ABSTRACT
Aim To compare the accuracy of shoulder ultrasound (US) in diagnosing rotator cuff tendon tears between junior and experienced musculoskeletal (MSK) radiologists. Also, to compare the overall ultrasound accuracy referenced to MRI. Methods A retrospective data collection for patients with clinically suspected rotator cuff tears who underwent ultrasound from June 2021 - June 2023 was conducted. Patients who also performed MRI for the same shoulder were only included in the study. US and MRI images were evaluated by two MSK radiologists with different experience levels. The diagnosis of rotator cuff tears was done on MRI through consensus. Ultrasound accuracy referenced to MRI was calculated for each radiologist. A second consensus was conducted for US images to calculate the overall US accuracy. The percentage of agreement and Cohen's kappa correlation coefficient were calculated before and after the US consensus. Results Forty-one patients were included in the study, 12 (29.3%) males and 29 (70.7%) females, with a mean age of 49.6 years. Sensitivity, specificity, positive (PPV) and negative predictive values (NPV), and accuracy of US interpreted by junior vs experienced MSK radiologists for supraspinatus full thickness tears (FTTs) were 100% vs 91%, 90% vs 93%, 79% vs 83%, 100% vs 97%, and 93% vs 93%, respectively. After the second consensus, sensitivity, specificity, PPV and NPV, and accuracy were 91%, 90%, 77%, 96%, and 90%, respectively. Conclusion The accuracy of shoulder ultrasound in diagnosing supraspinatus FTTs by junior MSK radiologists compares well to the more experienced ones, but not for partial thickness tears (PTTs).
