HP) with nivolumab (Nivo) in randomized, placebo (Pbo)-controlled trials

Abstract

Background: Tumor HP has been suggested to occur in some patients with solid tumors subsequent to PD-1/L1 inhibitor monotherapy; however, reports are based on nonrandomized, single-arm studies. Previous post-hoc analysis of HP with nivo was performed in the phase III ATTRACTION-2 trial, which randomized patients with gastric cancer who received ≥2 prior lines of therapy to nivo or pbo. Using the pbo arm as a surrogate for natural course of disease progression, nivo was not associated with HP at≥20,≥50, or≥100% tumor growth rates. The current analysis expands this framework to assess HP in patients with extensive disease small cell lung cancer (ED SCLC) in the randomized, pbo-controlled CheckMate 451 trial. Methods: CheckMate 451 was a phase III double-blind study that evaluated maintenance treatment in patients with ED SCLC without disease progression after platinumbased 1L chemotherapy. Patients were randomized to nivo 240 mg, nivo 1 mg/kg + ipilimumab 3 mg/kg, or pbo, within 9 weeks from the last dose of 1L chemotherapy (or≥11 weeks for those receiving PCI or whole brain radiotherapy); HP analysis focused on the nivo (n=280) and pbo (n=275) arms. Tumor assessments occurred every 6 weeks for the first 36 weeks, then every 12 weeks until disease progression. HP was calculated as changes in size of the primary lesion (sum of longest diameters [SLD]), as assessed retrospectively among patients who had baseline and≥1 on-treatment tumor measurement available (nivo, n=177; pbo, n=175). Results: Baseline characteristics were balanced between treatment arms in the analysis population. Median increase in tumor size from baseline to first on-treatment assessment in the maintenance period was 2% with nivo vs 17% with pbo. Compared to the pbo arm, fewer patients on nivo had SLD increases of ≥20% (27% vs 46%), ≥50% (10% vs 22%), and≥100% (3% vs 6%) at the first on-treatment scan. Conclusions: In this analysis, nivo was not associated with HP in the pbo-controlled CheckMate 451 trial. These data are consistent with the previous analysis of ATTRACTION-2, suggesting that reports of HP with immunotherapy may reflect some patients' natural course of disease.