A modified Patte classification system for rotator cuff tendon retraction to predict reparability and tendon healing...

The purpose of this study was to propose a modified Patte classification system for tendon retraction, including the cut-off points for predicting reparability and rotator cuff healing after arthroscopic rotator cuff repair (ARCR) and assess its prediction accuracy and measurement reliability.

This retrospective study included 463 consecutive patients scheduled to undergo ARCR for full-thickness supraspinatus tears. Receiver operating characteristic (ROC) curve analysis was used to determine the cut-off points for predicting reparability and tendon healing. The modified Patte classification system, in which these cut-off points were combined with the original Patte classification, classified the tendon retraction as stages I-V. The prediction accuracy of reparability and tendon healing was assessed using the area under the curve (AUC). Measurement reliability was determined using Cohen's κ statistics.

Of the 402 included patients, 32 rotator cuff tears were irreparable and 71 of the remaining 370 were diagnosed with healing failure. ROC analysis determined the cut-off point of reparability at the medial one-fifth and that of tendon healing at the medial one-third of the humeral head. The AUC of the modified Patte classification for predicting reparability and tendon healing was 0.897 (excellent) and 0.768 (acceptable), respectively. Intra-rater reliability was almost perfect (mean κ value: 0.875), and inter-rater reliability was substantial (0.797).

Diagnostic performance of the modified Patte classification system was excellent for reparability and acceptable for rotator cuff healing, with high measurement reliability. The modified Patte classification system can be easily implemented in clinical practice for planning surgical procedures and counselling patients in the day-by-day clinical work.

Level III.