The Perfect Circle Technique Shows Poor Inter-rater Reliability in Measuring Anterior Glenoid Bone Loss on Magnetic Resonance Imaging

Document Type

Journal Article

Publication Date



Arthroscopy, sports medicine, and rehabilitation








PURPOSE: To evaluate the reliability of the perfect circle methodology for measurement of glenoid bone loss in patients with anterior glenohumeral instability. METHODS: We performed a chart review of retrospectively collected patients who underwent isolated arthroscopic anterior labral repair between January 1 and June 30, 2021, using our institution's electronic medical records. The inclusion criteria included isolated anterior shoulder instability with anterior labral repair and corroborated tears on magnetic resonance imaging. A total of 9 raters, either sports or shoulder and elbow fellowship-trained orthopaedic surgeons, each evaluated the affected shoulder magnetic resonance imaging scans twice, with a minimum of 2 weeks between measurements. Measurements followed the "perfect circle" technique and included projected anterior-to-posterior glenoid diameter, amount of posterior bone loss, and percentage of posterior bone loss. Intrarater reliability and inter-rater reliability were then determined by calculating intraclass correlation coefficients (ICCs). RESULTS: Ten consecutive patients meeting the selection criteria were chosen for inclusion in this analysis. Average estimated bone loss for the cohort was 2.45 mm, and the mean estimated glenoid diameter of the involved shoulder was 28.82 mm. The average percentage of bone loss measured 8.54%. The ICC for interobserver reliability was 0.55 for the perfect circle diameter and 0.17 for the anterior bone loss measurement (poorly to moderately reliable). The ICC for intraobserver reliability was 0.69 for the perfect circle diameter and 0.71 for anterior bone loss (moderately reliable). CONCLUSIONS: The perfect circle technique for estimating anterior glenoid bone loss on magnetic resonance imaging was found to have moderate intrarater reliability; however, reliability between observers was found to be moderate to poor. LEVEL OF EVIDENCE: Level IV, diagnostic case series.


School of Medicine and Health Sciences Student Works