Reliability of Acromiohumeral Interval Measurement on Nonstandardized Shoulder Radiographs by Reporting Radiographers: A Retrospective Service Evaluation

Introduction: The acromiohumeral interval (AHI) on shoulder radiographs is widely used as an indirect marker of chronic rotator cuff (RC) pathology, but routine anteroposterior (AP) radiographs are often nonstandardized, and reliability in day-to-day reporting is uncertain. This service evaluation assessed inter- and intraobserver reliability of AHI measurement by reporting radiographers and examined correspondence with ultrasound for identifying clinically significant RC tears. Methods: Seventy-three nonstandardized AP shoulder radiographs from a UK NHS Trust atraumatic shoulder pathway (age 18-70 years) with a Trust-performed shoulder ultrasound report within six months were independently assessed by three reporting radiographers. Each observer measured AHI (mm, one decimal place) and recorded the Hamada grade; measurements were repeated after two weeks. Reliability was assessed using intraclass correlation coefficients (ICCs). Measurement error was quantified using the standard error of measurement (SEM) and the minimal detectable change at 95% confidence (MDC95). Ultrasound reports were dichotomized as positive for full-thickness (complete) or massive RC tear and negative for partial-thickness tear, tendinopathy, or no tear. Radiographs were dichotomized as AHI ≤6.0 mm (positive) vs. >6.0 mm (negative). Results: Intraobserver reliability was good to excellent (ICC, 0.88-0.98 (good to excellent)). Interobserver reliability was excellent in Trial 1 (ICC, 0.95 (excellent)) and good to excellent in Trial 2 (ICC, 0.92 (good to excellent)). SEM was <1 mm, and MDC95 was approximately 1.6-2.4 mm. The Hamada agreement was weak to moderate. Using the definitions above, AHI ≤6.0 mm showed a mean sensitivity of 42% and a mean specificity of 94% (mean across observers and trials), with an overall accuracy above 70%. Conclusions: Reporting radiographers measured AHI on nonstandardized AP radiographs with good-to-excellent reliability. Although sensitivity is limited (as expected for an indirect marker of advanced disease), high specificity supports AHI ≤6.0 mm as a rule-in feature for chronic full-thickness or massive RC tears. Standardized landmark guidance and focused training on common pitfalls (arm abduction, os acromiale) may further improve consistency.