Introduction
Methods and materials
Study population
TACE procedure
Combination therapy with Lenvatinib
Anti‑viral therapy
Assessment of safety
Follow-up surveillance
MRI protocol
Image features
Radiomics feature extraction
Feature selection and classifier modeling
Statistical analysis
Results
Characteristics of study cohorts
Clinical characteristic | Patient (n = 61) |
---|---|
Age | |
≥ 65 | 25 |
< 65 | 36 |
Gender | |
Male | 55 |
Female | 6 |
Weight | |
≥ 60 kg | 23 |
< 60 kg | 38 |
BCLC | |
A | 3 |
B | 19 |
C | 39 |
Child–Pugh | |
A | 55 |
B | 6 |
Tumor numbers | |
1 | 12 |
≥ 2 | 39 |
Tumor size (cm) | |
HBV | |
Positivity | 57 |
Negative | 4 |
AFP (ng/ml) | |
≥ 400 | 32 |
< 400 | 29 |
GGT (U/L) | |
≥ 60 | 45 |
< 60 | 16 |
ALP (U/L) | |
≥ 125 | 26 |
< 125 | 35 |
AST (U/L) | |
≥ 40 | 46 |
< 40 | 15 |
ALT (U/L) | |
≥ 50 | 34 |
< 50 | 27 |
ALB (g/L) | 39.1 ± 5.2 |
TBIL (µmol/L) | 18.9 ± 15.0 |
PT (s) | 12.4 ± 1.2 |
NLR | 3.0 ± 1.9 |
LMR | 4.0 ± 2.1 |
PLR | 154.6 ± 76.7 |
PNI | 47.8 ± 7.3 |
SII | 671.1 ± 565.6 |
Category | Response evaluation using mRECIST | ORR (%) | DCR (%) | |||
---|---|---|---|---|---|---|
CR | PR | SD | PD | |||
ALL case (n = 61) | 14 | 35 | 7 | 5 | 80.3 | 91.8 |
BCLC stage A/B (n = 22) | 2 | 14 | 4 | 2 | 72.7 | 90.9 |
BCLC stage C (n = 39) | 12 | 21 | 3 | 3 | 84.6 | 92.3 |
Scanners | GE medical systems (3.0 T) | Siemens | UIH |
---|---|---|---|
N = 27 | N = 20 | N = 14 | |
T1 TE (ms) | 1.4–2.4 | 2.0 | 1.4 |
T1 TR (ms) | 3.9–4.9 | 4.4 | 3.7 |
T1 slice thickness (mm) | 5.0 | 5.0 | 5.0 |
T2 TE (ms) | 90 | 80 | 70 |
T2 TR (ms) | 2200–3000 | 2000 | 2400 |
T2 slice thickness (mm) | 7.0 | 6.0 | 7.0 |
FOV (mm) | 384 × 286 | 380 × 269 | 375 × 306 |
Predictive factors related to PFS
Characteristic | Univariate | Multivariate | ||||
---|---|---|---|---|---|---|
HR | 95%CI | P value | HR | 95%CI | P value | |
Age | 1.00 | (0.97–1.0) | 0.6 | |||
Sex | 4.2e−10 | (0-Inf) | 1 | |||
Tumor size | 1.00 | (0.93–1.1) | 0.59 | |||
Tumor number | 3.70 | (1.1–12.0) | 0.035* | 4.64 | 1.03–20.88 | 0.045 |
HBV | 0.33 | (0.1–0.99) | 0.048* | |||
AFP | 0.94 | (0.42–2.1) | 0.88 | |||
GGT | 1.10 | (0.42–2.7) | 0.89 | |||
ALP | 0.83 | (0.38–1.8) | 0.65 | |||
AST | 0.58 | (0.24–1.4) | 0.23 | |||
Child–Pugh | 1.00 | (0.23–4.3) | 0.99 | |||
BCLC | 1.40 | (0.63–3.0) | 0.42 | |||
PLR | 1.50 | (0.66–3.4) | 0.33 | |||
LMR | 1.80 | (0.76–4.2) | 0.18 | |||
PNI | 1.30 | (0.55–3.0) | 0.57 | |||
SII | 1.70 | (0.75–3.8) | 0.21 | |||
NLR | 0.62 | (0.21–1.8) | 0.39 | |||
Location | 1.70 | (0.80–3.8) | 0.16 | |||
Tumor margin | 1.00 | (0.44–2.4) | 0.93 | |||
Types of gross growth | 1.80 | (1.1–3.0) | 0.038* | 1.15 | 0.6–2.17 | 0.676 |
ADC intensity display | 1.00 | (0.11–2.4) | 0.93 | |||
Signal homogeneity on T2WI | 0.82 | (0.36–1.8) | 0.63 | |||
Intensity enhancement on arterial phase | 0.35 | (0.15–0.8) | 0.013* | 0.24 | 0.09–0.64 | 0.004 |
Expansion of arterial phase range | 0.35 | (0.15–0.8) | 0.81 | |||
Type of enhancement | 0.56 | (0.24–1.3) | 0.17 | |||
Arterial peritumoral enhancement | 0.86 | (0.35–2.1) | 0.73 | |||
Enhancing tumor capsule | 0.77 | (0.35–1.7) | 0.53 | |||
Hepatic capsule bulge | 0.54 | (0.24–1.2) | 0.14 | |||
Scarring centrally | 0.34 | (0.08–1.5) | 0.15 | |||
Hemorrhage | 0.53 | (0.07–4.2) | 0.55 | |||
Necrosis | 2.20 | (0.91–5.5) | 0.079 | |||
Fat | 0.22 | (0.03–1.8) | 0.15 | |||
Massive vein tumor thrombosis | 0.85 | (0.44–1.6) | 0.62 | |||
Metastasis in distant areas | 1.76 | (0.83–3.7) | 0.14 |
Predictive performance of nine models
Varibales and models | Sensitity | Specificity | Accuracy |
---|---|---|---|
Clinical model | 0.27 | 0.95 | 0.71 |
T1WI | 0.45 | 0.8 | 0.68 |
T1WI arterial phase | 0.01 | 0.95 | 0.61 |
T1WI portal venous phase | 0.64 | 0.85 | 0.77 |
T1WI delay phase | 0.27 | 1 | 0.74 |
T2WI | 0.55 | 0.9 | 0.77 |
ADC | 0.54 | 0.85 | 0.76 |
DWI(b = 800) | 0.72 | 0.45 | 0.55 |
Combined models | 0.99 | 0.95 | 0.71 |