Gulshan B Sharma, Presenter: Nothing to Disclose

Melissa D. Kang MD, Abstract Co-Author: Nothing to Disclose

Paul Harkey MD, Abstract Co-Author: Nothing to Disclose

Douglas D. Robertson MD, PhD, Abstract Co-Author: Nothing to Disclose

Glenoid version is important for glenohumeral stability. Version is most accurately measured using cross-sectional imaging. Accurate measurement requires inclusion of the scapula’s medial edge. Most shoulder MRI & CTs reduce field of view (FOV) excluding this edge. This study’s purpose was to 1) measure accuracy & reliability of version measurements made from shoulder FOV cross-sectional images which replicated clinical MR and CT shoulder studies and 2) discover a reference point within the FOV to substitute for the excluded medial edge.

Twenty cadaver scapulae were CT scanned. A custom positioning device simulated scapular position within a patient’s supine thorax. CT slices through the glenoid center was selected. Selected images were re-reconstructed using a clinical shoulder study FOV excluding the scapula’s medial edge. Three observers twice measured glenoid version on each selected image. Measurements were made using a modified method of Friedman substituting the scapular medial edge point for 1) the most medial scapular point and 2) the inferior glenoid vault apex. True glenoid version was measured from the actual specimens. Method of Friedman measurements were made on full scapula images.

Shoulder study field of view version measurements using the most medial scapula point averaged –11 ± 3 deg and the vault apex averaged –7.5 ± 3 deg. True glenoid version averaged –0.5 ± 3 deg. Mean difference between true version and vault apex referenced version was 7.1 ± 3 deg. Adding 7 deg to vault apex referenced measurements corrected the error (p>0.9). Interobserver reliability had an interclass correlation coefficient (ICC) of 0.72 (p<0.01). Intraobserver reliability had a mean ICC of 0.82 (p<0.01). ±

Absence of the scapula’s medial edge on cross-sectional images adds false retroversion. Use of the scapula’s most medial point as a reference created greater inaccuracy than use of the glenoid vault apex. A single correction factor corrected values.

Use of this methodology can produce accurate version measurement for clinical shoulder MRI and CT studies which exclude the scapula’s medial edge.  

Sharma, G, Kang, M, Harkey, P, Robertson, D, Validity of Shoulder Cross-sectional Imaging for Glenoid Version Measurements.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12034687.html