Table 2.
Subgroup analysis for prospective cohort (our center vs. external centers).
| Variables | Levels | Total | Our center | External centers | Pvalue1 | |
|---|---|---|---|---|---|---|
| Overall | MVI− | 290 (58.5%) | 192 (50.3%) | 98 (86.0%) | <0.001 | |
| MVI+ | 206 (41.5%) | 190 (49.7%) | 16 (14.0%) | |||
| Tumor Size | 1.0–3.0 cm | MVI− | 79 (74.5%) | 56 (69.1%) | 23 (92.0%) | 0.022 |
| MVI+ | 27 (25.5%) | 25 (30.9%) | 2 (8.0%) | |||
| 3.0–5.0 cm | MVI− | 99 (68.8%) | 70 (61.4%) | 29 (96.7%) | <0.001 | |
| MVI+ | 45 (31.3%) | 44 (38.6%) | 1 (3.3%) | |||
| ≥5.0 cm | MVI− | 112 (45.5%) | 66 (35.3%) | 46 (78.0%) | <0.001 | |
| MVI+ | 134 (54.5%) | 121 (64.7%) | 13 (22.0%) | |||
| Tumor Number | 1 | MVI− | 253 (61.7%) | 160 (52.6%) | 93 (87.7%) | <0.001 |
| MVI+ | 157 (38.3%) | 144 (47.4%) | 13 (12.3%) | |||
| 2 | MVI− | 23 (45.1%) | 19 (42.2%) | 4 (66.7%) | 0.390 | |
| MVI+ | 28 (54.9%) | 26 (57.8%) | 2 (33.3%) | |||
| 3 | MVI− | 4 (40.0%) | 4 (40.0%) | 0 | Not Applicable | |
| MVI+ | 6 (60.0%) | 6 (60.0%) | 0 | |||
| >3 | MVI− | 10 (40.0%) | 9 (39.1%) | 1 (50.0%) | 1.000 | |
| MVI+ | 15 (60.0%) | 14 (60.9%) | 1 (50.0%) | |||
| BCLC Staging | 0A | MVI− | 257 (61.2%) | 164 (52.2%) | 93 (87.7%) | <0.001 |
| MVI+ | 163 (38.8%) | 150 (47.8%) | 13 (12.3%) | |||
| B | MVI− | 33 (43.4%) | 28 (41.2%) | 5 (62.5%) | 0.283 | |
| MVI+ | 43 (56.6%) | 40 (58.8%) | 3 (37.5%) | |||
| AFP Level | ≦20 | MVI− | 156 (69.6%) | 114 (63.7%) | 42 (93.3%) | <0.001 |
| MVI+ | 68 (30.4%) | 65 (36.3%) | 3 (6.7%) | |||
| 20–400 | MVI− | 69 (54.3%) | 44 (45.4%) | 25 (83.3%) | <0.001 | |
| MVI+ | 58 (45.7%) | 53 (54.6%) | 5 (16.7%) | |||
| ≧400 | MVI− | 65 (44.8%) | 34 (32.1%) | 31 (79.5%) | <0.001 | |
| MVI+ | 80 (55.2%) | 72 (67.9%) | 8 (20.5%) |
MVI, microvascular invasion; AFP, alpha fetoprotein. 1principles for selecting P values and statistics: (1) For continuous variables, if they met normal distribution, we used T-test results; otherwise, we used Wilcoxon results. (2) For categorical variables, we used chi-square test or Fisher exact probability method. 2. Data description method: (1) For continuous variables, if they satisfied normal distribution, we selected the mean (standard deviation); otherwise, we selected the median (interquartile range). (2) For categorical variables, they were described as N (%) under different categories.