Table 2.
Radiological characteristics of the study population according to pathological microvascular invasion in the training set and validation set
| Variables | Training set (n = 93) | Validation set (n = 40) | p-valueb | ||||
|---|---|---|---|---|---|---|---|
| MP (n = 31) | MN (n = 62) | p-valuea | MP (n = 14) | MN (n = 26) | p-valuea | ||
| LI-RADS major features | |||||||
| Tumor size (cm) | … | … | … | ||||
| < 2.0 | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |||
| ≥ 2.0 | 31 (100) | 62 (100) | 14 (100) | 26 (100) | |||
| Non-rim arterial phase hyperenhancement | 0.598 | 0.533 | 1.000 | ||||
| Absent | 2 (6.5) | 2 (3.2) | 0 (0) | 2 (7.7) | |||
| Present | 29 (93.5) | 60 (96.8) | 14 (100) | 24 (92.3) | |||
| Non-peripheral washout | 0.716 | 0.694 | 0.788 | ||||
| Absent | 7 (22.6) | 12 (19.4) | 4 (28.6) | 5 (19.2) | |||
| Present | 24 (77.4) | 50 (80.6) | 10 (71.4) | 21 (80.8) | |||
| Enhancing capsule | 0.537 | 0.864 | 0.732 | ||||
| Absent | 12 (38.7) | 20 (32.3) | 5 (35.7) | 10 (38.5) | |||
| Present | 19 (61.3) | 42 (67.7) | 9 (64.3) | 16 (61.5) | |||
| LI-RADS ancillary features (favoring HCC in particular) | |||||||
| Non-enhancing capsule | 0.906 | 0.646 | 0.850 | ||||
| Absent | 27 (87.1) | 56 (90.3) | 11 (78.6) | 23 (88.5) | |||
| Present | 4 (12.9) | 6 (9.7) | 3 (21.4) | 3 (11.5) | |||
| Nodule-in-nodule architecture | 0.397 | 0.602 | 0.750 | ||||
| Absent | 28 (90.3) | 59 (95.2) | 12 (85.7) | 24 (92.3) | |||
| Present | 3 (9.7) | 3 (4.8) | 2 (14.3) | 2 (7.7) | |||
| Mosaic architecture | 0.007 | 0.021 | 0.889 | ||||
| Absent | 13 (41.9) | 44 (71.0) | 5 (35.7) | 19 (73.1) | |||
| Present | 18 (58.1) | 18 (29.0) | 9 (64.3) | 7 (26.9) | |||
| Fat in mass, more than adjacent liver | 0.734 | 0.795 | 0.737 | ||||
| Absent | 24 (77.4) | 46 (74.2) | 11 (78.6) | 18 (69.2) | |||
| Present | 7 (22.6) | 16 (25.8) | 3 (21.4) | 8 (30.8) | |||
| Blood products in mass | 0.730 | 0.828 | 0.442 | ||||
| Absent | 23 (74.2) | 48 (77.4) | 9 (64.3) | 19 (73.1) | |||
| Present | 8 (25.8) | 14 (22.6) | 5 (35.7) | 7 (26.9) | |||
| LI-RADS ancillary features (favoring malignancy, not HCC in particular) | |||||||
| Restricted diffusion | … | … | … | ||||
| Absent | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |||
| Present | 31 (100) | 62 (100) | 14 (100) | 26 (100) | |||
| Mild-moderate T2 hyperintensity | … | … | … | ||||
| Absent | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |||
| Present | 31 (100) | 62 (100) | 14 (100) | 26 (100) | |||
| Corona enhancement | 0.001 | 0.012 | 0.578 | ||||
| Absent | 15 (48.4) | 50 (80.6) | 5 (35.7) | 21 (80.8) | |||
| Present | 16 (51.6) | 12 (19.4) | 9 (64.3) | 5 (19.2) | |||
| Fat sparing in the solid mass | 1.000 | 1.000 | 0.895 | ||||
| Absent | 29 (93.5) | 57 (91.9) | 13 (92.9) | 23 (88.5) | |||
| Present | 2 (6.5) | 5 (8.1) | 1 (7.1) | 3 (11.5) | |||
| Iron sparing in solid mass | 0.598 | 0.602 | 0.384 | ||||
| Absent | 29 (93.5) | 60 (96.8) | 12 (85.7) | 24 (92.3) | |||
| Present | 2 (6.5) | 2 (3.2) | 2 (14.3) | 2 (7.7) | |||
| Non-LI-RADS imaging features | |||||||
| Cirrhosis | 0.660 | 0.641 | 0.720 | ||||
| Absent | 15 (48.4) | 33 (53.2) | 7 (50.0) | 15 (57.7) | |||
| Present | 16 (51.6) | 29 (46.8) | 7 (50.0) | 11 (42.3) | |||
| Tumor margin | 0.015 | 0.037 | 0.572 | ||||
| Smooth | 14 (45.2) | 44 (71.0) | 6 (42.9) | 21 (80.8) | |||
| Non-smooth | 17 (54.8) | 18 (29.0) | 8 (57.1) | 5 (19.2) | |||
| Tumor capsule | 0.024 | 0.026 | 0.895 | ||||
| Complete | 14 (45.2) | 43 (69.4) | 5 (35.7) | 20 (76.9) | |||
| Incomplete/absent | 17 (54.8) | 19 (30.6) | 9 (64.3) | 6 (23.1) | |||
| TTPVI | 0.019 | 0.049 | 0.586 | ||||
| Absent | 16 (51.6) | 47 (75.8) | 7 (50.0) | 22 (84.6) | |||
| Present | 15 (48.4) | 15 (24.2) | 7 (50.0) | 4 (15.4) | |||
Data are expressed as numbers with percentages in parentheses. Statistically significant results (p < 0.05) are marked in bold. Ellipses indicated all observations (133/133) showed the feature of greater than 2.0 cm, restricted diffusion, and mild-moderate T2 hyperintensity
HCC hepatocellular carcinoma, LI-RADS Liver Imaging Reporting, and Data System, MN microvascular invasion negative, MP microvascular invasion positive, TTPVI a two-trait predictor of venous invasion
a Comparison between the MP group and the MN group
b Comparison between the training set and validation set