Network Meta-Analysis: Noninvasive Imaging Modalities for Identifying Clinically Significant Portal Hypertension

Background Although measurement of the hepatic venous pressure gradient (HVPG) is the current reference standard for obtaining portal venous pressures, several noninvasive imaging-based modalities have been proposed as alternatives. Aims We performed a systematic review and meta-analysis to compare the diagnostic accuracy of noninvasive imaging approaches for identifying clinically significant portal hypertension (CSPH). Methods Two independent reviewers conducted a literature search of PubMed, SCOPUS, and the Cochrane Library from inception until January 5, 2021. The following imaging modalities were compared to HVPG: computed tomography (CT), magnetic resonance imaging (MRI), magnetic resonance elastography, ultrasound, transient elastography (TE), shear wave elastography (SWE), acoustic radiation force impulse (ARFI) imaging, contrast-enhanced ultrasound (CEUS), and subharmonic-aided pressure estimation (SHAPE). Sensitivity, specificity, diagnostic odds ratio (DOR), and area under the curve (AUC) for summary receiver operating characteristic were calculated using both frequentist random effects and Bayesian network meta-analytic approaches. Results We analyzed 45 studies of 5678 patients. A broad overlapping confidence interval (CI) of DOR was observed among different imaging modalities: ARFI (30.5; 95% CI 12.7–73.3), CEUS and SHAPE (21.1; 95% CI 6.4–69.8), TE of liver stiffness (21.1; 95% CI 13.3–33.5), CT and MRI (13.7; 95% CI 7.40–25.4), SWE of liver stiffness (10.5; 95% CI 5.2–21.1), and ultrasound (9.5; 95% CI 4.9–18.4). The AUC of all imaging methods exceeded 0.8, indicating very good performance. At a cutoff of 80% specificity, TE, CEUS, and SHAPE exceeded 80% sensitivity. Conclusion Overall, noninvasive imaging modalities perform well for identifying CSPH. Clinicians should consider these noninvasive and cost-efficient tests when diagnosing CSPH.