CECT-Based Radiomic Nomogram of Different Machine Learning Models for Differentiating Malignant and Benign...

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Differentiating benign from malignant solid renal masses is crucial for appropriate patient management.
• Accurate preoperative diagnosis can prevent unnecessary surgeries and guide timely treatment for renal tumors.

Purpose of the Study:

• To evaluate the diagnostic performance of a radiomic nomogram derived from contrast-enhanced computed tomography (CECT) for distinguishing benign and malignant solid renal masses.
• To develop and validate a noninvasive tool for renal mass characterization.

Main Methods:

• Radiomic features were extracted from CECT scans of 122 patients with pathologically confirmed renal masses.
• Machine learning models, including logistic regression, were trained and validated to build a radiomic signature.
• A clinical model was developed using significant clinical characteristics.
• A combined radiomic nomogram integrating clinical factors and radiomic signature was constructed.

Main Results:

• The logistic regression-based radiomic model achieved an area under the curve (AUC) of 0.952 (training) and 0.887 (test).
• The clinical model showed AUCs of 0.854 (training) and 0.747 (test).
• The combined radiomic nomogram demonstrated excellent discriminative performance with AUCs of 0.973 (training) and 0.900 (test).

Conclusions:

• The CECT-based radiomic nomogram is a promising noninvasive method for differentiating malignant from benign solid renal masses.
• This tool can assist clinicians in making informed decisions regarding treatment strategies for renal masses.
• The developed nomogram provides valuable insights for the accurate diagnosis of renal tumors.