Development and Validation of a Prognostic Nomogram for Prognosis in Patients With Renal Artery Stenosis

Abstract

Background and ObjectiveRenal artery stenosis (RAS) is associated with an increased risk of renal function deterioration (RFD). Our previous study showed that renal cortical blood perfusion assessed by contrast-enhanced ultrasound (CEUS) was an important related factor for RFD in RAS patients. Based on several conventional related factors confirmed by previous studies, we aimed to establish and verify a CEUS+ scoring system to evaluate the risk of RFD at 1 year of follow-up in RAS patients.MethodsThis study was a single-center retrospective study. A total of 497 elderly RAS patients (247 in the training group and 250 in the verification group) admitted to the Beijing Hospital from January 2016 to December 2019 were included. The baseline characteristics of the patients on admission (including general conditions, previous medical history, blood pressure, blood creatinine, RAS, and cortical blood perfusion in the affected kidney) and renal function [glomerular filtration rate (GFR)] at 1-year of follow-up were collected. We used the univariate and multivariate logistic regressions to establish a CEUS+ scoring system model, the receiver operating characteristic (ROC) curve and area under the curve (AUC) to evaluate prediction accuracy, and the decision curve analysis and nomogram to evaluate the clinical application value of CEUS+ scoring system model.ResultsAmong the 497 patients enrolled, 266 (53.5%) were men, with an average age of (51.7 ± 19.3) years. The baseline clinical-radiomic data of the training group and the verification group were similar (all p > 0.05). Multivariate logistic regression analysis results showed that age [Odds ratio (OR) = 1.937, 95% confidence interval (CI): 1.104–3.397), diabetes (OR = 1.402, 95% CI: 1.015–1.938), blood pressure (OR = 1.575, 95% CI: 1.138–2.182), RAS (OR = 1.771, 95% CI: 1.114–2.816), and area under ascending curve (AUCi) (OR = 2.131, 95% CI: 1.263–3.596) were related factors for the renal function deterioration after 1 year of follow-up (all p < 0.05). The AUC of the ROC curve of the CEUS+ scoring system model of the training group was 0.801, and the Youden index was 0.725 (specificity 0.768, sensitivity 0.813); the AUC of the ROC curve of the validation group was 0.853, Youden index was 0.718 (specificity 0.693, sensitivity 0.835). There was no significant difference in ROC curves between the two groups (D = 1.338, p = 0.325). In addition, the calibration charts of the training and verification groups showed that the calibration curve of the CEUS+ scoring system was close to the standard curve (p = 0.701, p = 0.823, both p > 0.10).ConclusionThe CEUS+ scoring system model is helpful in predicting the risk of worsening renal function in elderly RAS patients.