Skip to main content
. 2022 Mar 1;32(7):4687–4698. doi: 10.1007/s00330-022-08585-x

Table 4.

Univariable and multivariable logistic regression analysis with respect to MRI morphologic tumor response at follow-up

Variables Univariable model Multivariable model
Demographics OR Sign. OR Sign.
Sex (male) 0.70 0.59
Age (years) > 70 0.91 0.85
Liver disease
BCLC tumor stage (C vs. B) 0.31 0.16
Hepatitis (yes/no) 1.93 0.37
Cirrhosis (yes/no) 1.03 0.96
Previous treatments
TKI (yes/no) 0.44 0.23
Resection (yes/no) 0.56 0.41
TACE (yes/ no) 2.19 0.35
Ablation (yes/no) 4.30 0.18
Radioembolization approach
Sequential lobar therapy (yes/no) 2.31 0.25 5.8 0.07
Dose reduction (yes/no) 0.73 0.68
Tumor characterization
Ln TTV (in mL) 0.52 0.03 0.4 0.01
Ln ETV (in mL) 0.73 0.23
ETV/TTV 4.39 0.15
ETV/TTV > 50% 3.41 0.07 6.3 0.02
Hepatopulmonary shunt in % 1.08 0.23 1.3 0.01
Performance of the multivariable model
Significance p = 0.001
Nagelkerke’s pseudo-R2 (range, competing models) 0.44 (0.30-0.44)

BCLC, Barcelona Clinic Liver Cancer stage; ETV, enhancing tumor volume; HR, hazard ratio; Ln, natural logarithm; ml, milliliter; Sign., significance; TACE, transarterial chemoembolization; TKI, tyrosine kinase inhibitor; TTV, total tumor volume