A Novel Multimodal Radiomics Model for Predicting Prognosis of Resected Hepatocellular Carcinoma

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Hepatocellular carcinoma (HCC) prognosis prediction is crucial for patient management.
• Current prediction models may not fully leverage multimodal imaging data.

Purpose of the Study:

• To develop and validate a novel model for predicting liver cancer prognosis using integrated MRI and CT radiomics data.
• To assess the model's performance in forecasting disease-free survival (DFS) and overall survival (OS).

Main Methods:

• A retrospective study of 103 HCC patients.
• Extraction of 1,217 radiomics features from CT and MR images.
• Development of multimodal radiomics scores and nomograms integrating imaging and clinical data.
• Validation of prediction models using a separate cohort.

Main Results:

• Multimodal radiomics scores independently predicted prognosis (DFS and OS) in HCC patients (p < 0.05).
• Integrated CT, MRI, and clinical data demonstrated high predictive performance (C-index for DFS: 0.858 training, 0.704 validation; OS: 0.893 training, 0.738 validation).
• Calibration curves confirmed the multimodal model's consistency and clinical utility.

Conclusions:

• Multimodal (MRI/CT) radiomics models are effective visual tools for predicting liver cancer prognosis.
• This approach shows significant potential for improving preoperative treatment decisions in HCC patients.