Computed tomography-based radiomics nomogram model for predicting adherent perinephric fat.

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Adherent perinephric fat (APF) is a significant factor in renal carcinoma surgery.
• Accurate prediction of APF is crucial for surgical planning and patient outcomes.
• Current predictive methods may lack sufficient accuracy and specificity.

Purpose of the Study:

• To investigate the predictive value of a computed tomography (CT)-based radiomics nomogram model for APF.
• To develop and validate a model integrating radiomics features and clinical factors for APF prediction.
• To compare the diagnostic performance of the radiomics nomogram model against other predictive models.

Main Methods:

• Retrospective analysis of 220 renal carcinoma patients, divided into training (n=153) and validation (n=67) cohorts.
• Extraction of radiomics features from plain CT scans and generation of a radscore.
• Development of a radiomics nomogram model using multivariate logistic regression, incorporating selected radiomics features and clinical risk factors.
• Evaluation of model performance using Receiver Operating Characteristic (ROC) curves and Area Under the Curve (AUC) analysis.

Main Results:

• Thirteen radiomics features demonstrated good predictive effect for APF.
• The radscore model achieved an overall AUC of 0.966.
• The radiomics nomogram model, integrating radiomics and clinical factors, achieved higher AUC values (overall: 0.981; training: 0.997; validation: 0.949) compared to the clinical model.
• The radiomics nomogram model exhibited superior diagnostic accuracy and specificity.

Conclusions:

• The CT-based radiomics nomogram model demonstrates high prediction ability for APF.
• This model, incorporating radiomics features and clinical risk factors, holds significant clinical application value for APF prediction.
• The radiomics nomogram model offers improved diagnostic efficiency and accuracy over traditional clinical models.