【摘要】 目的 构建并验证基于双源双能CT(DSDECT)定量参数预测输尿管镜取石术(URL)清石率的列线图模型。方法 回顾性选取218例行URL治疗的肾结石患者，将患者按7:3的比例分为训练集(n=153)和验证集(n=65),将训练集根据术后1个月的结石清除情况分为残石组(n=43)和清石组(n=110)。通过多因素logistic回归分析筛选影响术后结石残留的独立危险因素，并基于这些因素建立列线图预测模型，分析预测模型的预测性能。结果 多因素logistic回归分析显示，鹿角型结石、多发结石、结石CT值、双能比、双能差及双能量指数(DEI)是URL术后结石残留的独立危险因素(P<0.05)。基于这些因素建立的列线图预测模型在训练集和验证集中均表现出良好的区分能力，曲线下面积(AUC)分别为0.92[95%置信区间(CI)0.88～0.95]和0.91(95%CI 0.87～0.93)。Hosmer-Lemeshow检验结果显示模型拟合良好(训练集：χ~2=4.352,P=0.315;验证集：χ~2=3.389,P=0.422),校准曲线进一步证实该模型校准度良好；决策曲线分析(DCA)结果显示，列线图预测模型预测URL患者术后结石残留的净收益更高。结论 基于DSDECT定量参数构建的预测模型预测URL患者术后结石残留具有良好的效能和显著的临床应用价值，筛选出的预测变量为鹿角型结石、多发结石、结石CT值、双能比、双能差以及DEI。更多还原

【Abstract】 Objective To construct and validate a nomogram model for predicting the stone clearance rate of ureteroscopic lithotomy(URL) based on quantitative parameters from dual-source dual-energy computed tomography(DSDECT). Methods A retrospective selection was made of 218 patients with kidney stones who underwent URL treatment.The patients were divided into the training set(n=153) and the validation set(n=65) in a 7:3 ratio.The training set was divided into the residual stone group(n=43) and the stone clearance group(n=110) according to the stone clearance at 1 month after operation.Multivariate logistic regression analysis was used to screen the independent risk factors affected postoperative residual stones,and a nomogram prediction model was established based on these factors to analyze the prediction performance of the prediction model. Results Multivariate logistic regression analysis showed that staghorn calculi,multiple calculi,stone CT value,dual-energy ratio,dual-energy difference and dualenergy index(DEI) were independent risk factors for residual stones after URL(P<0.05).The nomogram prediction model based on these factors showed good discrimination ability in the training set and the validation set,and the area under the curve(AUC)were 0.92 [95% confidence interval(CI)0.88-0.95]and 0.91(95% CI 0.87-0.93),respectively.The results of Hosmer-Lemeshow test showed that the model fitted well(training set:χ~2=4.352,P=0.315;validation set:χ~2=3.389,P=0.422).The calibration curve further confirmed that the model had good calibration.The results of decision curve analysis(DCA) showed that the nomogram prediction model had a higher net benefit in predicting postoperative residual stones in URL patients. Conclusion The prediction model based on DSDECT quantitative parameters demonstrated good performance and significant clinical application value in predicting postoperative residual stones in URL patients.The identified predictive variables are staghorn calculi,multiple calculi,stone CT value,dualenergy ratio,dual-energy difference and DEI.更多还原