RESUMO
Background: Although some evidence suggests that machine learning algorithms may outperform classical statistical methods in prognosis prediction for several orthopaedic surgeries, to our knowledge, no study has yet used machine learning to predict patient-reported outcome measures after rotator cuff repair. Purpose: To determine whether machine learning algorithms using preoperative data can predict the nonachievement of the minimal clinically important difference (MCID) of disability at 2 years after rotator cuff surgical repair with a similar performance to that of other machine learning studies in the orthopaedic surgery literature. Study Design: Case-control study; Level of evidence, 3. Methods: We evaluated 474 patients (n = 500 shoulders) with rotator cuff tears who underwent arthroscopic rotator cuff repair between January 2013 and April 2019. The study outcome was the difference between the preoperative and 24-month postoperative American Shoulder and Elbow Surgeons (ASES) score. A cutoff score was calculated based on the established MCID of 15.2 points to separate success (higher than the cutoff) from failure (lower than the cutoff). Routinely collected imaging, clinical, and demographic data were used to train 8 machine learning algorithms (random forest classifier; light gradient boosting machine [LightGBM]; decision tree classifier; extra trees classifier; logistic regression; extreme gradient boosting [XGBoost]; k-nearest neighbors [KNN] classifier; and CatBoost classifier). We used a random sample of 70% of patients to train the algorithms, and 30% were left for performance assessment, simulating new data. The performance of the models was evaluated with the area under the receiver operating characteristic curve (AUC). Results: The AUCs for all algorithms ranged from 0.58 to 0.68. The random forest classifier and LightGBM presented the highest AUC values (0.68 [95% CI, 0.48-0.79] and 0.67 [95% CI, 0.43-0.75], respectively) of the 8 machine learning algorithms. Most of the machine learning algorithms outperformed logistic regression (AUC, 0.59 [95% CI, 0.48-0.81]); nonetheless, their performance was lower than that of other machine learning studies in the orthopaedic surgery literature. Conclusion: Machine learning algorithms demonstrated some ability to predict the nonachievement of the MCID on the ASES 2 years after rotator cuff repair surgery.
RESUMO
BACKGROUND: Few studies have evaluated the clinical results of surgical treatment of rotator cuff tears in young patients and most of the publications are only case series and with a small number of evaluated individuals. The aim of this study is to compare the clinical outcomes of patients undergoing arthroscopic rotator cuff repair according to age at the time of the procedure. HYPOTHESIS: Patients with 50 years of age or younger undergoing surgical treatment of rotator cuff tear have similar clinical outcomes to older patients. MATERIALS AND METHODS: Retrospective cohort study comparing results obtained after surgical treatment of rotator cuff tears between patients aged 50 years or younger and the older patients by the ASES and UCLA functional scales. Patients undergoing arthroscopy full-thickness rotator cuff repair were included. RESULTS: We evaluated 390 shoulders (377 patients), 94 aged 50 years or younger (median=46.5 years) and 296 aged over 50 years (median=60 years). Both groups significantly improved with the procedure after 24 months of follow-up, according to the ASES and UCLA scales (p<0.001). The groups did not differ in the scores obtained in the preoperative assessments and at 24 months of follow-up. The score obtained on the ASES scale at 24 months of follow-up had a median of 87.2 (IQR=38) among patients aged 50 years or younger and 90 points (IQR=26.4) among older patients (p=0.253). The scores obtained by the UCLA scale were 31 points (IQR=9) and 33 points (IQR =7) respectively (p=0.156). DISCUSSION: Our results showed that, after 24 months, the functional results of arthroscopic full-thickness rotator cuff repair did not differ between patients younger than 50 years and older patients. These results are similar to those found by other authors. Both groups of patients achieved significant improvement after the surgical procedure, achieving approximately 90 points on the ASES scale and 32 points on the UCLA scale. LEVEL OF EVIDENCE: III Retrospective cohort study.