The Ratio of the Transverse to Longitudinal Diameter of the Glenoid Projection Is of Good Predictive Value for Defining the Reliability of Critical Shoulder Angle in Nonstandard Anteroposterior Radiographs
Jingyi Hou, Fangqi Li, Xiang Zhang, Yuanhao Zhang, Yaping Yang, Yiyong Tang, Rui Yang
Abstract
PURPOSE: To investigate (1) whether measurement of the critical shoulder angle (CSA) is still reliable in some nonstandard anteroposterior radiographs; and (2) whether the ratio of the transverse to longitudinal diameter of the glenoid projection (RTL) could be used to determine the reliability of the CSA in nonstandard anteroposterior radiographs. METHODS: The authors analyzed radiology records from 2017 to 2019 for patients with computed tomography (CT) scans of the shoulder. According to the Suter-Henninger classification system, each CT scan underwent 3-dimensional (3D) reconstructions to obtain 8 digitally reconstructed radiographs (DRRs), including 1 type A1 film and 7 type D1 films with different rotation angles. CSA and RTL were measured on all films, and 2 blinded reviewers evaluated DRRs. The relationship between RTL and CSA was determined by Pearson correlation test. The threshold value was determined by receiver operating characteristic (ROC) analyses using RTL as predictors and defined reliable CSA as criterion in the logistic regression. Intra- and inter-rater reliability were assessed by the intraclass correlation coefficient (ICC). Discriminative capacity was calculated with ROC analyses and area under the curve (AUC). An optimal measurement cutoff with resultant sensitivity and specificity was calculated. RESULTS: A total of 86 patients were included, mean age 40.70 ± 18.25 years, 19 of whom had rotator cuff tears (RCTs) and 67 without RCTs. Spearman's correlation test revealed that the deviation of CSA was positively correlated with RTL, correlation coefficient 0.544 (P < .001) in all patients and 0.686 (P < .001) only in patients with RCTs. ROC analysis of all patients showed that the AUC of the RTL was 0.90, and the optimal cutoff point was 0.25 (sensitivity 88%, specificity 87%). ROC analysis only of patients with RCTs showed that the AUC of the RTL was 0.95, and the optimal cutoff point was 0.22 (sensitivity 95%, specificity 92%). CONCLUSION: The ratio of the transverse to longitudinal diameter of the glenoid projection (RTL) is of good predictive value in defining the reliability of the CSA in malposition films. Based on the results, the CSA can be considered reliable if its RTL is <0.25. LEVEL OF EVIDENCE: III, retrospective cohort study investigating a diagnostic test.