Patients with an isolated diagnosis of rotator cuff impingement syndrome were prospectively entered into the study. Each of the 23 subjects was refractory to conservative therapy, had preoperative roentgenograms, and underwent an open acromioplasty. The roentgenograms included anteroposterior, axillary, 30 degrees caudal tilt, and supraspinatus outlet views. The roentgenograms were measured by four independent readers. The separate views were then scored for reliability, and the correlation of the measurements with intraoperative acromial measurements was assessed. Interobserver reliability was highest for the caudal tilt view (0.84) and lowest for the axillary view (0.09). The supraspinatus and caudal tilt views correlated significantly with distinct intraoperative measurements of acromial spur size. We continue to advocate the evaluation of both views for preoperative assessment of the acromial spur in the rotator cuff impingement syndrome.