These authors assessed the inter- and intraobserver reliability of the modified Walch classification using 3-dimensional (3D) computed tomography (CT) imaging as determined by experienced shoulder surgeons.
In Group 1 96 cases involving all modified Walch classification categories were evaluated by 12 readers.In group 2 98 cases involving posterior glenoid deformity categories [B2, B3, C1, C2] evaluated by 11 readers.
Interobserver reliability showed fair to moderate agreement for both groups (kappa of .41 to 43). Group 2 had a kappa of .37 to .38. Intraobserver reliability showed substantial agreement for group 1 (kappa of .61 to .63).For group 2, intraobserver reliability showed moderate to substantial agreement (.51 to .61).
Their findings suggest that cases with a spectrum of posterior glenoid bone loss and/or dysplasia can be harder to distinguish by modified Walch type because of a lack of defined thresholds.
The characteristics of the different types of glenoidhumeral arthritic pathoanatomy used in this study are shown below
A1 Centered humeral head, minor glenoid erosion.
A2 Centered humeral head, major central glenoid erosion defined by a line drawn from the anterior to posterior rims of the glenoid transecting the humeral head.
B1 Posteriorly subluxated humeral head, with no or minor posterior glenoid erosion.
B2 Posteriorly subluxated humeral head, posterior glenoid erosion with biconcavity and no dysplasia.
B3 Posteriorly worn glenoid that is monoconcave with little or no biconcavity due to posterior and central glenoid erosion, without dysplasia.
C1 Dysplastic glenoid with high degrees of retroversion due to dysplasia rather than glenoid erosion
C2 Dysplastic glenoid with acquired posterior glenoid erosion creating glenoid biconcavity and posterior subluxation of the humeral head.
D Glenoid anteversion or anterior humeral head subluxation.
Note that the criteria above require the observer to differentiate between "minor" and "major" and "high degrees" and lesser degrees.
The diagram below demonstrates the problem of classifying pathoanatomies that lie in between A1 and A2, between A1 and B1, between B1 and B2, and in between A2 and B2.
The fact is that arthritic glenohumeral pathoanatomy varies widely among shoulders and the attempts of a surgeon to push this variability into discrete pigeon holes is influenced by his or her own experience in looking at the images. As the authors point out "even internally consistent readers are applying criteria to select a particular modified Walch type in different ways when compared to each other".
It may be clearer to describe the pathoanatomy of each shoulder in terms of measurable characteristics: (1) glenoid retroversion in degrees
(2) the amount of decentering of the humeral head on the glenoid as a percentage and
(3) the amount and location of glenoid bone loss.
