RESUMO
Early detection of liver graft fibrosis is crucial for risk stratification to identify patients for liver biopsy and timely treatment. However, diagnostic accuracy of noninvasive tests (NITs) remains unclear. Thus, this study sought to evaluate diagnostic accuracy of NITs in assessing liver allograft fibrosis and compare the differences in specificities and sensitivities among NITs. Medline and Embase databases were searched to include articles on diagnostic tests in liver transplantation (LT) patients with fibrosis. A meta-analysis on diagnostic test accuracy was conducted in a random-effects model. Sensitivities and specificities among the diagnostic tests were compared, and threshold values were calculated where applicable. A total of 25 articles were included. Vibration-controlled transient elastography (VCTE) met the minimum diagnostic accuracy requirements, yielding sensitivity, specificity, and diagnostic odds ratios of 0.9 (CI, 0.8-1.0), 0.9 (CI, 0.8-1.0), and 379.6 (CI, 45.8-1728.7), respectively. In the threshold assessment, the optimal cutoff was 9.30 kPa with a sensitivity, specificity, and area under the curve of 0.7 (CI, 0.5-0.9), 0.9 (CI, 0.8-0.9), and 0.9 (CI, 0.8-0.9), respectively. For significant fibrosis, acoustic radiation force impulse (ARFI) was superior to FibroTest (LabCorp [Burlington, NC]) and magnetic resonance elastography (MRE) in sensitivity. VCTE was superior to FibroTest in specificity. For advanced fibrosis, ARFI was superior to the Fibrosis-4 Index (FIB-4) in sensitivity. VCTE was superior to the AST to Platelet Ratio Index (APRI), MRE, and FIB-4 in specificity. In cirrhosis, VCTE was superior to APRI in specificity (P = 0.004) with comparable sensitivity. This study demonstrates the potential of VCTE and ARFI as diagnostic tools for fibrosis in LT recipients compared with blood-based NITs, which were shown to be less optimal.