Radiosensitising strategies in locally advanced head and neck cancer: a systematic review and network meta-analysis of randomised controlled trials (RADIANCE-NMA)

Background Radiosensitiser enhance radiotherapy efficacy in locally advanced head and neck squamous cell carcinoma (HNSCC). Although cisplatin-based chemoradiotherapy is the standard-of-care, over two dozen radiosensitising strategies have been evaluated in randomised trials with no network meta-analysis comparing all drug classes. Methods PubMed, Embase, Cochrane CENTRAL, and ClinicalTrials.gov were searched from inception to 26 June 2026 for Phase IIb–III randomised controlled trials evaluating concurrent radiosensitiser therapy in adults with locally advanced head and neck squamous cell carcinoma (nasopharyngeal carcinoma excluded). Studies lacking overall survival data and induction-only comparisons were excluded. The primary outcome was overall survival (OS). All treatment-effect estimates are reported as hazard ratios (HRs) with 95% confidence intervals (CIs). Data were independently extracted by two reviewers. Risk of bias in individual trials was assessed with the Cochrane Risk of Bias (RoB 2) tool and confidence in the network estimates using the Confidence in Network Meta-Analysis (CINeMA) framework. Random-effects network meta-analysis was done with frequentist (netmeta) and Bayesian (gemtc) frameworks. Between-study heterogeneity was quantified using the I 2 statistic and τ 2 ; publication bias and small-study effects were assessed using comparison-adjusted funnel plots and Egger's regression test. This study is registered with the Open Science Framework (https://osf.io/cygsh). Findings A total of 44 trials (13,768 patients; 84% male) were included in the primary analysis, with 23 additional trials in sensitivity analyses (67 total; 19,527 patients). Risk of bias showed some concerns in 39 of 44 (89%) primary trials on the Cochrane RoB 2 tool. Four treatments showed significant overall survival (OS) benefit versus radiotherapy alone: cisplatin plus 5-fluorouracil (hazard ratio (HR) 0.35; 95% confidence interval (CI) 0.18–0.66), nimotuzumab plus weekly cisplatin (0.48; 0.29–0.81), weekly cisplatin (0.57; 0.44–0.74), and three-weekly cisplatin (0.79; 0.65–0.95). In the network estimate, weekly cisplatin was favoured over three-weekly (HR 0.73; 0.58–0.93); the underlying randomised trials were designed as non-inferiority comparisons. Between-study heterogeneity was moderate (I 2 = 48.9%; τ 2 = 0.036; Cochran Q=37.15, df=19, p=0.008), with no detected inconsistency (design-by-treatment global test p=0.88). The comparison-adjusted funnel plot and Egger's regression test showed no evidence of small-study effects, but are interpreted as exploratory given fewer than ten trials per edge on most comparisons. Grade 3–4 toxicity was highest with weekly cisplatin (pooled 71.6% across contributing trials) and lowest with immunotherapy combinations. Meta-regression confirmed enrolment era as a significant effect modifier (β = 0.23 post-2005 versus pre-2005; p=0.010), consistent with the transition from two-dimensional or conformal radiotherapy in earlier trials to intensity-modulated radiotherapy in modern trials. Interpretation Platinum-based multi-agent chemoradiotherapy and weekly cisplatin are the most effective radiosensitising strategies for locally advanced HNSCC. Weekly cisplatin appears at least non-inferior to the widely used three-weekly schedule and may be favoured in network-integrated estimates, though confidence intervals do not exclude clinical equivalence. Neither cetuximab-based regimens (with radiotherapy or added to chemoradiotherapy) nor immune-checkpoint inhibitors added to cisplatin-based chemoradiotherapy significantly improved survival beyond standard cisplatin chemoradiotherapy. These findings provide a unified evidence framework for treatment selection and trial design.