MOIO: Interim analyses of a phase III randomized non inferiority trial assessing reduced frequency immune checkpoint inhibitors in responding patients with advanced/metastatic cancer

MOIO

Patient / Population Intervention / Exposition Vergleich Endpunkt

In patients with advanced or metastatic solid tumours responding to immune checkpoint inhibitors, reduced dose-intensity immunotherapy was not non-inferior for progression-free survival compared with standard dose-intensity immunotherapy.

N

340Patienten

Design

Academic, multicentre, phase III randomised non-inferiority trial across 36 French centres in patients with locally advanced or metastatic solid tumours in partial or complete response after 6 months of standard-dose immunotherapy (melanoma in complete response excluded), randomised to standard dose-intensity immunotherapy or reduced dose-intensity immunotherapy every 3 months at the same dose (NCT05078047).

Endpunkt

Progression-free survival, with non-inferiority defined by a hazard ratio margin of 1.3.

ErgebnisIn the modified intent-to-treat set (N=308), median PFS 25 months (95% CI 19.4-NE) with standard vs 21.9 months (95% CI 12.1-31.6) with reduced immunotherapy (HR 1.35, 95% CI 0.9-2.0; p=0.0635), exceeding the prespecified futility boundary. Reduced immunotherapy was associated with a higher risk of death (HR 2.0, 95% CI 1.1-3.9; p=0.0143). PFS HR 2.9 (1.3-6.7) in urothelial, 1.6 (0.8-3.1) in lung, 0.7 (0.3-1.5) in renal cell and 0.6 (0.1-2.4) in head and neck cancer. Beyond first line HR 2.9 (0.9-4.9). Grade 3-4 immune-related AEs 10.5% vs 9.5% and SAEs 26.3% vs 25.0% with standard vs reduced immunotherapy.

Gravis G, Barthelemy P, Korakis I, et al. Presented at ESMO Congress 2026, abstract LBA10, 26 October. Not yet published.

Diskussion & Kritik

At this interim analysis, reduced dose-intensity immunotherapy crossed the prespecified futility boundary for PFS with a hazard ratio of 1.35. It also showed a signal of worse overall survival and no meaningful safety advantage. Follow-up is short at a median of 18.6 months, only 340 of up to 646 planned patients were randomised, and the abstract does not say why the modified intent-to-treat set is smaller than the randomised population. Blinding is not stated, and the subgroup estimates by tumour type and treatment line have wide confidence intervals, so they generate hypotheses only. The abstract does not report the number of events, response duration or which checkpoint inhibitors were used.