A Multiparametric Fusion Radiomics Signature Based on Contrast-Enhanced MRI for Predicting Early Recurrence of Hepatocellular Carcinoma
Table 2
The predictive performance of the clinico-radiological model, radiomics signatures using MR sequences and the predictive model.
Different models
Training dataset (n = 211)
Validation dataset (n = 91)
SENS%
SPEC%
ACC%
AUC (95% CI)
SENS%
SPEC%
ACC%
AUC (95% CI)
Clinico-radiological model
80.3
87.2
83.4
0.90 0.83–0.92
77.3
80.9
79.1
0.85 0.76–0.90
T2WI
87.8
65.6
78.0
0.83 0.78–0.89
69.8
67.4
68.5
0.74 0.64–0.85
DWI
94.8
61.3
79.9
0.81 0.75–0.88
88.6
53.2
70.3
0.75 0.65–0.85
Arterial phase signature
89.7
70.2
81.0
0.85 0.80–0.91
88.6
70.2
79.1
0.79 0.69–0.89
Portal venous phase signature
89.7
64.9
78.7
0.81 0.75–0.87
90.9
59.6
74.7
0.80 0.70–0.89
Fusion radiomics signature
89.7
69.1
80.6
0.85 0.80–0.91
90.9
70.2
80.2
0.79 0.68–0.89
Predictive model
91.5
78.7
85.8
0.91 0.87–0.95
88.6
74.5
81.3
0.87 0.79–0.94
A fusion radiomics signature was developed with arterial phase images and portal venous phase images. The predictive model consisted of a fusion radiomics signature and a clinico-radiological model. T2WI: T2-weighted imaging, DWI: diffusion-weighted imaging, SENS: sensitivity, SPEC: specificity, ACC: accuracy, and AUC: area under the curve.
