椎动脉支架植入术后再狭窄列线图预测模型的构建及验证

目的]探讨椎动脉狭窄支架植入术后再狭窄发生的危险因素,从而构建列线图预测模型并进行验证。 [方法]收集2016年6月—2023年6月新疆医科大学第一附属医院收治的272例经数字减影血管造影(DSA)证实椎动脉狭窄并行支架植入术患者的临床资料并进行回顾性研究,根据行支架植入术的时间划分为建模组(2016年6月—2021年12月,220例)和验证组(2022年1月—2023年6月,52例)。在建模组中,随访期间根据CT血管造影(CTA)或者DSA结果将患者分为支架内狭窄(ISR)组(50例)和非支架内狭窄组(170例)。基于LASSO回归及多因素Logistic回归分析患者所发生ISR的独立危险因素构建列线图预测模型,并通过ROC曲线下面积(AUC)评价预测模型的预测能力,通过临床决策曲线评价预测模型的临床应用价值。 [结果]建模组中220例行椎动脉支架植入术的患者中有50例发生了ISR,ISR的发生率为22.72%。LASSO回归及多因素Logistic回归分析提示ESSEN卒中风险评分(ESRS)分值高、合并高同型半胱氨酸血症(HHcy)、合并颈内动脉和/或对侧椎动脉中度以上狭窄、低密度脂蛋白胆固醇(LDLC)水平≥1.8 mmol/L、术后椎动脉收缩期峰值流速(PSV)偏低及较小的支架直径是ISR发生的的危险因素。基于上述6个变量因素构建列线图预测模型,列线图预测椎动脉支架植入术后ISR发生建模组的AUC为0.857(95%CI:0.799～0.915),验证组的AUC为0.847(95%CI:0.732～0.961),提示模型具有较好的区分度。 [结论]本研究所建立的列线图预测模型可较好地预测椎动脉支架植入术患者发生ISR的危险程度,有助于临床医师发现ISR发生的高危患者,及时做出干预,使患者得到更大获益。

Construction and verification of prediction model of restenosis nomogram after vertebral artery stenting

Aim To investigate the risk factors of restenosis after stent implantation in patients with vertebral artery stenosis and construct the prediction model of nomogram. Methods The clinical data of 272 patients with digital subtraction angiography (DSA) confirmed vertebral artery stenosis and stent implantation admitted to the First Affiliated Hospital of Xinjiang Medical University from January 2016 to June 2023 were collected and retrospectively studied. According to the time of stent implantation, 272 patients were divided into modeling group (from January 2016 to December 1,0 cases) and verification group (from January 2022 to June 3,2 cases). In the modeling group, patients were divided into in-stent restenosis (ISR) group (50 cases) and non-ISR group (170 cases) according to CT angiography (CTA) or DSA results. Based on the independent risk factors of ISR analyzed by LASSO regression and multiple Logistic regression, a nomogram prediction model was constructed. The predictive ability of the prediction model was evaluated by the area under the receiver operating characteristic curve (AUC). The clinical application value of the prediction model was evaluated by clinical decision curve. Results In the modeling group, ISR occurred in 50 of 220 patients undergoing vertebral artery stenting, and the incidence of ISR was 22.72%. LASSO regression and multivariate Logistic regression analysis suggested that high ESSEN stroke risk score(ESRS), hyperhomocysteinemia (HHcy), moderate or higher stenosis of internal carotid artery and/or contralateral vertebral artery, low density lipoprotein cholesterol (LDLC) ≥1.8 mmol/L, low postoperative peak systolic velocity (PSV) of vertebral artery and small stent diameter were risk factors for ISR. A nomogram prediction model was built based on the above six variable factors. The nomogram predicted that the AUC of ISR after vertebral artery stenting was 0.857(95%CI:0.799～0.915) in the modeling group, and the AUC of the verification group was 0.847(95%CI:0.732～0.961), which suggested that the model had a good degree of differentiation. Conclusion The prediction model established in this study can better predict the risk degree of ISR in patients with metal stent implantation of vertebral artery, which is helpful for clinicians to find high-risk patients with ISR and make timely intervention, so that patients can get greater benefits.