Introduction: Ultrasonography offers a fast and inexpensive method to evaluate the rotator cuff in the office setting. However, the accuracy of ultrasound is highly user dependent. The purpose of this study is to investigate the learning curve of an orthopaedic surgeon in using ultrasound to diagnose rotator cuff tears.
Methods: A sports medicine fellowship trained orthopaedic surgeon was taught how to perform and interpret an ultrasound examination of the shoulder by a musculoskeletal radiologist. In this prospective study, subjects were patients who presented to the office with shoulder pain suspected to be consistent with rotator cuff pathology, either complete or partial tears. The surgeon was blinded to magnetic resonance imaging (MRI) results and performed the ultrasound after performing a physical exam. Based on ultrasound and exam, the surgeon assessed if the rotator cuff was intact (no tear) or torn (having a partial or full thickness tear). Results were compared to MRI findings and arthroscopic findings (when available), and accuracy was evaluated over time to determine overall accuracy and if significant learning and improvement in accuracy took place over the time period studied.
Results: Eighty patients were enrolled in the study; seventy-four had an MRI within 3 months of the ultrasound and were available for evaluation. Nineteen patients underwent ultrasound, MRI, and arthroscopy. Ultrasound was able to accurately diagnose the correct rotator cuff pathology (no tear, a partial thickness tear, or a full thickness tear) in 61% of patients. Ultrasound accurately diagnosed simply the presence or absence of a tear in 74% of patients. There was a general trend toward improved accuracy as the investigator gained experience, with accuracy rates of approximately 51% for the first 40 patients, and 69% for the last 40 patients evaluated, although this difference was not statistically significant (p = 0.154).
Discussion: Ultrasound imaging requires significant training and practice to provide a clinically useful level of diagnostic accuracy. The applicability of this procedure for diagnosing primary rotator cuff tears in an orthopaedic office setting may be limited by the time and volume required for the practitioner to approach the accuracy reported for diagnostic ultrasound and MRI in the literature.