前交叉韧带断裂患者并发膝关节软骨损伤的列线图预测模型构建

Construction of nomogram prediction model for knee joint cartilage injury in patients with anterior cruciate ligament rupture

目的分析前交叉韧带断裂患者并发膝关节软骨损伤的相关因素并构建列线图预测模型。方法回顾性分析2020年3月至2023年2月于天津二七二医院和中国人民解放军联勤保障部队第九八三医院行手术治疗的160例单侧前交叉韧带断裂患者的临床资料。根据是否并发膝关节软骨损伤分为有损伤组97例和无损伤组63例。通过受试者工作特征(ROC)曲线分析各因素的最佳截断值;采用Logistic多元回归模型分析前交叉韧带断裂患者并发膝关节软骨损伤的独立危险因素;构建预测前交叉韧带断裂患者并发膝关节软骨损伤的列线图模型,其内部验证采用校正曲线,预测效能采用决策曲线进行评估。结果有损伤组患者体质量指数(BMI)、半月板损伤比例、扭伤次数、受伤时间均高于无损伤组[(24.15±2.52)kg/m^(2)比(22.84±3.13)kg/m^(2)、77.32%(75/97)比17.46%(11/63)、(2.64±0.90)次比(1.17±0.64)次、(19.15±3.77)d比(12.92±3.14)d],差异均有统计学意义(P<0.05)。ROC曲线分析结果显示:BMI、扭伤次数、受伤时间的最佳截断值分别为22.9 kg/m^(2)、1次、16 d。BMI(>22.9 kg/m^(2))、半月板损伤(有)、扭伤次数(>1次)、受伤时间(>16 d)是前交叉韧带断裂患者并发膝关节软骨损伤的独立危险因素,也是构建列线图模型的预测因子。内部验证结果显示,列线图模型预测前交叉韧带断裂患者并发膝关节软骨损伤的C-index为0.819(95%CI 0.715~0.883)。观测值与预测值间一致性良好。列线图模型预测前交叉韧带断裂患者并发膝关节软骨损伤的阈值>0.14,列线图模型提供的临床净收益均高于BMI、半月板损伤、扭伤次数、受伤时间。结论基于BMI、半月板损伤、扭伤次数、受伤时间,构建了预测前交叉韧带断裂患者并发膝关节软骨损伤的列线图模型,该模型对于前交叉韧带断裂患者并发膝关节软骨损伤有较好的预测价值,可用于识别前交叉韧带断裂患者中易并发膝关节软骨损伤的高危患者。

Objective To analyze the relevant factors of knee joint cartilage injury in patients with anterior cruciate ligament rupture and construct a nomogram prediction model.Methods The clinical data of 160 patients with unilateral anterior cruciate ligament rupture who underwent surgical treatment from March 2020 to February 2023 at Tianjin 272 Hospital and the Ninety-Eighty-Third Hospital of the People′s Liberation Army Joint Logistics Support Force were retrospectively analyzed.The patients were divided into injured group(97 cases)and non injured group(63 cases)based on whether there was concurrent knee joint cartilage injury.The optimal cutoff values of each factor were analyzed by the receiver operating characteristic(ROC)curve.Using a multiple Logistic regression model to analyze the independent risk factors of knee joint cartilage injury in patients with anterior cruciate ligament rupture;construct a nomogram model for predicting knee joint cartilage injury in patients with anterior cruciate ligament rupture.The internal validation of the nomogram model was validated using calibration curves,and the predictive performance of the nomogram model is evaluated using decision curves.ResultsThe body mass index(BMI),rate of meniscus injury,number of sprains and injury time in injured group were significantly higher than those in non injured group:(24.15±2.52)kg/m^(2) vs.(22.84±3.13)kg/m^(2),77.32%(75/97)vs.17.46%(11/63),(2.64±0.90)times vs.(1.17±0.64)times,(19.15±3.77)d vs.(12.92±3.14)d,and there were statistical differences(P<0.05).The ROC curve analysis results show that the optimal cutoff values for BMI,number of sprains and injury time were 22.9 kg/m^(2),once and 16 d,respectively.BMI(>22.9 kg/m^(2)),meniscus injury(with),number of sprains(>1 time)and injury time(>16 d)were independent risk factors for knee joint cartilage injury in patients with anterior cruciate ligament rupture,and they were also predictive factors for building nomogram model.The internal validation results show that the nomogram model predicts a C-index of 0.819(95%CI 0.715 to 0.883)for patients with anterior cruciate ligament rupture complicated by knee cartilage injury.The consistency between the observed values and the predicted values was good.The nomogram model predicts a threshold of over 0.14 for knee joint cartilage injury in patients with anterior cruciate ligament rupture,and the clinical net benefits provided by the column chart model were higher than BMI,meniscus injury,number of sprains and injury time.ConclusionsThis study constructs a nomogram model based on BMI,meniscus injury,number of sprains,and injury time to predict knee joint cartilage injury in patients with anterior cruciate ligament rupture.The model has good predictive value for knee joint cartilage injury in patients with anterior cruciate ligament rupture,and can be used to identify high-risk patients who are prone to knee joint cartilage injury in patients with anterior cruciate ligament rupture.