Exactech (South Korea)

Abstract ObjectiveKnee osteoarthritis (OA) is a leading cause of disability worldwide, with early identification of structural changes critical for improving patient outcomes. While magnetic resonance imaging (MRI) provides rich spatial detail, its interpretation remains challenging due to complex anatomy, subtle lesion presentation, and limited voxel-level annotations. Meanwhile, radiology reports encode semantic and diagnostic insights that are typically underutilized in imaging AI pipelines. In this work, we introduce CMANet, a Cross-Modal Attention Network that integrates 3D knee MRI volumes with their corresponding free-text radiology reports for joint OA severity classification and lesion segmentation. CMANet introduces four key innovations: (1) an asymmetric cross-modal attention mechanism that enables bidirectional information flow between image and text, (2) a weakly supervised anatomical alignment module linking report phrases to MRI regions, (3) a multi-task prediction head for simultaneous OA grading and voxel-level lesion detection, and (4) interpretable attention pathways for tracing predictions to report language and anatomical structures. Evaluated on a dataset of 642 patients with paired MRI and radiology reports, CMANet achieved significant improvements over unimodal baselines—boosting KL-grade classification AUC from 0.769 to 0.871 (Δ =0.102, p=0.004) and increasing Dice scores for cartilage and BML lesion segmentation. The model also demonstrated generalizability in predicting 2-year OA progression (AUC=0.804) and achieved improved alignment between anatomical regions and textual descriptions. These results highlight the potential of multimodal learning to enhance diagnostic accuracy, spatial localization, and explainability in musculoskeletal imaging.

This study's purpose is to compare clinical and radiographic outcomes of primary anatomic total shoulder arthroplasty (aTSA) and primary reverse total shoulder arthroplasty (rTSA) patients with osteoarthritis (OA) and an intact rotator cuff with no previous history of shoulder surgery using a single platform TSA system at a minimum follow-up of 2 years. A total of 370 aTSA patients and 370 rTSA patients matched for age, sex, and length of follow-up from an international multi-institutional Western Institutional Review Board approved registry with a minimum 2-year follow-up were reviewed for this study. All patients had a diagnosis of OA, an intact rotator cuff, and no previous shoulder surgery. All patients were evaluated and scored preoperatively and at latest follow-up using six outcome scoring metrics and four active range of motion measurements. Mean follow-up was 41 months, and the mean age was 73 years. Preoperatively, the rTSA patients had lower outcome metrics and less motion. Postoperatively, aTSA and rTSA patients had similar clinical outcomes, motion, and function, with the only exception being greater external rotation in aTSA exceeding the minimal clinically important difference. Pain relief was excellent, and patient satisfaction was high in both groups. Humeral radiolucent lines were similar in both groups (8%). Complications were significantly higher with aTSA (aTSA = 4.9%; rTSA = 2.2%; P = 0.045), but revisions were similar (aTSA = 3.2%; rTSA = 1.4%; P = 0.086). At a mean of 41 month follow-up, primary aTSA and rTSA patients with OA and an intact rotator cuff with no previous history of shoulder surgery had similar clinical and radiographic outcomes. Greater external rotation was noted in aTSA patients at follow-up. However, aTSA patients had a significantly greater rate of complications compared with rTSA patients. rTSA is a viable treatment option in patients with an intact rotator cuff and no previous shoulder surgery, offering similar clinical outcomes with a lower complication rate. Level III.