Computed tomography radiomics-derived nomogram for prediciting early renal function decline after partial nephrectomy...

Area of Science:

• Radiology
• Oncology
• Nephrology

Background:

• Localized renal cell carcinoma (RCC) necessitates surgical intervention, often partial nephrectomy (PN), with potential for early renal function decline.
• Accurate prediction of postoperative renal function is crucial for patient management and surgical planning.

Purpose of the Study:

• To develop and validate CT-based radiomics-clinical nomograms for predicting early renal function decline after PN in localized RCC patients.
• To integrate dual-region radiomics features (tumor and normal kidney) with clinical factors for enhanced predictive accuracy.

Main Methods:

• A multicenter retrospective study of 1440 localized RCC patients undergoing PN.
• Extraction of radiomics features from tumor and ipsilateral normal renal parenchyma across CT phases, followed by LASSO regression.
• Development of preoperative and preoperative-intraoperative nomograms combining radiomics signatures and clinical predictors, validated internally and externally.

Main Results:

• Selected radiomics features from both tumor and normal kidney demonstrated significant predictive ability.
• The preoperative-intraoperative nomogram achieved high discrimination (AUC up to 0.962) and robust calibration.
• Key predictors included age, diabetes, eGFR, RENAL score, ischemia time, and radiomics signature.

Conclusions:

• CT-based radiomics signatures integrated with clinical parameters accurately predict early renal function decline post-PN.
• The developed nomograms offer a valuable tool for individualized risk stratification and optimizing surgical decisions in RCC patients.