颅骨缺损三维塑型钛网颅骨成形术后并发症列线图风险预测模型的构建

Construction of a risk prediction model for postoperative complications in cranioplasty with three-dimensional plastic titanium mesh for skull defect

目的构建颅骨缺损三维塑型钛网颅骨成形术后并发症的列线图风险预测模型。方法通过简单随机抽样法选取2018-03-01—2023-02-28在安徽医科大学附属六安医院行三维塑型钛网颅骨成形术治疗的颅骨缺损患者156例,对其临床资料进行回顾性分析,根据术后并发症发生情况分为并发症组(24例)和无并发症组(132例),通过单因素方差分析2组一般资料,将有统计学差异的指标纳入二元Logistic回归模型中,分析影响术后并发症的相关因素,并构建预测模型。结果二元Logistic回归分析显示,骨窗凹陷(轻度凹陷/明显凹陷)、去骨瓣减压至修补时间、颅骨缺损面积是影响颅骨缺损三维塑型钛网颅骨成形术后并发症发生的危险因素(P<0.05)。纳入上述因素构建研究模型的整体预测准确率为91.7%,Hosmer-Lemeshow拟合度检验分析显示其具备较好拟合优度(χ^(2)=15.321,P=0.053)。进一步Bootstrap法内验证研究模型预测AUC为0.771,敏感度为100.0%,特异性为51.5%。结论骨窗凹陷、去骨瓣减压至修补时间、颅骨缺损面积是影响颅骨缺损三维塑型钛网颅骨成形术后并发症发生的危险因素,基于上述因素建立的风险预测模型具有较好预测效能。

Objective To construct a nomogram risk prediction model for postoperative complications in cranioplasty with three-dimensional plastic titanium mesh for skull defects.Methods The clinical data of 156 patients with skull defect admitted to Lu’an Hospital Affiliated to Anhui Medical University underwent 3D plastic titanium mesh cranioplasty from March 1,2018 to February 28,2023 were retrospectively analyzed,and were divided into complicated group(24 cases)and uncomplicated group(132 cases)according the occurrence of post-operative complications.The general data were analyzed through one-way variance analysis,and the indicators with statistical differences were included in the binary Logistic regression model,the related factors affecting postoperative complications were analyzed,and the prediction model was constructed.Results Binary Logistic regression analysis showed that bone window depression(mild depression/obvious depression),decompression time to repair,and skull defect area were risk factors affecting the occurrence of complications after three-dimensional titanium mesh skull plasty(P<0.05).The overall prediction accuracy of the model constructed by including the above factors was 91.7%,and the Hosmer-Lemeshow fit test analysis showed a good goodness of fit(χ^(2)=15.321,P=0.053).A further internal Bootstrap method validation study model predicted the AUC of 0.771,the sensitivity was 100.0%and the specificity was 51.5%.Conclusion Bone window depression,bone valve decompression time to repair,and skull defect area are risk factors that affect the occurrence of complications after three-dimensional skull plasty,and the risk prediction model established based on the above factors has good predictive efficacy.