Prediction of histologic grade of hepatocellular carcinoma using dual-layer spectral-detector computed tomography...

Area of Science:

• Radiology and Imaging
• Oncology
• Medical Physics

Background:

• Multi-parameter imaging aids tumor grading.
• Dual-layer spectral-detector computed tomography (DLCT) quantifies tumor properties.
• Region of interest (ROI) placement impacts DLCT parameter measurement and clinical diagnosis.

Purpose of the Study:

• Compare two ROI plotting methods on DLCT for differentiating hepatocellular carcinoma (HCC) histologic grades.
• Evaluate the impact of ROI selection on spectral parameter measurements for HCC.
• Determine the optimal method for predicting HCC tumor grade using DLCT.

Main Methods:

• Retrospective study of 48 HCC patients undergoing DLCT.
• Measured CT attenuation, electron density relative to water (EDW), normalized effective atomic number (NZeff), and normalized iodine density (NID) using planar sketching (PS) and volumetric analysis methods.
• Analyzed differences between arterial phase (AP) and venous phase (VP) parameters (∆CT, ∆EDW, ∆NZeff, ∆NID).
• Utilized t-tests, Mann-Whitney U test, Spearman correlation, and ROC curve analysis.

Main Results:

• Volumetric analysis yielded significantly lower mean spectral parameters (CT_AP, NZeff_AP, NID_AP) and AP-VP differences (∆CT, ∆EDW, ∆NZeff) compared to PS (P<0.05).
• Volumetric analysis achieved the highest AUC (0.918) for ∆NZeff in differentiating HCC grades, outperforming PS (AUC = 0.853).

Conclusions:

• DLCT spectral parameters offer a novel, clinically recommended method for evaluating HCC histological differentiation.
• ROI plotting methods significantly influence spectral parameter measurements.
• Comprehensive ROI selection covering the entire tumor is crucial for accurate HCC grading prediction.