A 10-Gene Signature Identified by Machine Learning for Predicting the Response to Transarterial Chemoembolization in Patients with Hepatocellular Carcinoma
Table 3
Cox regression for OS of patients receiving TACE treatment (n = 104).
Feature
Univariate analysis
Multivariate analysis
Hazard ratio
95% CI (HR)
value
Hazard ratio
95% CI (HR)
value
Gender (male VS female)
1.30
0.40–4.20
0.66
Age (>VS ≤ 50)
1.07
0.59–1.94
0.83
Predicted response (nonresponse VS response)
2.12
1.16–3.87
0.01
1.93
1.02–3.64
0.04
HBV (positive VS negative)
0.75
0.18–3.10
0.69
AFP (>VS ≤ 300 ng/ml)
1.35
0.74–2.45
0.33
ALT (>VS ≤ 50 U/L)
0.85
0.46–1.55
0.59
Cirrhosis (presence VS absence)
7.15
0.98–51.99
0.05
Main tumor size (>VS ≤ 5 cm)
1.88
1.03–3.42
0.04
1.40
0.67–2.93
0.38
Multinodular (presence VS absence)
1.17
0.58–2.38
0.66
BCLC stage (B and C VS 0 and A)
2.76
1.45–5.22
<0.005
1.98
0.93–4.24
0.08
HBV, hepatitis B virus; AFP, alpha fetoprotein; ALT, alanine aminotransferase; BCLC stage, Barcelona clinic liver cancer staging (there were no BCLC-D patients included in the research) .