HYPOTHESIS Glenoid orientation likely plays an important role in shoulder mechanics and has been associated with glenohumeral instability, rotator cuff tears, and total shoulder arthroplasty outcome. Glenoid version and inclination measurements are widely used, but these measurements are 2-dimensional, and orientation of the central face must be inferred from 2 points on the outer rim. This study tested the hypothesis that sphere fitting provides an optimal assessment of glenoid face 3-dimensional orientation, and clinically important differences will exist in some individuals between sphere fit orientation and traditional version measurement. MATERIALS AND METHODS Computed tomography scans of 20 normal glenoids were obtained. Glenoid version was measured from resliced, pure transverse images at a series of glenoid heights. Separately, the 3-dimensional surface of each glenoid face was reconstructed. A sphere was fit to the glenoid face, and its orientation was described by 2 angles analogous to version and inclination. RESULTS "Sphere fit version" averaged -3.2° ± 3.4° (negative indicating retroversion), and "sphere fit inclination" averaged 1.3° ± 4.7°. Absolute differences between sphere fit version and the standard mid-glenoid version averaged 1.5° (maximum 4.0°) across patients. The glenoids were, on average, 5.5° more retroverted at the 80% height than at the 20% height. DISCUSSION Three-dimensional aspects of the glenoid, including the spiraling twist evidenced by increased retroversion superiorly, are accounted for in this novel sphere-fitting approach for assessing glenoid orientation. CONCLUSIONS For most normal glenoids, midglenoid version appears to provide an adequate measure of glenoid central face orientation. Sphere fitting has potential utility in optimizing glenoid implant alignment in total shoulder arthroplasty.