Related Experiment Video
Updated: Aug 6, 2026
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
A Fusion Model of Time-dependent Diffusion MRI and Serum Biomarkers for Differentiating ICC from HCC: A Prospective
Xing-Qing Qin1, Yu-Chen Wei2, Yong-Mei Huang1
1Department of Radiology, The First Affiliated Hospital of Guangxi Medical University, Nanning, China (X.Q.Q., Y.C.W., Y.M.H., Y.Y.Y., M.N.L., J.Y.L.).
Rationale And Objectives:
To investigate the feasibility of time-dependent diffusion MRI (td-dMRI) derived microstructural parameters and serum markers versus gadoxetic acid-enhanced MRI (EOB-MRI) in the differential diagnosis of intrahepatic cholangiocarcinoma (ICC) and hepatocellular carcinoma (HCC), and to assess the value of non-enhanced MRI.
Materials And Methods:
78 patients (53 HCC, 25 ICC) undergoing preoperative EOB-MRI and td-dMRI were prospectively enrolled. Td-dMRI data were processed using the IMPULSED model to obtain microstructural parameters, including cell diameter (d), extracellular diffusion coefficient (Dex), intracellular volume fraction (Vin), and cellularity. The correlation between d and pathological cell diameter was assessed with postoperative immunohistochemistry. Clinical data and MRI features were analyzed, with logistic regression identifying independent ICC predictors. Receiver operating characteristic (ROC) curve, decision curve analysis (DCA), and calibration analysis were performed to evaluate diagnostic performance.
Results:
ICC patients exhibited elevated CA125, CA15-3, and CA19-9 levels but normal PIVKA-II levels. MRI revealed arterial phase hyperenhancement without clearance in the portal venous and transitional phases and rim arterial phase hyperenhancement (rim APHE). The ICC group exhibited significantly lower d, Vin, and cellularity, but higher Dex. Normal PIVKA-II, rim APHE, d, and Dex were independent ICC predictors. The clinic-EOB-MRI model (normal PIVKA-II + rim APHE), the td-dMRI model (d + Dex), and the fusion model (normal PIVKA-II + d + Dex) achieved AUCs of 0.75, 0.81, and 0.86, respectively.
Conclusion:
The fusion model integrating td-dMRI parameters with serum biomarkers represents a promising non-invasive imaging tool and may offer valuable insights for treatment decision-making.