Intraoperative 3D Tomosynthesis of the Elbow in a Simulated Prone-Position Setup: A Cadaveric Study

Background/Objectives: Fractures of the distal humerus and olecranon often require surgical intervention due to their complexity and intra-articular fracture pattern. Precise implant positioning and articular surface reconstruction are crucial for optimal outcomes. Intraoperative 3D imaging offers advantages over 2D fluoroscopy by providing enhanced visualization of fracture reduction and implant alignment. However, limited space in the prone position restricts the use of standard 3D imaging with complete orbital rotation. This cadaver study investigates the feasibility and intra-articular screw detection performance of intraoperative 3D tomosynthesis with reduced orbital rotation at the elbow. Methods: In this cadaver study, six distal humeri and six olecranons were examined using an isocentric 3D C-arm with a complete 200° acquisition. The disarticulated upper extremities were positioned to reproduce the prone surgical setup. Reduced-angle datasets ranging from 200° to 100° were reconstructed retrospectively from the complete acquisition, with and without metal artifact reduction (MAR). Double-plate osteosynthesis was applied to the distal humerus, while an anatomic olecranon plate was used for fixation on the olecranon. To evaluate detection capabilities, screw positions were deliberately modified to include both correct placements and intra-articular malpositions. Image quality (5-point Likert scale) and certainty of intra-articular screw assessment were evaluated by three blinded raters; sensitivity and specificity were calculated relative to the complete 200° imaging-based reference condition. Results: Image quality, sensitivity, and specificity decreased with reduced orbital rotation. MAR significantly improved image quality and specificity, but not sensitivity. At an orbital rotation of 160°, MAR-reconstructed tomosynthesis achieved acceptable image quality (3.36 ± 0.53), high sensitivity (94.9%), and high specificity (93.9%) for detecting intra-articular screws. Conclusions: Intraoperative tomosynthesis with MAR provides acceptable image quality and high performance for intra-articular screw detection at the elbow under the present experimental conditions. This approach may support intraoperative assessment in prone-position elbow surgery when full orbital rotation is not feasible. Clinical studies are needed to confirm its value in routine practice.