Interdepartmental imaging protocol for clinically based three-dimensional computed tomography can provide accurate measurement of glenoid version

Abstract

Conventional computed tomography (CT) is not accurate for glenoid version measurement. This study sought to examine the feasibility of an interdepartmental protocol implemented between orthopedic surgery and radiology departments for acquisition of anatomic axial CT images and to validate the glenoid version measured through such a protocol. Data of 30 conventional CT scans of 10 normal and 20 osteoarthritic glenoids were transferred to clinical 3-dimensional imaging software by a radiology technician trained for the study. The technician independently reoriented the scapulae to generate anatomic CT images. A separate team of orthopedic researchers used laboratory-based 3-dimensional reconstruction software (Mimics; Materialise, Leuven, Belgium) to generate anatomic axial images. Three independent examiners measured glenoid version on the conventional CT, reoriented anatomic CT, and Mimics images at the superior, middle, and inferior levels. Data were analyzed using the Mimics data as the "gold standard." Reoriented anatomic CT images generated by the technician resulted in almost identical version measurements to the Mimics images in both normal and arthritic glenoids. The conventional CT images had poor agreement with the Mimics images in normal glenoids but had good agreement in arthritic glenoids. Both normal and arthritic glenoids had increased retroversion superiorly (P < .05), and this phenomenon was significantly exaggerated on the conventional CT images (P < .05). This study demonstrated that an interdepartmental protocol can produce reoriented anatomic axial CT images on which true glenoid version can be accurately measured. Such an institutional protocol would help surgeons accurately evaluate glenoid version preoperatively with reduced workload and expense.