Table 3.
The diagnostic performance of EM, GM, LSM, APRI, and FIB-4 in evaluate liver fibrosis stages in training and test cohort
| Stage and method | AUC | p value * | p value ** | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | LR+ | LR− |
|---|---|---|---|---|---|---|---|---|---|
| Training cohort | |||||||||
| S4 | |||||||||
| APRI | 0.715 (0.663–0.767) | < .001 | < .001 | 65.9 | 66.3 | 56.8 | 75.0 | 1.9 | 0.5 |
| FIB-4 | 0.690 (0.636–0.744) | < .001 | < .001 | 69.1 | 61.2 | 55.8 | 72.4 | 1.9 | 0.6 |
| LSM | 0.926 (0.899–0.953) | < .001 | < .001 | 84.8 | 91.2 | 79.8 | 93.3 | 5.8 | 0.1 |
| GM | 0.994 (0.984–0.999) | / | < .001 | 95.1 | 95.6 | 96.3 | 95.8 | 21.3 | 0.0 |
| EM | 1.0 (0.992–1.0) | / | / | 100.0 | 100.0 | 100.0 | 100.0 | / | 0.0 |
| ≥ S3 | |||||||||
| APRI | 0.778 (0.730–0.827) | < .001 | < .001 | 71.3 | 75.8 | 73.9 | 72.8 | 2.6 | 0.3 |
| FIB-4 | 0.745 (0.695–0.795) | < .001 | < .001 | 69.0 | 66.3 | 67.9 | 65.5 | 1.9 | 0.5 |
| LSM | 0.906 (0.876–0.937) | < .001 | < .001 | 86.8 | 85.3 | 86.7 | 84.7 | 6.0 | 0.2 |
| GM | 0.992 (0.981–0.997) | / | < .001 | 96.9 | 95.0 | 97.1 | 95.0 | 19.0 | 0.0 |
| EM | 0.993 (0.983–0.998) | / | / | 97.4 | 95.8 | 98.1 | 97.0 | 23.4 | 0.0 |
| ≥ S2 | |||||||||
| APRI | 0.781 (0.73–0.832) | < .001 | < .001 | 75.2 | 73.3 | 84.8 | 58.2 | 2.8 | 0.4 |
| FIB-4 | 0.729 (0.673–0.785) | < .001 | < .001 | 62.8 | 73.3 | 79.7 | 54.4 | 2.0 | 0.4 |
| LSM | 0.906 (0.873–0.940) | < .001 | < .001 | 82.6 | 87.7 | 89.5 | 75.6 | 4.2 | 0.2 |
| GM | 0.992 (0.972–1.0) | / | < .001 | 95.7 | 94.4 | 97.5 | 95.0 | 22.3 | 0.0 |
| EM | 0.994 (0.981–0.998) | / | / | 98.8 | 96.0 | 98.1 | 97.0 | 25.8 | 0.0 |
| Test cohort | |||||||||
| S4 | |||||||||
| APRI | 0.716 (0.617–0.815) | < .001 | < .001 | 60.7 | 75.6 | 55.4 | 78.3 | 1.8 | 0.4 |
| FIB-4 | 0.698 (0.598–0.798) | < .001 | < .001 | 60.7 | 68.3 | 53.9 | 74.0 | 1.7 | 0.5 |
| LSM | 0.884 (0.821–0.947) | .003 | < .001 | 83.6 | 78.0 | 72.7 | 84.5 | 4.0 | 0.3 |
| GM | 0.897 (0.831–0.940) | .002 | < .001 | 86.0 | 83.6 | 87.8 | 81.3 | 5.2 | 0.2 |
| EM | 0.921 (0.897–0.951) | .01 | .005 | 89.1 | 87.0 | 92.8 | 81.3 | 6.9 | 0.1 |
| GM + LSM | 0.937 (0.907–0.970) | / | .013 | 89.0 | 92.5 | 90.2 | 87.3 | 12.1 | 0.1 |
| GM + EM | 0.950 (0.917–0.972) | / | / | 90.1 | 94.3 | 94.9 | 88.0 | 15.7 | 0.1 |
| ≥ S3 | |||||||||
| APRI | 0.741 (0.645–0.838) | .001 | .001 | 58.3 | 72.2 | 65.0 | 64.3 | 1.7 | 0.5 |
| FIB-4 | 0.721 (0.622–0.821) | < .001 | < .001 | 58.3 | 68.5 | 64.9 | 62.2 | 1.6 | 0.5 |
| LSM | 0.898 (0.839–0.956) | .004 | .001 | 83.3 | 74.1 | 80.3 | 74.5 | 3.9 | 0.3 |
| GM | 0.885 (0.827–0.919) | .003 | .004 | 87.7 | 84.3 | 92.1 | 76.8 | 5.6 | 0.2 |
| EM | 0.910 (0.853–0.952) | .022 | .004 | 85.2 | 90.0 | 93.1 | 73.3 | 8.1 | 0.2 |
| GM + LSM | 0.927 (0.893–0.958) | / | .016 | 87.8 | 85.8 | 90.5 | 78.2 | 8.2 | 0.1 |
| GM + EM | 0.932 (0.899–0.961) | / | / | 89.9 | 87.9 | 90.7 | 80.3 | 8.1 | 0.2 |
| ≥ S2 | |||||||||
| APRI | 0.796 (0.711–0.881) | .001 | < .001 | 63.6 | 73.9 | 81.0 | 53.9 | 2.0 | 0.4 |
| FIB-4 | 0.801 (0.711–0.890) | .001 | < .001 | 66.7 | 75.4 | 82.8 | 57.9 | 2.3 | 0.4 |
| LSM | 0.896 (0.834–0.957) | .004 | .001 | 75.8 | 85.5 | 87.0 | 72.7 | 3.2 | 0.2 |
| GM | 0.882 (0.825-0.918) | .003 | < .001 | 87.1 | 83.3 | 89.9 | 70.6 | 5.3 | 0.1 |
| EM | 0.907 (0.849–0.950) | .022 | .007 | 87.9 | 83.2 | 90.2 | 76.0 | 5.5 | 0.2 |
| GM + LSM | 0.920 (0.886–0.951) | / | .019 | 88.0 | 88.2 | 92.5 | 87.3 | 6.6 | 0.1 |
| GM + EM | 0.930 (0.899–0.962) | / | / | 90.0 | 87.8 | 94.2 | 77.6 | 7.2 | 0.1 |
Data in parentheses are 95% confidence intervals
*Compared with GM in training cohort and compared with GM + LSM in testing cohort
**Compared with EM in training cohort and compared with GM + EM in testing cohort
GM gray scale modality, EM elastogram modality, LSM liver stiffness measurement, GM + EM, gray scale modality and elastogram modality, GM + LSM gray scale modality and liver stiffness measurement, NPV negative predictive value, PPV positive predictive value, LR+ positive diagnostic likelihood ratio, LR− negative diagnostic likelihood ratio