BackgroundComputed tomography (CT) and magnetic resonance imaging (MRI) are essential diagnostic tools but often provoke significant anxiety in patients, which may compromise image quality and patient cooperation. Non-pharmacological interventions such as music have been proposed as feasible alternatives to anxiolytic medications. However, the evidence supporting music interventions in radiology settings has not been systematically synthesized.ObjectiveThis meta-analysis aimed to evaluate the effect of music intervention on anxiety in patients undergoing CT or MRI.MethodsA systematic search was conducted in PubMed, EMBASE, Cochrane Library, Web of Science, and CINAHL from inception to March 1, 2026. Randomized controlled trials (RCTs) evaluating music interventions versus control conditions (e.g., usual care or headphones only) in patients undergoing CT or MRI were included. Outcomes included anxiety [State-Trait Anxiety Inventory (STAI), Visual Analog Scale (VAS), or Child Anxiety Metacognition Scale (CAM-S)], heart rate (HR), and peripheral oxygen saturation (SpO2). Standardized mean differences (SMDs) or weighted mean differences (WMDs) with 95% confidence intervals (CIs) were pooled using random-effects models. Subgroup analyses were performed by measurement tool, imaging modality, and patient population.ResultsSix RCTs (n=972 patients) were included. Music significantly reduced post-intervention anxiety compared with control (SMD = -1.94, 95% CI: -3.59 to -0.28, p=0.02, I²=98.5%). Analysis of pre-to-post change in anxiety also favored music (SMD = -0.74, 95% CI: -1.22 to -0.26, p=0.003, I²=80.5%). Music did not significantly affect HR (WMD = -3.99 bpm, 95% CI: -9.65 to 1.67, p=0.17, I²=94.6%) or SpO2 (WMD = 0.30%, 95% CI: -0.26 to 0.86, p=0.30, I²=74.6%). Subgroup analyses suggested potential effect modification by measurement tool and imaging modality, but heterogeneity remained high. Sensitivity analyses confirmed robustness of anxiety outcomes.ConclusionPooled estimates suggest a reduction in anxiety favoring music, but the extreme heterogeneity (I² = 98.5%) indicates that the studies do not share a common effect size; therefore, the pooled estimate may not represent a clinically meaningful common effect. Music appears safe and low-cost, but conclusions should be drawn cautiously. Future high-quality RCTs with standardized protocols are needed to optimize implementation strategies.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO, identifier CRD420261356881.