TABLE 4.
Diagnostic performance of hepatocellular carcinoma with microvascular invasion in the training and testing cohorts.
| True positive | False positive | False negative | True negative | Sensitivity | Specificity | Positive-predictive value | Negative-predictive value | Accuracy | |
|---|---|---|---|---|---|---|---|---|---|
| Training cohort (n = 68) | |||||||||
| Radiomics model | 38 | 5 | 0 | 25 | 100 | 83.3 | 88.4 | 100 | 92.6% |
| SUVmax | 20 | 5 | 18 | 25 | 52.6 | 83.3 | 80 | 58.1 | 66.2 |
| ADC | 26 | 8 | 12 | 22 | 68.4 | 73.3 | 76.5 | 64.7 | 70.6 |
| Hypovascular | 31 | 11 | 7 | 19 | 81.6 | 63.3 | 73.8 | 73.1 | 73.5 |
| Non-smooth margin | 16 | 6 | 22 | 24 | 42.1 | 80 | 72.7 | 52.2 | 58.8 |
| Hybrid model | 36 | 0 | 2 | 30 | 94.7 | 100 | 100 | 93.8 | 97.1 |
| Testing cohorts (n = 29) | |||||||||
| Radiomics model | 16 | 3 | 3 | 7 | 84.2 | 70 | 84.2 | 70 | 79.3 |
| SUVmax | 14 | 3 | 5 | 7 | 73.7 | 70 | 82.4 | 58.3 | 72.4 |
| ADC | 13 | 6 | 2 | 8 | 86.7 | 57.1 | 68.4 | 80 | 72.4 |
| Hypovascular | 13 | 6 | 6 | 4 | 68.4 | 40 | 68.4 | 40 | 58.6 |
| Non-smooth margin | 12 | 5 | 7 | 5 | 63.2 | 50 | 70.6 | 41.7 | 58.6 |
| Hybrid model | 19 | 3 | 1 | 7 | 95 | 70 | 86.4 | 87.5 | 86.7 |