Effect of reduced orbital rotation on image quality and intra-articular screw detection in intraoperative 3D imaging of proximal humerus plate fixation: a cadaveric study

Abstract Background Proximal humerus fractures are common injuries in the elderly, frequently treated with plate osteosynthesis. Intraoperative 2D fluoroscopy provides only limited visualization of the humeral head’s complex anatomy, often leading to undetected intra-articular screw perforations. Intraoperative 3D imaging provides enhanced assessment but is technically limited by the confined operating field and the risk of collisions between the C-arm, patient, and operating table. This cadaveric study systematically investigated how progressive reduction of orbital rotation during intraoperative 3D imaging of the proximal humerus affects image quality and intra-articular screw detection accuracy in plate osteosynthesis. Methods Five fresh-frozen cadaveric shoulders were scanned using an isocentric C-arm. Full 200° scans served as the gold standard. From these datasets, reconstructions with reduced orbital rotations (100°–200°, in 20° increments) were generated under three imaging conditions: (1) without a metal implant, (2) with plate osteosynthesis and extra-articular screw configuration, and (3) with plate osteosynthesis and intra-articular screw configuration. Both configurations were reconstructed with and without metal artifact reduction (MAR), resulting in 30 datasets per specimen. Three blinded raters independently assessed subjective image quality, visualization of the articular surface, intra-articular screw detection, and diagnostic certainty. Results A reduction of orbital rotation had a significant overall effect on subjective image quality and assessability of articular surfaces (p < 0.001). Post-hoc analysis showed no further significant improvement in image quality beyond 160° of orbital rotation. MAR significantly enhanced image quality and surface visualization (p < 0.001). Sensitivity in detecting intra-articular screws was preserved down to an orbital rotation of 160°, while specificity was unaffected by rotation reduction (p = 0.519). MAR did not significantly influence screw detection accuracy. Conclusions Intraoperative 3D imaging with a 160° orbital rotation yields sufficient image quality, enabling the reliable identification of intra-articular screws during plate osteosynthesis of the proximal humerus. The additional use of MAR enhances visualization in the presence of metal without increasing the detection rate of misplaced screws. 3D imaging with reduced orbital rotation minimizes spatial demands and collision risks, supporting its practical application as an effective intraoperative imaging technique for the shoulder.