Influence of radiographic viewing perspective on the reverse shoulder arthroplasty angle

The reverse shoulder arthroplasty (RSA) angle quantifies inclination of the inferior glenoid and, thus the correction required during reverse total shoulder arthroplasty. There are no data on the effect of two-dimensional (2D) radiographic projection on the RSA angle or on the relationships between 2D true anterior-posterior (AP) views and their three-dimensional (3D) counterparts. Digitally reconstructed radiographs of the 2D true AP view of N = 68 scapulae, a subset in controlled ante-/retroversion and extension/flexion views, and their corresponding 3D anatomic models were analyzed. The RSA angle was measured on true AP images with the glenoid in profile at the intersection of lines defined by the supraspinatus fossa and inferior glenoid rim. On altered viewing perspectives, the glenoid face was visible, and thus the anterior and posterior rims. Since it was often unclear which was anterior and posterior, for consistency the RSA angle was measured at the most medial and lateral rims, and glenoid midpoint to determine the influence of measurement location. The 3D RSA angle was measured on 3D models using semi-automated techniques. Data were analyzed to determine the effects of viewing perspective and measurement location on the RSA angle and to compare 2D true AP to 3D measures. The 2D RSA angle was 18.1 ± 7.1° (range: 1.1° to 35.3°), while the 3D RSA angle was 10.1 ± 7.3° (−8.1° to 25.7°) ( P < .001). Ante-/retroversion views had large effects on the RSA angle. The lateral rim was the most susceptible to error (up to 25.8 ± 6.6°) and the glenoid midpoint was least susceptible (less than 5.6 ± 6.5°). Extension/flexion was also influential, but the magnitudes were generally much less than ante/retroversion. Trends due to viewing perspective differed between the medial rim, lateral rim, and glenoid midpoint. The glenoid midpoint maintained the flattest distribution with the smallest errors across the views. Inter- and intra-rater reliability in measuring RSA angles was good to excellent (≥ 0.754). The 2D RSA angle experiences viewing perspective errors when not measured on a true AP radiograph. The glenoid midpoint provided the most consistent and smallest maximum error, with good to excellent reliability. The 3D underestimated the 2D RSA angle on true AP images by an average of −8.0°, with similar variability. When using 2D imaging, a true AP image is desirable; otherwise, use the glenoid midpoint to minimize viewing perspective errors. Comparisons between radiographic measures and 3D preoperative planning should consider the bias between the two techniques, and 3D measures should be evaluated for their measurement techniques within the respective preoperative planning softwares to ensure consistency and reliability among manufacturers.