Pathologic diagnosis of biliary atresia on liver biopsy: is tissue the issue?

Although a rare disease occurring in 1 : 5000 to 1 : 18 000 newborns, biliary atresia is the most common cause of pathologic infant jaundice, accounting for more than 30% of all neonates presenting with cholestasis,1 cirrhosis and end-stage liver disease; it is the most common indication for liver transplantation in children. Biliary atresia is reported across all countries and racial backgrounds but appears to be more prevalent in infants of Asian or African descent.2 This disease is characterized by a progressive course of fibroobliterative obstruction of the extrahepatic and intrahepatic bile ducts. It is difficult to identify during the perinatal period, thus even though recent data have suggested long-term survival without transplantation can be accomplished in more children if the Kasai procedure (hepatoportoenterostomy) is performed at a very early age (30–45 days old). Diagnosis during such an early period is not common, and therefore the diagnosis tends to be delayed without appropriate treatment. While the diagnosis of biliary atresia can be aided by ultrasonography, hepatobiliary scintigraphy, magnetic resonance cholangiopancreatography (MRCP), and endoscopic retrograde cholangiopancreatography (ERCP), liver biopsy is generally considered the gold standard.3 However, liver biopsy interpretation in biliary atresia still represents a diagnostic challenge, even for experienced pathologists. A recent paper published in JGH reports from India on histopathologic features and their diagnostic accuracy for biliary atresia by pre-laparotomy liver biopsy.4 In the Indian subcontinent, biliary atresia constitutes approximately 25–34% of all cases of neonatal cholestasis;5, 6 early diagnosis is crucial as timing of surgery correlates with outcome.7 Biliary atresia typically destroys the main extrahepatic bile ducts. Thus, the early histopathology of the intrahepatic biliary tree is characterized by features reflecting downstream biliary obstruction, such as bile ductular reaction with accompanying neutrophilic infiltrates in the portal tracts and bile plugs in the bile ductal lumens. These changes are usually seen in newborn infants younger than 80 days. As disease progresses, portal and periportal fibrosis gradually develops. Cholate stasis, a unique feature of cholestatic periportal hepatocytes, can be seen as hepatocellular swelling with cytoplasmic clumping at the portal-periportal junction. Accumulated intermediate filaments forming eosinophilic intrahepatocytic strands, otherwise known as Mallory hyaline, in the periportal hepatocytes is a common finding. In addition, abnormal copper deposition in the periportal hepatocytes can often be demonstrated by a special stain for copper or copper-associated protein, because copper is normally excreted into the bile canaliculi by periportal hepatocytes.1 Cumulative effects of persistent bile flow obstruction, recurrent ascending cholangitis, and progressive destruction of the intrahepatic biliary tree by the primary inflammatory process lead to paucity or even total loss of the interlobular bile ducts. Therefore biopsies obtained in the later course of the disease usually show less ductular reaction but more impressive ductopenia. The etiology of biliary atresia remains unknown, although a variety of theories have been proposed, including genetic, immune mediated mechanisms, infection and other environmental factors, as recently discussed in these columns.8 Genetic predisposition such as alpha-1-antitrypsin deficiency, mutation in the Jagged 1 gene, certain HLA haplotypes, etc. has been postulated; however, it is not certain how the disease is precipitated, given the proposed genetic background. Infectious agents, particularly viruses have been suggested to play a certain role in the pathogenesis of biliary atresia. It has been tantalizing to hypothesize that virus, especially cytomegalovirus, human herpes virus, reovirus, and rotavirus infected during the perinatal period may trigger biliary epithelial injury and subsequent chronic biliary obstruction via an immune-mediated process.3 However, most of the evidence suggesting this hypothesis was circumstantial and built on animal models that need further studies on human to support this theory. It is likely these factors all play a role in contributing to the development of this disease, but the pathogenesis may be complex and dynamic with sophisticated interplay between host and environmental factors. While the interpretation of liver biopsy plays a pivotal role in the diagnostic algorithm, accuracy of percutaneous liver biopsy in the diagnosis of extrahepatic biliary atresia ranges between 60 to 95% according to various studies.9 Such variation of diagnostic accuracy may in part be attributed to the experience of the pathologists, but may also partly depend on when the biopsy is taken during the course of the disease. In reality, not all needle liver biopsies show the typical obstructive findings, particularly in later stages as mentioned earlier. In addition, the interpretation can be complicated by equivocal histological changes and/or atypical clinical and imaging findings. In particular, biliary atresia can develop in the form of the so-called ‘early severe biliary atresia’,1, 10 which usually shows portal tracts with paucity of bile ducts and features reminiscent of ductal plate malformation in the percutaneous liver biopsy; this may cause diagnostic problems in pathology interpretation. A previous study has suggested the expression of CD56 in bile ducts and reacting ductules may help in the diagnosis of early biliary atresia when ductular reaction is present.11 CD56 is an isoform of N-CAM (neural cell adhesion molecule) that serves as an adhesion molecule for neural cells and natural killer cells. It is also a marker of immature cells committed to the biliary lineage and is expressed by reactive bile ductules in human liver diseases. Along the same line, Libbrecht et al. also found patches of immature biliary cells that express CD56/N-CAM in interlobular bile ducts with some cases of chronic liver disease, including a subset of biliary atresia. These findings indicate that progenitor cells could contribute to the repair of interlobular bile ducts, as interlobular bile duct cells normally do not express CD56.12 Overall, the results point to the direction that CD56 may be an ancillary diagnostic tool in equivocal cases, although further investigations with a large controlled cohort may be needed to validate the diagnostic utility of CD56 immunohistochemistry in biliary atresia. In their recent report, Rastogi and associates' article show that the histologic parameters most indicative of biliary atresia are ductular reaction, bile duct and ductular plugs, and portal fibrosis. In particular, ductular reaction was the most important histologic feature in distinguishing biliary atresia from neonatal hepatitis.4 Both the sensitivity and specificity for diagnosing biliary atresia by percutaneous liver biopsy in this study reach 88%, with an accuracy of 88%. This result is comparable to a previous study.13 Nonetheless, the accuracy of diagnosing biliary atresia by liver biopsy varies, depending on various studies reported.14 As the authors in the current study point out, ductular reaction should be interpreted with caution and should not be interpreted in isolation, because this histologic feature may also be seen in neonatal hepatitis. However, in the latter condition it seldom reaches the extent seen in biliary atresia, so over-interpreting mild to moderate ductular reaction in neonatal hepatitis may lead to the misdiagnosis of biliary atresia. Gene expression studies have shown a distinct signature in biliary atresia, with increased expression of proinflammatory genes and genes involved in morphogenesis, tissue remodeling, and fibrogenesis.15, 16 In the hope of developing novel biomarkers for early screening and diagnosis, and designing effective therapeutic targets, further investigative approaches including proteomic studies using clinical samples as well animal models are necessary to unveil the pathogenesis and etiology at the molecular, cellular, and phenotypic levels. Meanwhile, liver biopsy is the best we have, but it is not the final word on diagnosis as tissue manifestation of this disease changes with time.