Methodological and Reporting Rigor in Non-Inferiority and Equivalence Trials for Multimodality Cancer Treatment: Lessons from Breast Cancer Radiotherapy.

Abstract

Objective: To investigate the methodological and reporting quality of non-inferiority (NI) or equivalence trials in breast cancer radiotherapy, providing suggestions for future NI/equivalence trials.

Study design and setting: Prospective phase III randomized controlled trials comparing different radiation modalities in breast cancer and designed as NI/equivalence were identified. Extracted data included the trial design, analysis, and reporting characteristics. The relationship between trial numbers and publication year was assessed. Trials with pre-specified NI margins as absolute risk differences (ARDs) were reevaluated using margins as relative risks.

Results: Twenty-one studies were included. The number of publications increased over time. Trial interventions mainly involved dose fractionation and radiation volume. The primary endpoint was local or locoregional recurrence in 15, toxicity in 5, and both in 1 study. Reporting gaps included: Not specifying the trial as NI/equivalence in the title/abstract (n = 6); inadequate justification for the NI/equivalence design rationale (n = 10) or margins (n = 12); absence of both intention-to-treat and per-protocol analysis (n = 12); no reporting of p values for NI/equivalence tests (n = 12) or margins with confidence intervals (n = 5). Fifteen studies failed to meet their planned accrual target, mostly owing to overestimation of event rates in the control group. Among 8 trials with 9 comparisons claiming NI with pre-specified margins as ARDs, 4 comparisons were classified as inconclusive when using the margins as relative risks.

Conclusion: Recently, NI/equivalence trials have dramatically increased in breast cancer radiotherapy; however, there is substantial room for improvement in their methodological and reporting quality.