Critical Shoulder Angle and Height-to-Roof Ratio Measurement on Shoulder Radiography: A Comparison for Predicting Rotator Cuff Tears

Purpose or Learning Objective: To investigate the diagnostic accuracy of the critical shoulder angle (CSA) and height-to-roof ratio (H/R) measurements for detecting rotator cuff tears (RCTs).

Methods or Background: A total of 116 patients having a correct anteroposterior view shoulder radiography, shoulder magnetic resonance (MR) imaging (acquired no more than 3 months from radiography), and no history of trauma, shoulder osteoarthritis, or neoplasia were investigated retrospectively. In all of these patients, CSA and H/R were determined, and MR images were assessed for the presence of RCTs.

The patients were classified into four categories (0 = no tears; 1 = supraspinatus degeneration sine tear; 2 = supraspinatus partial-thickness tear; and 3 = supraspinatus full-thickness tear) and then divided into two groups: “no tears” (categories 0 and 1: 56 patients) and “tears” (categories 2 and 3: 60 patients). CSA and H/R were compared between the two groups (no tears and tears), using the Mann-Whitney U test. Age was also compared between the two groups.

In the tear group, CSA and H/R were compared between male and female, right and left shoulder, and patients < 60 and > 60 years of age, using the U test. Receiver operating characteristic (ROC) analysis was performed for CSA, H/R, and age to predict the strength of the tests. The area under the curve (AUC) of CSA and H/R were compared.

Results or Findings: CSA (33.1 versus 36.9) and H/R (0.65 versus 0.69) were significantly larger in the tear group than the no tear group.

The tear group patients were significantly older than the no tear patients. In the tear group, H/R was lower in female patients and CSA results were larger in the left shoulder.

For CSA, ROC analysis reported a sensitivity of 83.3% and a specificity of 82.1%, with a threshold value > 35.1. For H/R, ROC analysis reported a sensitivity of 82.1% and a specificity of 62.5%, with a threshold value > 0.65. Comparison of ROC curves showed a significant difference between AUCs.

Conclusion: CSA and H/R ratio measurement had good diagnostic performance for RCTs, with cutoff values of 35.1 and 0.65 degrees, respectively. CSA proved to be more sensitive and specific than H/R for predicting supraspinatus tears. Measuring CSA on plain radiographs may provide clinicians with a readily available and reliable method for detecting RCTs in daily practice.

Publication History

Article published online:
02 June 2022