Ensuring Superior Reporting of Non-Inferiority Radiotherapy Clinical Trials: A Systematic Review.

Non-inferiority clinical trials are used to investigate whether a new treatment is not worse than a control by an acceptable amount. These new treatments can be more convenient, cost-effective, or tolerable than the standard of care. Although the frequency of this trial design is increasing, the consistency of reporting and interpretation of these trials can vary. The aim of this systematic review was to assess the methodologic quality of non-inferiority radiotherapy clinical trials. A literature search of PubMed, EMBASE and Cochrane databases of randomized controlled radiotherapy trials with non-inferiority hypotheses published in English between January 2000 and July 2021 was performed by an information scientist. Exclusion criteria included study protocols, and follow-up or interim analyses. Records were screened and full texts were reviewed by two authors with discrepancies settled by a third reviewer or by consensus. Descriptive statistics pertaining to study size, primary cancer type, type of comparison, endpoint, and statistical measures were performed. This review was prospectively registered on PROSPERO: CRD42021270644. Of 260 records identified on initial search, 44 (17%) trials were included after full-text review. The mean number of participants was 876 (IQR: 216.25–912.5). All were published after 2004, and open label. Four (9%) trials were funded by industry. The most common primary cancer type was breast (n=10, 23%). Most studies (n=37, 84%) had 2 treatment arms. The most common types of interventions included altered radiation fractionation (n=13, 30%) and changes to systemic therapy agents (n=10, 23%). Six (14%) studies compared radiation to another treatment modality (e.g., surgery, radiofrequency ablation), and 3 (7%) studies examined the omission of radiation. The most common primary endpoints were progression-free survival (n=12, 27%) and local/regional control (n=12, 27%). Forty (91%) studies reported the non-inferiority margin, and only 5 (11%) studies provided a statistical rationale for the margin. The median non-inferiority margin was 10% (IQR: 5–10%). Sample size calculation was reported in 42 (95%) studies, and 42 (95%) studies reported confidence intervals of the primary endpoint. Both intention-to-treat and per protocol analyses were reported in the majority (52%) of studies. In 23 (52%) trials, a non-inferiority claim of the primary endpoint was reached. Clinical trials with a non-inferiority design are relatively uncommon in radiation oncology, and have been used for variable types of comparisons and primary endpoints. Standardization of required data to report and statistical methodology would be helpful to ensure proper interpretation of trial results. [ABSTRACT FROM AUTHOR]