Gadoxetic Acid-Enhanced MRI-Based Radiomic Models for Preoperative Risk Prediction and Prognostic Assessment of Proliferative Hepatocellular Carcinoma

Zuyi Yan ¹, Zixin Liu ¹, Guodong Zhu ²

⁰Nantong University, Nantong 226006, Jiangsu, China (Z.Y., Z.L., M.L.); Department of Radiology, Nantong Third People's Hospital, Nantong 226006, Jiangsu, China (Z.Y., Z.L., M.L., J.Z., M.L., J.J., T.Z., X.Z.); Department of Radiology, Affiliated Nantong Hospital 3 of Nantong University, Nantong 226006, Jiangsu, China (Z.Y., Z.L., M.L., J.Z., M.L., J.J., T.Z., X.Z.).

Academic Radiology +

|

August 24, 2024

View abstract on PubMed

Summary

This study developed radiomics models and a nomogram to predict and stratify risk for proliferative hepatocellular carcinoma (HCC). The delta radiomics model demonstrated the best performance in predicting patient survival outcomes.

Area Of Science

Radiology
Oncology
Medical Imaging

Background

Proliferative hepatocellular carcinoma (HCC) presents significant clinical challenges due to its high invasiveness and poor prognosis.
Accurate preoperative risk prediction and prognostic stratification are crucial for managing proliferative HCC.

Purpose Of The Study

To investigate the preoperative risk prediction and prognostic value of different radiomics models and a nomogram for proliferative HCC.
To assess the diagnostic efficacy and clinical utility of these predictive models.

Main Methods

Radiomics features were extracted from T1-weighted imaging (T1WI) across multiple phases (arterial, hepatobiliary).
Original and delta radiomics features were used to construct machine learning models.
A nomogram was developed integrating the optimal radiomics model score with independent clinical risk factors (BCLC stage).

Main Results

The developed nomogram demonstrated good predictive performance, with an AUC of 0.838 (training) and 0.801 (validation).
BCLC stage and the combined radiomics score were identified as independent predictors of proliferative HCC.
The delta radiomics (DR) model showed the best performance in independently predicting recurrence-free survival (RFS) and overall survival (OS), excelling in prognostic risk stratification.

Conclusions

The nomogram effectively predicts proliferative HCC.
Various radiomics models and the nomogram provide valuable prognostic stratification for patients with proliferative HCC.
The delta radiomics model shows particular promise for survival prediction in this patient group.

Keywords:

Hepatocellular carcinoma Machine learning Magnetic resonance imaging Radiomics