Diagnostic performance of CT and MRI in mandibular trauma assessment – a prospective comparative study

To compare the diagnostic performance of computed tomography (CT) and magnetic resonance imaging (MRI) for surgically relevant parameters in mandibular trauma, evaluating five in-house-optimized MRI protocols with a dedicated 15-channel mandibular coil. Fifteen patients with 30 acute mandibular fractures underwent 3T MRI using five optimized protocols: UTE, StarVIBE, Dark Bone, DESS, and STIR. Three independent observers qualitatively rated image quality, fracture line visibility, cortical delineation, and bone-to-soft-tissue contrast on a five-point visual analog scale (5 = best, 1 = lowest). Quantitative analysis included measurement of the maximum distance of the fracture gap. Descriptive statistics and inter-observer agreement (Krippendorff’s α) were calculated. CT enabled rapid and accurate fracture detection in all cases. Among MRI protocols, UTE (85% of fractures correctly identified and accurately diagnosed) and StarVIBE (75%) demonstrated the best diagnostic performance, with excellent fracture line visibility, sharp cortical delineation, and strong inter-observer agreement (α = 0.7–1.0.7.0). Dark Bone and DESS provided variable results, while STIR was the least suitable for fracture depiction despite good soft-tissue contrast. Quantitative assessment of fracture gap distance showed the highest accuracy with StarVIBE, demonstrating minimal deviation compared to CT (0.13–0.21 mm; p > 0.05). While CT remains the standard in emergency settings and for delineating complex osseous pathologies, MRI, combining CT-like and water-sensitive sequences, offers a promising radiation-free alternative for simultaneous visualization of hard and soft tissues. This approach may facilitate comprehensive radiation-free trauma evaluation, improve detection of concomitant injuries, and further support individualized clinical decision-making. Swiss National Clinical Trials Portal: SNCTP000006343, ClinicalTrials.gov ID: NCT07012850.