These authors evaluated the accuracy of glenoid version and inclination measurements using 2D CT scans compared with a validated 3D software program and its influence on decision making on implant selection.
They found that preoperative glenoid retroversion on reformatted 2D CT scans was 11.9° ± 9.6° and mean superior inclination was 10.7° ± 8.6°. When the 3D software was used the results were slightly different, glenoid retroversion averaged 15.1° ± 10.6° and superior inclination averaged 8.9° ± 9.9°.
Surgeon observers decided from glenoid measurements and rotator cuff integrity whether to implant an anatomic or reverse shoulder arthroplasty. A reverse was selected if there was
(1) grade 3 and 4 fatty infiltration according to the Goutallier classification
(2) superior glenoid inclination greater than 10° or
(3) glenoid retroversion greater than 27° of the neoglenoid as well as subluxation of the humerus greater than 80%.
Decisions based on the 2D analysis were compared to the decisions made according to the 3D planning. The hypothetical choice of anatomic versus total shoulder was changed in 7 of 50 patients after the 3D planning.
In all cases where the decision was altered from an anatomic to a reverse prosthesis, 3D measurements demonstrated an increased glenoid retroversion of more than 8° compared with 2D.
In all cases where the decision was altered from a reverse to an anatomic prosthesis a decrease of more than 10° of glenoid inclination was observed on the 3D reconstructions.
Comment: In our practice, we do not depend on CT measurements to define the choice between anatomic and reverse total shoulders. Rather the choice is based on characteristics such as pseudoparalysis, instability, bone stock, and patient activity levels. We do not find arbitrary thresholds, such as 10° of inclination or 27° of inclination, to be of clinical utility.
Comment: In our practice, we do not depend on CT measurements to define the choice between anatomic and reverse total shoulders. Rather the choice is based on characteristics such as pseudoparalysis, instability, bone stock, and patient activity levels. We do not find arbitrary thresholds, such as 10° of inclination or 27° of inclination, to be of clinical utility.
In most cases the information necessary for prosthesis selection can be gathered from a good history, physical examination and standardized plain radiographs, avoiding the cost and radiation exposure of CT scans and complex reconstructions requiring proprietary software
