CT-based radiomics for predicting pathological grade in hepatocellular carcinoma

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Hepatocellular carcinoma (HCC) grading is crucial for treatment planning.
• Accurate prediction of HCC grade pre-operatively remains a challenge.

Purpose of the Study:

• To develop and validate radiomics models for predicting hepatocellular carcinoma (HCC) pathological grade.
• To assess the diagnostic performance of these models using contrast-enhanced CT (CECT).

Main Methods:

• Radiomics models were constructed using tumor segmentation and feature extraction from CECT scans.
• Univariate analysis and least absolute shrinkage and selection operator (LASSO) regression were employed.
• Combined models incorporated clinical factors and radiomics scores, evaluated using ROC curves, calibration curves, and decision curve analysis (DCA).

Main Results:

• The arterial and portal venous phase (AP+VP) radiomics model demonstrated the best performance (AUC = 0.981 training, 0.842 validation).
• This model was used to build a predictive nomogram.
• Calibration curves and DCA confirmed satisfactory prediction efficiency.

Conclusions:

• CECT-based radiomics models can effectively differentiate between low- and high-grade HCC.
• These models offer good diagnostic performance for HCC pathological grading.