Kunze KN, Polce EM, Ranawat AS, Randsborg PH, Williams RJ 3rd, Allen AA, Nwachukwu BU, Pearle A, Stein BS, Dines D, Kelly A, Kelly B, Rose H, Maynard M, Strickland S, Coleman S, Hannafin J, MacGillivray J, Marx R, Warren R, Rodeo S, Fealy S, O'Brien S, Wickiewicz T, Dines JS, Cordasco F, and Altcheck D
Background: Understanding specific risk profiles for each patient and their propensity to experience clinically meaningful improvement after anterior cruciate ligament reconstruction (ACLR) is important for preoperative patient counseling and management of expectations., Purpose: To develop machine learning algorithms to predict achievement of the minimal clinically important difference (MCID) on the International Knee Documentation Committee (IKDC) score at a minimum 2-year follow-up after ACLR., Study Design: Case-control study; Level of evidence, 3., Methods: An ACLR registry of patients from 27 fellowship-trained sports medicine surgeons at a large academic institution was retrospectively analyzed. Thirty-six variables were tested for predictive value. The study population was randomly partitioned into training and independent testing sets using a 70:30 split. Six machine learning algorithms (stochastic gradient boosting, random forest, neural network, support vector machine, adaptive gradient boosting, and elastic-net penalized logistic regression [ENPLR]) were trained using 10-fold cross-validation 3 times and internally validated on the independent set of patients. Algorithm performance was assessed using discrimination, calibration, Brier score, and decision-curve analysis., Results: A total of 442 patients, of whom 39 (8.8%) did not achieve the MCID, were included. The 5 most predictive features of achieving the MCID were body mass index ≤27.4, grade 0 medial collateral ligament examination (compared with other grades), intratunnel femoral tunnel fixation (compared with suspensory), no history of previous contralateral knee surgery, and achieving full knee extension preoperatively. The ENPLR algorithm had the best relative performance (C-statistic, 0.82; calibration intercept, 0.10; calibration slope, 1.15; Brier score, 0.068), demonstrating excellent predictive ability in the study's data set., Conclusion: Machine learning, specifically the ENPLR algorithm, demonstrated good performance for predicting a patient's propensity to achieve the MCID for the IKDC score after ACLR based on preoperative and intraoperative factors. The femoral tunnel fixation method was the only significant intraoperative variable. Range of motion and medial collateral ligament integrity were found to be important physical examination parameters. Increased body mass index and prior contralateral surgery were also significantly predictive of outcome., Competing Interests: One or more of the authors has declared the following potential conflict of interest or source of funding: A.S.R. has received research support from DePuy and Stryker; consulting fees from Ceramtec, Medtronic, Moximed, Smith & Nephew, and Stryker; speaking fees from Ceramtec, Medtronic, Smith & Nephew, and Stryker Mako; and royalties from DePuy, Saunders/Mosby–Elsevier, Springer, and Stryker Mako and has stock/stock options in ConforMIS and Enhatch. R.J.W. has received research support from Histogenics; consulting fees from Arthrex, JRF Ortho, and Lipogems; royalties from Arthrex; hospitality payments from Stryker and has stock/stock options in BICMD, Cymedica, Engage Surgical, Gramercy Extremity Orthopedics, Pristine Surgical, and RecoverX. A.A.A. has received consulting fees from Arthrex and has stock/stock options in Pristine Surgical and Rom3. B.U.N. has received royalties from Remote Health. D.D. has received consulting fees and royalties from Zimmer Biomet. A.K. has received education payments from Arthrex. B.K. has received consulting and nonconsulting fees and royalties from Arthrex and speaking fees from Synthes GmbH. M.M. has received education payments from Arthrex. S.S. has received consulting fees from DePuy, Flexion Therapeutics, Pfizer, and Vericel and honoraria from JRF Ortho and Vericel. S.C. has received education payments from Pinnacle, consulting fees from Stryker, and nonconsulting fees from Smith & Nephew. J.H. has received hospitality payments from Smith & Nephew. R.W. has received royalties from Zimmer Biomet and nonconsulting fees from Arthrex. S.R. has received consulting fees from Flexion Therapeutics, nonconsulting fees from Smith & Nephew, honoraria from Fidia Pharma, and royalties from Zimmer Biomet and is a paid associate editor for The American Journal of Sports Medicine. S.F. has received royalties from Encore Medical. J.S.D. has received consulting fees from Arthrex, Merck Sharp & Dohme, Trice Medical, and Wright Medical; speaking fees from Arthrex; and royalties from Linvatec. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto., (© The Author(s) 2021.)