Preoperative acromiohumeral index and the Goutallier classification reliably predict rotator cuff integrity using computed tomography and magnetic resonance imaging as confirmed intraoperatively in shoulder arthroplasty patients

Hypothesis We hypothesized that both computed tomography (CT) and magnetic resonance imaging (MRI) could be used reliably to measure the acromiohumeral index (AHI) and supraspinatus muscle atrophy through the use of Goutallier classification. Additionally, we hypothesized that these measurements could be used in combination to predict rotator cuff status, as verified by intraoperative assessment. Methods This was a retrospective case series of a prospectively maintained single institution’s shoulder arthroplasty registry. Patients who underwent anatomic total shoulder arthroplasty or reverse total shoulder arthroplasty between March 2015 and July 2020 were eligible for inclusion if they had X-ray (XR), MRI, and CT scans performed within 6 months of surgery and rotator cuff integrity documented intraoperatively. AHI was measured for each patient using XR, MRI, and CT. Goutallier scores were assigned using both MRI and CT for supraspinatus muscle quality. Rotator cuff integrity was confirmed intraoperatively by the operating surgeon and reported to be (1) intact/attenuated or (2) torn. Intraclass correlation coefficients were calculated for each measurement and imaging modality and cut-point analysis was performed for predicting rotator cuff integrity. Results Of the patients, a total of 199 (132 anatomic total shoulder arthroplasty, 67 reverse total shoulder arthroplasty) met the inclusion criteria. Intraclass correlation coefficient values were as follows—Goutallier on CT: 0.76, Goutallier on MRI: 0.73, AHI on MRI: 0.81, AHI on XR: 0.76, and AHI on CT: 0.72. Median AHI measurements for patients with intact rotator cuffs were 7.50 mm on MRI, 7.80 mm on CT, and 8.65 mm on XR, whereas median AHI measurements for patients with torn rotator cuffs were 5.45 mm on MRI, 4.93 mm on CT, and 7.55 mm on XR. These intramodality differences were significantly different ( P < .001). Goutallier measurements on MRI and CT were significantly positively correlated (R = 0.93, P < 2.2e-16). The optimal cut point for determining rotator cuff integrity using AHI was 6.40 mm on MRI, 5.75 mm on CT, and 7.90 mm on XR. The optimal cut point for predicting rotator cuff integrity when using Goutallier score was 1.50 on MRI and 1.00 on CT. In the diagnosis of a rotator cuff tear (partial or full-thickness), MRI sensitivity was 92.7% and specificity was 33.1%. Conclusion Both CT and MRI provide reliable assessments of supraspinatus muscle quality and AHI in shoulder arthroplasty candidates. Median AHI measurements are significantly lower when using CT and MRI and these differences are amplified in the case of rotator cuff insufficiency. Furthermore, both CT and MRI predict intraoperative rotator cuff status more reliably than XR when using cut points for Goutallier scores and AHI measurements.