Abstract
BACKGROUND: Early-stage glottic carcinoma is characterized by high cure rate, but about 10% of patients still develop local recurrence after radiotherapy (RT). Although advances in local therapies have markedly improved larynx preservation, anterior commissure involvement (ACI) may worsen outcomes yet remains insufficient to fully guide personalized treatment. Here, we conducted a univariate meta-analysis and a Bayesian multivariate meta-analysis of multiple factors (MVMA-MF) to evaluate the prognostic impact of ACI and on local failure in early-stage glottic carcinoma. METHODS: We conducted a systematic review with univariate meta-analysis and MVMA-MF using data from studies of early-stage glottic carcinoma treated with definitive RT, incorporating clinical information, risk factors and survival outcomes. A systematic literature search was conducted in PubMed and Scopus on Sep 27, 2025. Eligible studies provided sufficient data to estimate odds ratios (ORs) for local control in relation to patient, tumor and treatment factors, including ACI. Data extraction followed PRISMA guidelines, and study quality was assessed using MINORS criteria. The primary endpoint of the study was the effect of ACI on local failure. Pooled ORs were calculated with a random-effects model, and heterogeneity was assessed using the Cochrane Q test and I (2) statistic. To address correlations among risk factors, we applied a Bayesian multivariate meta-analysis of multiple factors (MVMA-MF), testing three models (H, M0, U) with Markov Chain Monte Carlo algorithm. FINDINGS: Nine clinical trials representing 2527 patients were included. In univariate meta-analysis, the presence of ACI was significantly associated with an increased risk of local recurrence in early-stage glottic carcinoma treated with RT (OR 1.61, 95% CI 1.15-2.26, p = 0.006), with moderate heterogeneity (I (2) = 34%). In MVMA-MF including 14 risk factors, only T substage demonstrated a consistent and statistically significant association with recurrence across all models. ACI, smoking status and other risk factors showed non-significant associations, with wide credible intervals overlapping unity. Correlation analysis revealed mostly low-to-moderate inter-variable correlations, supporting the multivariable modeling approach. Comparison of univariate and multivariate estimates for ACI indicated that the unadjusted analysis suggested increased recurrence risk, whereas adjusted models yielded attenuated and non-significant associations (adjusted ORs 1.05-1.43, all 95% CrIs including unity). INTERPRETATION: This study represents the first application of both univariate and multivariate meta-analysis to clarify the prognostic role of ACI in early-stage glottic carcinoma treated with definitive RT. While univariate analysis suggested ACI increased recurrence risk, this effect was not confirmed in MVMA-MF, indicating the univariate findings may be confounded by correlation with other factors, particularly T substage. MVMA-MF approach accounted for heterogeneity, collinearity and incomplete reporting, providing more robust evidence. This work underscores the value of advanced multivariate methods in oncology meta-analysis and provides the strongest evidence to date that ACI should not be regarded as a fundamental prognostic criterion in early-stage glottic carcinoma.