The acromiohumeral distance (AHD) measured on conventional, true anteroposterior radiographs taken with the patient standing and the arm held in neutral rotation can predict the presence or absence of a large, chronic rotator cuff tears, and narrowing of the acromiohumeral distance can predict irreparability of rotator cuff tears. The relation between AHD measured on conventional radiographs and the integrity of the rotator cuff tendons is well known. Conversely, the correlation between the AHD on conventional radiographs and that measured on magnetic resonance imaging (MRI) or computed tomography (CT) scans is unknown. It has also not been determined which method correlates better with fatty infiltration of the rotator cuff muscles and might thereby be a better prognostic value for the outcome of rotator cuff repair. Conventional radiographs, plus MRI and or CT studies taken at the same time of patients with rotator cuff tears, shoulder instability, primary glenohumeral osteoarthritis, or idiopathic frozen shoulder, were randomly selected, and the AHD was measured both on conventional radiographs and on MRI or CT. The values were correlated with the stage of fatty degeneration of the rotator cuff muscles in the cases of rotator cuff tears. The AHD on conventional radiographs was larger than on MRI and CT, and correlation of the values measured (conventional radiography vs MRI and CT) could be demonstrated. For rotator cuff diseases, a significant but moderate correlation (r = 0.6; P < .05) between conventional and MRI was identified, and the conventional values were approximately 0.6 mm greater than the corresponding MRI values. It is, therefore, possible to use the AHD read on MRI or CT studies for prognostic reasons, provided the values obtained are converted accordingly and remembering that the correlation of the values is only moderate.