Multimodality 3D Tumor Segmentation in HCC Patients Treated with TACE

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Hepatocellular carcinoma (HCC) is a primary liver malignancy often treated with transarterial chemoembolization (TACE).
• Accurate tumor volume assessment is crucial for treatment planning and monitoring response to TACE.
• Multimodality imaging plays a key role, but concordance in volumetric measurements needs validation.

Purpose of the Study:

• To validate the volumetric concordance of a semiautomated 3D segmentation technique for HCC lesions.
• To compare measurements obtained from contrast-enhanced magnetic resonance imaging (CE-MRI), cone-beam computed tomography (CBCT), and multidetector CT (MDCT).
• To assess the reliability of this technique in patients with HCC undergoing TACE.

Main Methods:

• Retrospective analysis of 31 patients with unresectable HCC (40 target lesions).
• Inclusion of baseline CE-MRI, intraprocedural CBCT, and post-TACE MDCT scans.
• Application of a semiautomated 3D segmentation software based on non-Euclidean geometry and radial basis functions to calculate tumor volumes.

Main Results:

• Mean tumor volumes were comparable across CE-MRI, CBCT, and MDCT (18.72 ± 19.13 cm³, 21.26 ± 21.99 cm³, 19.88 ± 20.88 cm³, respectively).
• No significant differences in average tumor volumes were found between the imaging modalities (P > .577).
• Strong correlations were observed between volumetric measurements from all modality pairs (R² = 0.974–0.992), with deviations <10%.

Conclusions:

• The semiautomated 3D segmentation technique demonstrates high volumetric concordance for HCC lesions across CE-MRI, CBCT, and MDCT.
• This validated technique can reliably assess tumor volumes in HCC patients undergoing TACE.
• The findings support the use of this method for consistent tumor burden evaluation in HCC management.