This study aimed to develop and validate a nomogram to predict probability of in-stent restenosis (ISR) in patients undergoing percutaneous coronary intervention (PCI). Patients undergoing PCI with drug-eluting stents between July 2009 and August 2011 were retrieved from a cohort study in a high-volume PCI center, and further randomly assigned to training and validation sets. The least absolute shrinkage and selection operator (LASSO) regression model was used to screen out significant features for construction of nomogram. Multivariable logistic regression analysis was applied to build a nomogram-based predicting model incorporating the variables selected in the LASSO regression model. The area under the curve (AUC) of the receiver operating characteristics (ROC), calibration plot and decision curve analysis (DCA) were performed to estimate the discrimination, calibration and utility of the nomogram model respectively. A total of 463 patients with DES implantation were enrolled and randomized in the development and validation sets. The predication nomogram was constructed with five risk factors including prior PCI, hyperglycemia, stents in left anterior descending artery (LAD), stent type, and absence of clopidogrel, which proved reliable for quantifying risks of ISR for patients with stent implantation. The AUC of development and validation set were 0.706 and 0.662, respectively, indicating that the prediction model displayed moderate discrimination capacity to predict restenosis. The high quality of calibration plots in both datasets demonstrated strong concordance performance of the nomogram model. Moreover, DCA showed that the nomogram was clinically useful when intervention was decided at the possibility threshold of 9%, indicating good utility for clinical decision-making. The individualized prediction nomogram incorporating 5 commonly clinical and angiographic characteristics for patients undergoing PCI can be conveniently used to facilitate early identification and improved screening of patients at higher risk of ISR.

中文翻译：

---

经皮冠状动脉介入治疗患者支架内再狭窄风险预测列线图的开发和验证

本研究旨在开发和验证列线图，以预测接受经皮冠状动脉介入治疗 (PCI) 的患者支架内再狭窄 (ISR) 的概率。2009 年 7 月至 2011 年 8 月期间接受药物洗脱支架 PCI 的患者从一个高容量 PCI 中心的队列研究中检索出来，并进一步随机分配到训练和验证集。使用最小绝对收缩和选择算子（LASSO）回归模型筛选出构建列线图的重要特征。应用多变量逻辑回归分析来构建基于列线图的预测模型，该模型包含在 LASSO 回归模型中选择的变量。受试者工作特征 (ROC) 的曲线下面积 (AUC)，进行校准图和决策曲线分析（DCA）以分别估计列线图模型的判别、校准和效用。共有 463 名 DES 植入患者被纳入开发和验证组并随机分组。预测列线图由五个风险因素构成，包括先前的 PCI、高血糖、左前降支 (LAD) 支架、支架类型和没有氯吡格雷，这被证明对于量化支架植入患者的 ISR 风险是可靠的。开发集和验证集的 AUC 分别为 0.706 和 0.662，表明预测模型显示出预测再狭窄的中等辨别能力。两个数据集中的高质量校准图证明了列线图模型具有很强的一致性性能。此外，DCA 表明，当以 9% 的可能性阈值决定干预时，列线图在临床上是有用的，表明对临床决策有很好的效用。个性化预测列线图结合了接受 PCI 患者的 5 种常见临床和血管造影特征，可方便地用于促进早期识别和改进对 ISR 高风险患者的筛查。