Performance Characteristics, Intra- and Inter-operator Agreement of Transient Elastography in Pediatric Nonalcoholic Fatty Liver Disease

Abstract

Background and aims: Transient elastography (TE) is a valuable tool in assessment of hepatic steatosis and fibrosis using liver stiffness measurement (LSM) and controlled attenuation parameter (CAP), respectively. Although widely used in adults, little is known about performance characteristics and reproducibility of TE (using Fibroscan device) in evaluation of pediatric nonalcoholic fatty liver disease (NAFLD).

Methods: We prospectively recruited children with NAFLD. Three consecutive Fibroscan examinations were performed during the same visit-twice by a single expert operator and once by a different novice operator. Intra and inter-operator agreement was calculated using concordance correlation coefficient (CCC). Failure was defined as inability to obtain 10 valid measurements and examination was considered unreliable if LSM interquartile range/median was greater 30%.

Results: Fifty-one children (34 boys; median age 15 years) were recruited. Failure rates for expert and novice operator were 10% (5/51) and 12% (6/51) while unreliable readings were obtained in 2% (1/46) and 4% (2/45) of patients, respectively. Patients with failed/unreliable measurements were significantly more obese (median BMI 46.2 vs 33.1 kg/m2, P = 0.002) compared with those with reliable measurements. The intra-operator agreement was almost perfect for LSM and substantial for CAP values (CCC = 0.85 and 0.73, respectively). Inter-operator agreement was substantial for LSM and moderate for CAP values (CCC = 0.76 and 0.58, respectively). The inter-operator agreement in LSM did not vary significantly over time but showed an inverse correlation with BMI and CAP.

Conclusion: Our study demonstrated that use of TE in assessment of hepatic fibrosis and steatosis in children with NAFLD is highly reliable with low failure rate and highly reproducible with high intra- and inter-operator reproducibility.