#0758

Prediction model for postoperative acute kidney injury and chronic kidney disease in patients with renal cell carcinoma and venous tumor thrombus

Q. Jiao¹, X. Zhang^2,1

¹Nankai University, School of Medicine, Tianjin, China
²The Third Medical Center, Chinese PLA General Hospital, Department of Urology, Beijing, China

Introduction:

To develop nomograms to predict the risk of postoperative acute kidney injury (PO-AKI) and chronic kidney disease (CKD) in patients with renal cell carcinoma and venous tumor thrombus (RCC-VTT).

Material and methods:

A total of 353 consecutive postoperative patients with RCC-VTT were enrolled between Jan 2006 and June 2023. Clinicopathological, operative data and functional outcomes were collected and analyzed. Logistic regression was employed to develop predictive models incorporating risk factors. Model performance was assessed using the area under the curve (AUC), calibration, and decision curve analysis.

Results:

Among the 353 patients, the incidence of PO-AKI was 61.5%, with stage 2-3 PO-AKI observed in 15.0% overall and in 4.0% of cases with RCC and renal venous tumor thrombus, as well as 21.1% of RCC with inferior vena cava (IVC) tumor thrombus. Independent risk factors for PO-AKI included age >60 years, male sex, contralateral renal artery and IVC clamping time >20 minutes, BMI and preoperative eGFR. Tumor size was identified as the only independent protective factor against PO-AKI. Of 297 patients, 83 (27.9%) developed CKD, with BMI >24 kg/m2, age >60 years, hypertension, IVC cavectomy, PO-AKI, and adjuvant therapy identified as independent risk factors, preoperative eGFR emerged as the only independent protective factor. The AUCs for PO-AKI and CKD predictive models were 0.776 (95% CI, 0.728-0.824) and 0.863 (95% CI, 0.817-0.908), respectively, demonstrating good calibration and predictive accuracy.