Diagnostic Accuracy of a Telemedicine-Based Shoulder Examination Compared With MRI in Rotator Cuff Syndrome: A Cross-Sectional Study

Background Telemedicine is increasingly used in musculoskeletal care, but the accuracy of remote physical examination for rotator cuff syndrome, compared with magnetic resonance imaging (MRI), remains unclear. This study evaluated the diagnostic performance of telemedicine-based shoulder maneuvers against MRI, with secondary analyses including in-person examination. Methods A cross-sectional diagnostic accuracy study was conducted. Patients underwent standardized teleconsultations with Jobe, infraspinatus, and Bear Hug tests, each analyzed for pain and weakness. All patients also underwent the same in-person maneuvers. MRI of the affected shoulder, reviewed by blinded musculoskeletal radiologists, served as the reference standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and balanced accuracy were calculated. Secondary analyses included comparisons between telemedicine and in-person performance, and agreement between modalities. Results Thirty-two patients (mean age, 59.6 years) were included. Against MRI, telemedicine tests were highly sensitive: Jobe pain sensitivity was 96.2%, infraspinatus weakness sensitivity was 76.9%, and Bear Hug weakness sensitivity was 57.1%. In-person tests showed similar sensitivity for most maneuvers, with a tendency to classify fewer tendons as weak in MRI-negative shoulders. Paired analyses demonstrated minor differences for pain outcomes, but significant differences for weakness, with telemedicine being more sensitive and in-person examinations being more conservative in classifying weakness. Agreement between modalities was excellent for pain (κ up to 1.00) and moderate for weakness (κ, 0.11-0.47). Conclusion Telemedicine-based shoulder examination demonstrated high sensitivity compared with MRI, similar to in-person examination. Because all patients had MRI-confirmed pathology, and few true negative tendons were available, the ability of either modality to rule out rotator cuff tears based solely on clinical testing was limited. These findings support the use of telemedicine as a sensitive triage tool, while MRI remains necessary for definitive diagnosis and treatment planning.