Introduction ARFI™ (Acoustic Radiation Force Impulse) elastography is a widely applicable technique for the non-invasive assessment of liver fibrosis, which has been well validated in viral hepatitis patients. As with transient elastography, the predictive value of this technique falls in intermediate stages of fibrosis, and shear velocity readings may also be affected by a number of other factors. However, few studies have systematically exam:ined the causes of the observed variance with liver histology. We report the results of a large unselected series, in which liver stiffness and histology have been sampled simultaneously from the same region of liver tissue.

Methods One hundred and eighty six unselected, consecutive secondary care referrals underwent simultaneous elastography and liver biopsy from the same right lobe liver window, both performed or supervised by a single senior radiologist in all cases. ARFI shear velocity measurements were made using a standard 10 observation technique, and biopsies taken using an 18G Biopince™ needle. All biopsies were reviewed by a single specialist histopathologist. Clinical, laboratory, elastographic and histological data were analysed retrospectively. ARFI/histological variance (AHV) was defined as a difference of more than 1 Metavir or 2 Ishak stages from that predicted by ARFI, according to standard calibration.¹

Results Aetiologies were 99 viral hepatitis, 39 autoimmune (AILD) and other in 48. AHV was seen in 56(30.1%), of which 46(82%) showed a lower histological stage than predicted. AHV was not associated with age, gender, or ARFI measurement depths. Inflammation (ALT, necroinflammation), steatosis (US echogenicity, histology), suboptimal ARFI quality (IQR/median > 0.3) and AILD aetiology were significantly more common in AHV (p = 0.01, 0.007, <0.001 and 0.018, respectively). Two or more of these variables were present in 61% of variants, compared with 26.9% of non-variants (p < 0.001).

Conclusion These simultaneous paired data show that ARFI/ histological variance is common and influenced by aetiology, inflammation, steatosis and technical quality. It is more common in active autoimmune liver disease than in viral hepatitis. Assuming that histology is a true “Gold standard”, taking these 4 factors into account when interpreting ARFI scores will assist in assessing the predictive reliability of elastography, and hence in clinical decision making with regard to liver biopsy and treatment. Further prospective studies with paired ARFI/histology sampling are warranted to confirm these findings.

Reference

1. Friedrich-Rust M, et al. Radiology 2009:252: 595–604

Disclosure of Interest None Declared.