Can Deltoid Hounsfield Unit/Supraspinatus Hounsfield Unit Predict Postoperative Rotator Cuff Retear?

Retear is a common complication after arthroscopic rotator cuff repair (ARCR), yet available predictors are limited and combined models perform better than single factors. This study investigated whether the preoperative deltoid-to-supraspinatus Hounsfield unit (HU) ratio predicts retear and integrated it with established risk factors to develop an optimal predictive model.

This retrospective study included 243 patients with full-thickness rotator cuff tears who underwent arthroscopic repair from July 2021 to January 2025. Preoperative CT and MRI were obtained for all patients; postoperative MRI at 3-6 months evaluated tendon healing using the Sugaya classification (grades I-III: intact healing, n = 208; grades IV-V: retear, n = 35). Statistical analyses used SPSS. After normality testing, continuous variables were compared using t-test or Mann-Whitney U test, and categorical variables using chi-square or Fisher's exact test. Variables with group differences were screened by elastic net regression, then entered into binary logistic regression. ROC analysis determined optimal cutoffs, and 1000 bootstrap samples provided an optimism-corrected C-statistic for internal validation.

Multivariate analysis identified supraspinatus fatty infiltration (FI) (p < 0.01), the supraspinatus occupation ratio (SspOR) (p < 0.01), and the deltoid-to-supraspinatus HU ratio (D/S HU) (p < 0.01) as independent predictors of rotator cuff retear. In the univariate analysis, tear width, preoperative Patte stage, and the presence of concomitant frozen shoulder differed significantly between the intact healing and retear groups; however, none of these variables remained independent predictors in the multivariate model. The D/S HU was significantly lower in the intact healing group than in the retear group (0.87 ± 0.17 vs. 1.15 ± 0.29), further supporting its value as an independent predictor of retear. The optimal cutoff values for FI, SspOR, and the D/S HU were Goutallier Stage 2, 46.02%, and 1.02, respectively. The combination of a D/S HU > 1.02 and FI ≥ Goutallier Stage 2 provided the best predictive performance, with an area under the curve (AUC) of 0.868, sensitivity of 77.1%, specificity of 81.2%, and accuracy of 89.3%.

The D/S HU was significantly associated with postoperative rotator cuff retear, and a ratio greater than 1.02 was identified as an effective threshold for predicting retear. Moreover, the combination of a D/S HU > 1.02 and FI ≥ Goutallier Stage 2 yielded the strongest predictive performance. This preoperative risk assessment model may help surgeons optimize treatment planning and better anticipate postoperative outcomes.