脊柱骨质疏松性压缩性骨折经皮穿刺椎体成形后骨水泥渗漏风险分析及列线图预测模型的构建

Risk analysis of bone cement leakage after percutaneous puncture vertebroplasty for osteoporotic spinal compression fractures and construction of a predictive model with column line drawings

目的 探讨骨质疏松性脊柱压缩性骨折患者经皮穿刺椎体成形术治疗后骨水泥渗漏风险及列线图预测模型的构建。方法 选取2019年12月~2021年12月广元市第二人民医院就诊的236例骨质疏松性脊柱压缩性骨折患者为研究对象,根据经皮穿刺椎体成形术治疗后是否发生骨水泥渗漏分为渗漏组(n=58)和未渗漏组(n=178)。收集患者临床资料,分析骨水泥渗漏的相关因素;绘制受试者工作特征(ROC)曲线,分析各相关因素对骨水泥渗漏的预测价值;采用Logistic多元回归模型分析影响骨水泥渗漏的危险因素;采用R语言软件4.0“rms”包构建列线图预测模型。结果 渗漏组与未渗漏组患者在年龄、骨密度、椎体压缩程度、椎体终板/后壁完整、骨水泥黏度和骨水泥注射量方面的差异具有统计学意义(P<0.05)。年龄、骨密度、骨水泥注射量预测骨水泥渗漏的曲线下面积(AUC)分别为0.804、0.825、0.803;最佳截断值分别为71岁、0.67 g/cm~2、4.4 ml。年龄(>71岁)、骨密度(≤0.67 g/cm~2)、椎体压缩程度(重度)、椎体终板/后壁完整(否)、骨水泥黏度(低黏度)、骨水泥注射量(>4.4 ml)是骨水泥渗漏的独立危险因素。列线图模型预测骨水泥渗漏C-index为0.802(95%CI:0.689~0.868),阈值>0.19,列线图模型提供临床净收益。结论 年龄、骨密度、椎体压缩程度、椎体终板/后壁完整、骨水泥黏度、骨水泥注射量是骨水泥渗漏的独立危险因素,以此为预测因子构建的列线图预测模型具有临床应用价值。

Objective To analysis risk factor and to construct a line graph prediction model for bone cement leakage after percutaneous transluminal vertebroplasty treatment in patients with osteoporotic spinal compression fractures. Methods A total of 236 patients with osteoporotic spinal compression fractures who came to our hospital from December 2019 to December 2021 were selected for the study, and they were divided into a leakage group(n=58) and a non-leakage group(n=178) according to whether bone cement leakage occurred after percutaneous transluminal vertebroplasty treatment. The clinical data were collected to analyze the factors associated with bone cement leakage;The work receiver operating characteristic(ROC) curves of the subjects were drawn to analyze the predictive value of each relevant factor on bone cement leakage;The Logistic multiple regression model was used to analyze the risk factors affecting bone cement leakage;The R language software 4.0 “rms” package was used to construct the prediction model of column line diagram. Results The differences in age, bone density, degree of vertebral compression, vertebral endplate/posterior wall integrity, bone cement viscosity, and bone cement injection volume between patients in the leaky and non-leaky groups were statistically significant(P< 0.05). The area under curve(AUCs) for age, bone density, and cement injection volume to predict cement leakage were 0.804, 0.825, and 0.803, respectively;The best cutoff values were 71 years, 0.67 g/cm~2, and 4.4 ml, respectively. Age(>71 years), bone density(≤0.67 g/cm~2), vertebral compression(severe), vertebral endplate/posterior wall integrity(no), cement viscosity(low viscosity), and bone cement injection volume(>4.4 ml) were independent risk factors for bone cement leakage. The column line graph model predicted a C-index of 0.802(95% CI, 0.689-0.868) for cement leakage, with a threshold >0.19, and the column line graph model provided a net clinical benefit. Conclusion Age, bone density, degree of vertebral compression, vertebral endplate/posterior wall integrity, cement viscosity, and cement injection volume are independent risk factors for cement leakage, and the column line graph prediction model constructed with these predictors is of clinical application.