CHAPTER 4 Assessment and Differential Diagnosis of Pathological Features

Initial examination and reporting

Naked-eye examination and description of biopsy specimens

Although naked-eye examination and description are of limited diagnostic value, they reduce the possibility of specimen identification error. The pathologist should make sure that the whole specimen has been adequately sectioned by comparing the size of the sectioned and stained tissue with the measurement recorded on macroscopic examination. Naked-eye examination also helps in the selection of suitable areas for electron microscopy. Needle biopsy specimens from cirrhotic liver are often irregular in calibre or obviously nodular, with brown or green nodules separated by grey or white fibrous tissue. The specimen may easily break into smaller pieces on handling or processing. Specimens with metastatic tumour or primary hepatocellular carcinoma also often fragment. Normal liver, on the other hand, gives rise to cylinders of even colour and thickness, which do not fragment easily. Cholestasis imparts a green colour, whereas fatty liver is pale brown or yellow and may float in the fixative. In cholesterol ester storage disease and Wolman’s disease the specimen is bright orange; this should warn the pathologist of the need to keep some tissue for frozen sectioning and electron microscopy. A black or very dark brown colour is characteristic of the Dubin–Johnson syndrome. Metastatic tumour, like fibrous tissue, is often white. Congested liver is deep red in colour. The clinician should record whether the specimen was difficult to obtain from the patient and whether the liver felt hard when the needle was inserted, as in many cases of cirrhosis, or very hard, as in congenital hepatic fibrosis.

Routine microscopy

Routine microscopy of liver biopsies should include systematic assessment of overall structure, portal tracts and their contents, terminal hepatic venules, hepatocytes and sinusoidal cells. Some pathologists use a pro forma or checklist in order to avoid omitting relevant data.¹

The following sections are intended to help in the evaluation of pathological changes. Most of the information is also found in other parts of the book, under individual diseases. There is inevitably some repetition, because many of the listed features are found in combination. The final part of the chapter contains guidance on the differential diagnosis of acute cholestasis and of individual pathological findings.

Basic patterns of injury

Structural changes, collapse and fibrosis

Minor structural changes are difficult to assess in sections stained with haematoxylin and eosin (H&E), and may indeed be missed altogether. Examination of a connective tissue preparation is therefore often important. Normal liver tissue shows a hierarchy of ramifying portal tracts of varied sizes which are present in needle and wedge biopsy samples (Fig. 4.1). The subdivisions of these portal tracts parallel the hierarchy of hepatic artery and portal vein branches and bile ducts as they distribute throughout the liver and can thereby be roughly subdivided into segmental, area, conducting (septal) and terminal portal tract units (see Fig. 5.1). For detection of the most minor abnormalities an uncounterstained silver impregnation for reticulin is generally best, although pericellular fibrosis is most easily detected in sections stained for collagen.

Figure 4.1 Portal tract size variations.

Biopsies contain portal tracts ranging in size from larger conducting tracts (at left) to the small terminal tracts (right top and bottom) from which blood enters the parenchyma. (Wedge biopsy, chromotrope–aniline blue.)

Using these methods, an impression may be gained that, although portal tracts and terminal venules are normally related to each other, the portal tracts are enlarged and perhaps even linked by fibrous septa. This is consistent with mild chronic viral hepatitis or with one of the conditions in which portal changes typically predominate; these include biliary tract disease, haemochromatosis, congenital hepatic fibrosis and schistosomiasis. If on the other hand the reticulin framework of the parenchyma is distorted, lesions characterised by lobular damage should be considered. These include acute and chronic hepatitis as well as forms of biliary disease in which there is also hepatocellular damage, notably primary biliary cirrhosis. Venous congestion leads to regular condensation of perivenular reticulin.

Recent collapse and fibrosis are sometimes difficult to distinguish, even with the help of good collagen stains. A stain for elastic tissue can help to resolve this problem because the presence of elastic fibres outside the portal tracts is an indication of long-standing disease. Collagen stains are helpful for the recognition of blocked veins, for example in necrotic areas, alcoholic liver disease, venous outflow obstruction and epithelioid haemangioendothelioma. Collagen staining is important for the detection of pericellular fibrosis, as already indicated, and should therefore be used whenever there is substantial steatosis or a suspicion of steatohepatitis.

The histological diagnosis of cirrhosis is fully discussed in Chapter 10. Once cirrhosis has developed, the pattern of fibrosis is one of the features that may help to determine its cause. In primary or secondary biliary cirrhosis, for example, fibrosis expanding and linking the portal tracts is a more important early factor in pathogenesis than hepatocellular regeneration; this is reflected in the morphological picture of broad perilobular septa surrounding irregularly shaped islands of parenchyma (see Fig. 5.11). In hereditary haemochromatosis and chronic venous outflow obstruction the impression is also of fibrosis rather than regeneration as the principal pathogenetic factor. In these diseases with a long pre-cirrhotic phase of fibrosis, transected parenchymal peninsulas may be mistaken for true regenerative nodules. This is particularly common just deep to the liver capsule. Isolated subcapsular nodules in an otherwise not nodular biopsy should therefore be interpreted with caution.

Hepatocellular damage

Death of individual hepatocytes or small groups of these cells is loosely called focal necrosis, although the mechanism may in fact be apoptosis. The distinction cannot always be made easily by routine microscopy unless apoptotic bodies are seen. Focal necrosis is associated with accumulation of inflammatory cells of various types, including macrophages. Spotty necrosis is a term used for the same lesion in the context of acute hepatitis. Focal necrosis is a common finding which does not in itself indicate primary disease of the liver, because it is often part of a non-specific reaction to disease elsewhere in the body. While degenerating hepatocytes or cell fragments are sometimes seen within the focal inflammatory infiltrate, the inflammatory reaction is usually more obvious than the necrosis, and the latter is assumed to have taken place because of a gap in a liver-cell plate (liver-cell ‘dropout’).

Confluent necrosis

Confluent necrosis (see Fig. 8.4) refers to substantial areas of liver-cell death. The commonest cause of this type of necrosis in biopsy material is hepatitis, either viral or drug-related, in which case the necrosis is accompanied by an inflammatory reaction. Confluent necrosis with little or no inflammation is seen in hypoperfusion of the hepatic parenchyma, as in shock or left ventricular failure, and in heatstroke (see Fig. 12.2). Paracetamol (acetaminophen) poisoning produces a similar lesion. In all the above examples the necrosis is typically perivenular but it may, if severe and extensive, form bridges linking vascular structures (see below). Some poisons, including ferrous sulphate, typically cause periportal (zone 1) necrosis. Haphazardly distributed areas of necrosis are found in disseminated herpes virus infections (e.g. herpes simplex, varicella) (see Fig. 15.4) and in mycobacterial diseases. Tumour necrosis may be so extensive that no recognisable tumour tissue is present in the section; in such cases the reticulin pattern may help to establish a diagnosis.

Bridging necrosis

Bridging necrosis describes the location rather than the type of necrosis. It usually results from extensive necrosis of confluent type. The term has been used for necrosis linking any of the vascular structures, but it is now more often restricted to the linking of terminal hepatic venules (centrilobular veins) to portal tracts (Fig. 4.2). A possible explanation for this type of bridging is that it represents necrosis of acinar zones 3, which touch both the veins and the larger portal tracts (as in Fig. 3.1). Linking of portal tracts to each other is common in conditions in which portal tracts are widened, for example by chronic hepatitis or biliary tract disease; this is partly because the chance of obtaining a longitudinal section of a widened portal tract is greater than for one of normal width. Linking of perivenular areas to each other by is found in some examples of parenchymal hypoperfusion and venous outflow obstruction.

Figure 4.2 Acute hepatitis with bridging necrosis.

Collapsed reticulin here gives a false impression of chronic liver disease. A bridge or passive septum (arrowheads) links an expanded portal tract (P) with a terminal hepatic venule (V). (Needle biopsy, reticulin.)

Bridging of terminal hepatic venules to portal tracts is a fairly common feature of acute hepatitis of viral type, when the bridges contain few or no elastic fibres. It is also seen in exacerbations of chronic hepatitis. Old bridges contain elastic fibres as well as collagen fibres. Such bridging fibrosis is an important component both of the more severe examples of chronic viral hepatitis and of steatohepatitis. Contraction of collagen-rich bridges may produce rapid and severe distortion of the normal hepatic microstructure, with correspondingly rapid progression to cirrhosis.

Panlobular and multilobular necrosis

Panlobular (panacinar) and multilobular (multiacinar) necrosis are terms used to describe confluent necrosis involving entire single lobules or several adjacent lobules respectively. They are further discussed in Chapter 6.

Interface hepatitis (piecemeal necrosis)

Interface hepatitis (piecemeal necrosis) (see Figs 9.3, 9.4) is a process of inflammation and erosion of the hepatic parenchyma at its junction with portal tracts or fibrous septa. The term ‘interface hepatitis’ was introduced because the death of hepatocytes probably involves apoptosis rather than, or as well as, necrosis,²^–⁵ and because it takes place at the parenchymal–connective tissue interface. It is common in chronic viral hepatitis but is also found in other conditions (see Box 9.2). The inflammatory infiltrate is composed mainly of lymphocytes, with or without recognisable plasma cells, and is accompanied by fibrosis of the affected areas with new formation of collagens and other extracellular matrix components.⁶ The process is sometimes referred to as classical or lymphocytic piecemeal necrosis in order to distinguish it from biliary, ductular and fibrotic piecemeal necrosis, processes found in chronic biliary tract disease and described in the section on primary biliary cirrhosis in Chapter 5.

Cholestasis

In morphological terms, cholestasis is the presence of visible bile in tissue sections. It is also known as bilirubinostasis because the main component seen by light microscopy is bilirubin. Bile is rarely seen in normal liver, and then only in minute amounts; cholestasis should therefore be regarded as pathological. The location of the bile varies. The commonest is in dilated bile canaliculi between hepatocytes. This canalicular form of cholestasis, sometimes called acute cholestasis, may be accompanied by bile accumulation in the cytoplasm of hepatocytes and Kupffer cells. Canalicular cholestasis is typically perivenular. In contrast, in patients with chronic biliary tract disease, bile may accumulate in periportal hepatocytes. This is also known as cholate stasis because abnormal bile salts are thought to contribute to its pathogenesis.

In large bile-duct obstruction in adults, bile is not usually visible under the microscope within canals of Hering, bile ductules or bile ducts even though the biliary tree may be dilated. The commonest cause of ductular cholestasis is sepsis. Dense bile is also visible in ducts in different forms of ductal plate malformation and in extrahepatic biliary atresia.

Canalicular cholestasis

Canalicular cholestasis takes the form of bile plugs (bile thrombi) in dilated canaliculi (see Fig. 5.2). There is often brown or yellow pigment in nearby hepatocytes and Kupffer cells, but the distinction of this pigment from others such as lipofuscin and ceroid is not a serious practical problem; this is because the presence of bile in the canaliculi makes the diagnosis of cholestasis obvious. In general, cholestasis should only be diagnosed with great caution in the absence of bile plugs in canaliculi, although cytoplasmic liver-cell bilirubinostasis without canalicular bile is quite common after liver transplantation. The perivenular location of canalicular cholestasis is partly an artefact of paraffin embedding, but also reflects real functional differences between the various parts of the acinus.

The colour of bile under the microscope varies according to pigment concentration and the degree of oxidation. It may be dark brown, green or yellow, and is occasionally so pale as to make detection difficult at first glance. The Van Gieson stain, which stains bilirubin green, may then be helpful (Fig. 4.3). Pale counterstaining, as commonly used in Perls’ and Prussian blue methods for iron, also makes bile easier to see. Specific histochemical methods for bilirubin are rarely necessary in ordinary diagnostic work.