A predictive model for sepsis-associated acute kidney injury based on CEUS and biomarkers

Background: Sepsis-associated acute kidney injury (SA-AKI) is a common and serious complication in critically ill patients, leading to increased morbidity and mortality. This study aimed to develop a predictive model for early identification of SA-AKI based on contrast-enhanced ultrasound (CEUS) and biomarkers. Methods: This retrospective observational study included 152 septic patients admitted to the Surgical Intensive Care Unit (SICU) of Zhongshan Hospital from January 2021 to June 2022. Patients were divided into training and validation cohorts. Clinical data, CEUS-derived parameters, and biomarkers were collected within 24 hours of admission. A prediction model was constructed using LASSO and multivariate logistic regression. Model performance was assessed by AUROC and calibration. Results: SA-AKI occurred in 86 patients (56.6%). Four independent predictors—urinary neutrophil gelatinase-associated lipocalin (NGAL), serum cystatin C, renal resistive index (RRI), and medullary wash-in slope (WIS)—were identified. The model demonstrated excellent discriminative performance, with AUROCs of 0.956 in the training cohort and 0.927 in the validation cohort. Sensitivity and specificity were 82.8% and 91.1% in the training cohort, and 75.0% and 88.9% in the validation cohort. DCA demonstrated favorable net clinical benefit across a wide range of thresholds. Conclusion: We developed a robust, noninvasive model incorporating CEUS-derived microcirculatory metrics and renal biomarkers for early SA-AKI prediction in septic ICU patients. This tool may support timely risk stratification and intervention. External validation in multicenter prospective cohorts is warranted to confirm its clinical utility.