Algorithmic approaches to diagnosis have been widely used in medicine. Popularity of such approaches may have several explanations. First of all it reflects the logical pathway to diagnosis and at its core it is nothing more than formalized, practice tested deductive reasoning. Second, the algorithms are a useful practical method that can be readily taught to trainees and students; and for them it is easy to follow the outlines and schemes presented in a logical manner. Third, the logical sequences forming the basis of algorithms can be easily translated into computer languages and adapted to electronic instruction and/or keeping records. Fourth, algorithmic approach forces all of us to be mentally disciplined and systematically cover all facts, reducing thus the subjectivity of diagnosis. Finally, algorithmic data are easily presented in synoptic form.

Pathologists have been using algorithms in daily practice and synoptic reporting of liver biopsies has been widely accepted by hepatologists and pathologists alike (Ludwig 1992). Algorithmic approaches have been advocated by the leading hepatopathologists (Lefkowitch 2010; Theise and Saxena 2015; Torbenson 2015), who all promote an eclectic, team based, evidence tested and rather flexible approach. One should note that the algorithms included in such an approach may begin with a microscopic observation (e.g., predominantly portal inflammation, or steatosis), or a clinical-laboratory diagnosis (e.g., serologically confirmed viral hepatitis C). This applies especially to algorithms developed for diffuse liver disease. Tumors, cysts and other mass liver lesions are not included in these schemes and need to be diagnosed using an entirely different approach. Radiologic and or laparoscopic findings are critical for the diagnosis of lesions in this second group. Clinical-pathologic correlation is essential for proper diagnosis of both diffuse liver diseases and mass lesions.

Standardized grading and staging of several liver diseases, such as chronic viral hepatitis C, primary biliary cirrhosis, and steatohepatitis is also based on algorithms, which are constructed on the basis of detailed descriptions defining each grade and stage. These tabular systems generate numerical parameters which are very convenient for the follow up the diseases under observation or the evaluation of treatment. We will discuss them in the chapter devoted to specific liver disease.

In this chapter we will present our approach to stepwise diagnosis of liver diseases based on pattern recognition advocated by essentially all major hepatology texts. Schematically our approach is presented in Fig. 1. However, before we proceed, it is worth repeating some of crucial caveats, as follows:

Several histologic patterns can be recognized in the liver biopsy and accordingly the biopsy findings could be classified into several major categories as follows:

This group of changes includes several morphologic alterations none of which is diagnostic of a specific liver disease, individually or in aggregate (Burt et al. 2012). Such changes could involve hepatocytes, bile ducts, Kupffer cells, or hepatic vasculature. Torbenson (2015) lists them under the heading of “almost normal liver biopsy”. The biopsy findings are usually nonspecific and rarely if ever diagnostic (Table 1).

The most common “abnormality” is mild focal infiltration of some portal tracts with lymphocytes (Fig. 2). Since the normal portal tracts contains only a few lymphocytes, the assessment of portal lymphocytosis is subjective and essentially “in the eyes of the beholder”. We usually review all the portal tracts in the biopsy specimen and if only one or two out of 20 are infiltrated with lymphocytes, we tend to ignore that finding, especially if there are no clinical correlates to suggest chronic hepatitis. It is worth mentioning that infiltrates of portal lymphocytes occur quite often in the liver parenchyma around some other lesions, especially primary or metastatic carcinomas. Scattered lymphocytes or neutrophils may be seen inside the lobular parenchyma as well.

∎ Hepatocytes may show foci or single cell hydropic change or prominent anisonucleosis (variation in size of nuclei) outside of the normal range (Fig. 3). Glycogen may accumulate in the nuclei of hepatocytes accounting for the clearing up of nuclei, which are then referred to as glycogenated nuclei (Fig. 4). In the cytoplasm of scattered hepatocytes glycogen can form of pseudogroundglass inclusions, which impart the cytoplasm a finely granular appearance (Fig. 5). This change may be caused by a variety of stimuli, such as drugs, hormones, diabetes and even chronic passive congestion. Fatty change, which usually varies from one lobule to another may present in form of small or large droplets, i.e., as microvesicular or macrovesicular steatosis (Fig. 6). Brown cytoplasmic granules, especially in elderly or chronically ill patients usually represent lipofuscin. It must be distinguished from hemosiderin, which may be demonstrated with the Perl iron stain.

∎ Bile ducts may be increased in number in portal tracts (Fig. 7a). Such portal tracts may be infiltrated by neutrophils or lymphocytes. The increased number of bile ductules (cholangioles) may be associated with periductular fibrosis. It should be distinguished from ductular proliferation combined with acute pericholangitis and mild portal tract edema which is usually associated with mechanical biliary obstruction. Ductal and ductular proliferation and ductular reaction of biliary obstruction are best seen in slides immunohistochemically stained with antibodies to cytokeratin CK7 (Fig. 7b).

∎ Inflammatory cells may be seen in the portal tracts or lobules of patients who have some systemic disease, but also in some that have no signs of inflammation. Most often one deals with lymphocytes (Fig. 8). An increased number of Kupffer cells cannot be distinguished from an infiltration of bone marrow derived macrophages. In surgically removed tissue or in needle biopsies performed during surgical operation the liver parenchyma may contain aggregates of neutrophils (Fig. 9). This surgical acute hepatitis is related to mechanical handling of the liver and should not be confused with changes related to bacterial infections.

∎ Kupffer cells may be increased in number and prominent due their abundant cytoplasm. The cytoplasm of these cells may contain granules of brown pigment (Fig. 10a). This glycolipid-rich pigment, called ceroid, stains with PAS and is resistant to digestion with diastase (Fig. 10b). It is usually seen as evidence of a past liver cell injury (especially in drug induced liver injury), in which the remnants of liver cell cytoplasm were taken up by phagocytosis into the cytoplasm of resident macrophages. Such findings are sometimes referred to as resolving hepatitis pattern (Torbenson 2015).

∎ Sinusoids may show focal dilatation and congestion (Fig. 11). Furthermore these vascular spaces may contain a number of cells that are important for diagnosis. The pathology of hepatic sinusoids has been reviewed recently by Brunt et al. (2014).

The significance and the extent of circulatory changes cannot be always established from the limited sample obtained by needle biopsy. It is worth remembering that certain drugs, such as oral contraceptives and steroid hormones may cause sinusoidal dilatation and congestion. Additional clinical information provided such as echocardiogram showing congestive heart failure or hepatic vein outflow obstruction on ultrasound or cross sectional imaging would be relevant. Inside the sinusoids one may see reactive lymphocytes in various systemic viral diseases, such as infectious mononucleosis. In various leukemias sinusoids contain neoplastic white blood cells. In myelofibrosis the sinusoids may contain hematopoietic cells, indicative of extramedullary hematopoiesis.

∎ Laboratory findings indicative of a liver disease may be encountered in many patients with systemic infection or sepsis. Chronic pancreatic or biliary tract diseases, such as cholelithiasis, chronic cholecystitis or pancreatitis or cystic fibrosis may cause hepatic changes. Gastrointestinal diseases of unknown etiology, such as ulcerative colitis or Crohn’s disease, and some systemic autoimmune diseases such as systemic lupus erythematosus, and blood disorders such as sickle cell anemia may also affect the liver. Preoperative biopsies in candidates for heart or kidney transplantation may also show minor nonspecific changes. Table 2 contains a list of representative diseases that may cause non diagnostic minor reactive changes in the liver.

Normal liver tissue does not contain any visible bile, and accordingly any bile that can be seen in the liver biopsy should be considered as pathological. The most common causes of cholestasis are listed in Table 3.

In acute cholestasis, as encountered following short-lived bile duct obstruction, bile accumulates in the form of dot-like or short linear bile plugs in the dilated intercellular bile canaliculi. Retrograde pressure resulting from acute bile duct obstruction will be transmitted to the centrilobular area, and thus the bile plugs composed of yellow/brown bilirubin will first be seen in the centrilobular zone 3. Later on bile plugs may be seen in the entire lobule (Fig. 12). Bile may be seen inside the cytoplasm of liver cells and Kupffer cells. In chronic cholestasis bile plugs will be seen in the periportal zone 1. In addition to bilirubin these biliary plugs contain biliary salts (cholates) which may cause liver cell injury. Cholate stasis is thus associated with feathery degeneration of liver cells and regeneration of liver cells which form so called cholestatic hepatocellular rosettes around the centrally located bile filled canaliculi (Fig. 13). Destruction of liver cells accompanied by accumulation of bile in the extracellular spaces results in the formation of “bile lakes” (Fig. 14). In long term biliary obstruction bile may be seen in the lumen of biliary ducts, but this happens rarely. The most common cause of ductal cholestasis is sepsis, which may also cause intralobular cholestasis and can be seen as a sign of biliary tract infection. Bile may be seen in the lumen of distorted ducts in polycystic liver disease or parts of bile duct hamartoma included accidentally in the liver biopsy (Fig. 15). Ductal cholestasis is also seen in various congenital abnormalities of the biliary system in children.

Liver biopsies showing signs of cholestasis should be evaluated stepwise as indicated in Box 1 and also illustrated in Fig. 16.