Zhan Hongwei, Kang Xin, Zhang Xiaobo, Zhang Yuji, Wang Yanming, Yang Jing, Zhang Kun, Han Jingjing, Feng Zhiwei, Zhang Liang, Wu Meng, Xia Yayi, Jiang Jin
Department of Sports Medicine, Honghui Hospital, Xi'an Jiaotong University, Xi'an, Shaanxi, China. Electronic address: https://facebook.com/100091611350229.
Department of Sports Medicine, Honghui Hospital, Xi'an Jiaotong University, Xi'an, Shaanxi, China.
Arthroscopy. 2025 Jun;41(6):1896-1908.e2. doi: 10.1016/j.arthro.2024.07.028. Epub 2024 Aug 10.
To develop a machine-learning model to predict clinical outcomes after medial patellofemoral ligament reconstruction (MPFLR) and identify the important predictive indicators.
This study included patients who underwent MPFLR from January 2018 to December 2022. The exclusion criteria were as follows: (1) concurrent bony procedures, (2) history of other knee surgeries, and (3) follow-up period of less than 12 months. Forty-two predictive models were constructed for 7 clinical outcomes (failure to achieve minimum clinically important difference of clinical scores, return to preinjury sports, pivoting sports, and recurrent instability) using 6 machine-learning algorithms (random forest, logistic regression, support vector machine, decision tree, implemented multilayer perceptron, and K-nearest neighbor). The performance of the model was evaluated using metrics such as the area under the receiver operating characteristic curve, accuracy, specificity, and sensitivity. In addition, SHapley Additive exPlanation summary plot was employed to identify the important predictive factors of the best-performing model.
A total of 218 patients met criteria. For the best-performing models in predicting failure to achieve the minimum clinically important difference for Lysholm, International Knee Documentation Committee, Kujala, and Tegner scores, the area under the receiver operating characteristic curves and accuracies were 0.884 (good) and 87.3%, 0.859 (good) and 86.2%, 0.969 (excellent) and 97.0%, and 0.760 (fair) and 76.8%, respectively; 0.952 (excellent) and 95.2% for return to preinjury sports; 0.756 (fair) and 75.4% for return to pivoting sports; and 0.943 (excellent) and 94.9% for recurrent instability. Low preoperative Tegner score, shorter time to surgery, and absence of severe trochlear dysplasia were significant predictors for return to preinjury sports, whereas the absence of severe trochlear dysplasia and patellar alta were significant predictors for return to pivoting sports. Older age, female sex, and low preoperative Lysholm score were highly predictive of recurrent instability.
The predictive models developed using machine-learning algorithms can reliably forecast the clinical outcomes of MPFLR, particularly demonstrating excellent performance in predicting recurrent instability.
Level III, case-control study.
开发一种机器学习模型，以预测髌股内侧韧带重建（MPFLR）后的临床结果，并确定重要的预测指标。
本研究纳入了2018年1月至2022年12月期间接受MPFLR的患者。排除标准如下：（1）同期进行骨手术；（2）有其他膝关节手术史；（3）随访期少于12个月。使用6种机器学习算法（随机森林、逻辑回归、支持向量机、决策树、实现的多层感知器和K近邻）为7种临床结果（未能达到临床评分的最小临床重要差异、恢复到伤前运动水平、旋转运动、复发性不稳定）构建了42个预测模型。使用受试者操作特征曲线下面积、准确性、特异性和敏感性等指标评估模型的性能。此外，采用SHapley加性解释汇总图来确定表现最佳模型的重要预测因素。
共有218例患者符合标准。对于预测未能达到Lysholm、国际膝关节文献委员会、Kujala和Tegner评分的最小临床重要差异的最佳模型，受试者操作特征曲线下面积和准确性分别为0.884（良好）和87.3%、0.859（良好）和86.2%、0.969（优秀）和97.0%、0.760（中等）和76.8%；恢复到伤前运动水平的模型的受试者操作特征曲线下面积和准确性分别为0.952（优秀）和95.2%；恢复到旋转运动水平的模型的受试者操作特征曲线下面积和准确性分别为0.756（中等）和75.4%；复发性不稳定的模型的受试者操作特征曲线下面积和准确性分别为0.943（优秀）和94.9%。术前Tegner评分低、手术时间短和无严重滑车发育不良是恢复到伤前运动水平的重要预测因素，而无严重滑车发育不良和髌骨高位是恢复到旋转运动水平的重要预测因素。年龄较大、女性和术前Lysholm评分低是复发性不稳定的高度预测因素。
使用机器学习算法开发的预测模型可以可靠地预测MPFLR的临床结果，特别是在预测复发性不稳定方面表现出色。
III级，病例对照研究。
