Specimen Handling

At the time of the biopsy procedure, a needle liver biopsy specimen should be examined immediately for adequacy. It should be at least 1.5 cm long, and if it is not, another pass is recommended because adequate specimen size minimizes sampling error.^7,8 If a tumor is suspected, a touch preparation from the tissue can immediately determine specimen adequacy or a diagnosis. The specimen is then discharged into a Petri dish lined with lens paper, which has ideally been soaked in normal saline solution to prevent fresh tissue from adhering to it. Artifacts from squeezing or drying of the specimen should be avoided. Biopsy specimens should not be placed on dry gauze, which tends to dehydrate cells, resulting in a prominent nuclear artifact. Tissue squeezing distorts cells and elongates nuclei, which makes cytologic evaluation of the specimen difficult.

Gross characteristics of the tissue, such as color, consistency, and tendency to fragment or float in the fixative solution, are documented by the physician. Tumors or granulomas, for example, can be recognized as white areas in an otherwise reddish brown tissue. Gray-black discoloration is seen in Dubin-Johnson syndrome, rusty brown in hemochromatosis, green in cholestasis, yellow in fatty liver, dark red in congested liver, and variegated or dark brown in metastatic melanoma. Fragmentation of specimens, especially in a transjugular biopsy specimen or when a Menghini (suction) needle is used, often indicates advanced fibrosis or cirrhosis.

The routine fixative for liver biopsies is 10% neutral buffered formalin. Immersion of biopsies in saline causes discohesion of cells and distortion of the hepatocyte cords. The advantages of routine formalin fixation is that the formalin solution is stable, penetrates and fixes tissues well, is inexpensive, and allows subsequent application of most histochemical, immunohistochemical, and molecular pathology procedures. The characteristics of tissues fixed in formalin are well known. The disadvantage of formalin is the relative lack of cytologic detail compared with some other types of fixatives.

On the basis of the clinical diagnosis or possible differential diagnoses, the clinician determines which additional procedures may be required (Table 43.2): fixation in 3% buffered glutaraldehyde for electron microscopy; fresh, unfixed tissue for viral and mycobacterial cultures; rapid freezing in liquid nitrogen or a mixture of dry ice and isopentane for fat stains, certain immunohistochemical and enzyme activity studies,⁹ quantitative studies of hormone receptors, and isolation of genomic and viral DNA and RNA for molecular analyses¹⁰; and fixation in 1% periodic acid in 10% neutral buffered formalin at 4° C for 48 hours for evaluation of glycogen storage diseases.¹¹ When RNA recovery is needed, an alcohol-based fixative is better than a formalin fixative. When the patient is a child or a young adult, it may be beneficial to fix tissue in anticipation of possible ultrastructural studies.

Needle liver biopsy specimens should be placed immediately in the desired fixative, because the foundation of a good histologic preparation is rapid and complete fixation. Good preservation of tissue can be achieved by following standard guidelines. First, the volume of the fixative should be at least 15 to 20 times the volume of the tissue.¹² When transit to the laboratory is likely to involve much movement, the container should be filled to the brim with fixative. Second, sufficient time must be allowed for fixation to occur before processing is started. Because formalin penetrates most tissues at approximately 0.5 mm per hour at room temperature, approximately 4 hours are needed to penetrate a 2-mm-thick piece of tissue core. Fixation time may be shortened by application of heat, pressure, vacuum, agitation, or microwave techniques. Fragmentation of the central portion of the liver biopsy along the long axis is a sign of insufficient fixation time. Although prolonged formalin exposure (>24 hours) does not result in overfixation (i.e., hardening of tissue), it may reduce the availability of antigen sites for immunohistochemical studies. Tissues may afterward be stored in 10% (or lower) neutral buffered formalin or in 70% alcohol.¹³

Rush liver biopsy specimens should be manually processed to shorten the delay and meet the needs of critically ill or transplant patients; regular specimens are processed in an automated tissue processor. Table 43.3 compares the times for the manual and automatic methods. The application of microwave processing has significantly reduced processing time to as little as 15 minutes for a tiny biopsy specimen or to a 60- to 90-minute cycle for a larger biopsy specimen.¹⁴

Normal Microanatomy of the Liver

The functional unit of the liver is represented by the hepatic lobule of Kiernan or the hepatic acinus of Rappaport.¹⁵ A hepatic lobule consists of an efferent central vein with cords of hepatocytes radiating out toward multiple peripheral portal tracts. The changes in the lobule are classically described as being centrilobular, midzonal, or periportal.

The Rappaport acinus is a regular, three-dimensional structure in which blood flows from a central axis, formed by the terminal afferent portal venule and terminal hepatic arteriole in the portal tract, into the acinar sinusoids and then empties into several terminal hepatic venules at the periphery of the acinus. The acinus is subdivided into zones 1, 2, and 3, indicative of progressively decreasing tissue oxygenation (Fig. 43.1). The distance between two terminal hepatic venules represents the size of the acinus. The oxygen gradient, metabolic heterogeneity, and differential distribution of enzymes across the three zones of the acini help explain the zonal distribution of liver damage caused by hypoperfusion or ischemia and by certain toxic substances.

The hepatocyte is a polygonal epithelial cell with one or more centrally located, round nucleoli. The number of binucleate forms increases with age (Fig. 43.2). Some nuclei are larger than others, particularly in persons older than 60 years, indicating polyploidy.¹⁶ The significance of polyploidy is unknown. It is usually more prominent in the midzonal areas. Hepatocytes are arranged in one-cell-thick cords in adults, and they are separated by sinusoids, in which blood flows from the portal tracts to the terminal hepatic venules. In children as old as 5 or 6 years, the liver cells are arranged in two-cell-thick plates (Fig. 43.3). The presence of two-cell-thick plates and rosette formation in adults indicates hepatocyte regeneration (Fig. 43.4). Rare eosinophilic bodies or apoptotic bodies may indicate normal turnover of hepatocytes in otherwise apparently normal livers.

Glycogen accumulation in hepatocyte nuclei surrounding portal tracts produces a vacuolated appearance and is common in adolescents (Fig. 43.5). In adults, this appearance may be conspicuous in conditions such as glucose intolerance, Wilson disease, or diabetes mellitus. Cytoplasmic glycogen imparts a fine, reticulated, foamy appearance to the cytoplasm. The distribution of glycogen also shows diurnal and diet-related variations. An irregular distribution pattern may be found normally in biopsies and is not of diagnostic significance.

Lipofuscin pigment is normally seen in variable quantities in the centrilobular areas as periodic acid–Schiff (PAS)–positive, diastase-resistant, fine, light brown granules. There is a progressive increase of lipofuscin in individual hepatocytes with age (Fig. 43.6). The granules represent lysosomes that contain materials that cannot be further degraded; they are not present in recently regenerated hepatocytes. Bile is invisible under normal circumstances. Intracellular bile in cholestasis can be distinguished from lipofuscin by its greenish hue, lack of a granular appearance, and its frequent formation of bile thrombi in bile canaliculi in zone 3 hepatocytes (see Fig. 43.6). Large amounts of lipofuscin are difficult to distinguish from Dubin-Johnson pigment microscopically. Compared with lipofuscin, iron and copper are coarser, birefringent, and usually deposited in periportal hepatocytes. Hemosiderin and copper are abundant in the cytoplasm of hepatocytes during the first week of life, then gradually disappear, and should be absent before the age of 9 months. Small quantities of stainable iron are common in normal hepatocytes, particularly in older individuals.

Sinusoidal lining cells consist of specialized fenestrated endothelial cells and specialized macrophages or Kupffer cells, which are usually inconspicuous in normal biopsy specimens. Occasional lymphocytes or neutrophils may be present in the sinusoids. Between the sinusoidal lining cells and the hepatocytes lies the space of Disse, which contains plasma, scanty connective tissue, and perisinusoidal cells such as hepatic stellate cells (i.e., Ito cells or interstitial fat-storing cells) and pit cells (i.e., natural killer lymphocytes). Hepatic stellate cells are modified resting fibroblasts that can store fat and vitamin A and produce hepatocyte growth factor and collagen.^17,18 They play a significant role in hepatic fibrogenesis. When activated, hepatic stellate cells contain stainable desmin and actin in their cytoplasm that can be highlighted with an immunohistochemical stain for smooth muscle actin.¹⁹ Elastic fibers and basement membrane material are absent from normal sinusoids.^20,21

Smaller branches of the hepatic veins and the smallest efferent veins (or terminal hepatic venules) are in direct contact with the hepatocyte parenchyma. The terminal hepatic venules have very thin walls lined by endothelial cells. Thickening of the wall of terminal hepatic venules is often part of a pericellular fibrosis reaction and of central hyalin sclerosis in alcoholic liver disease.^22,23 It also may be seen focally in apparently normal individuals.

Most portal tracts contain a bile duct, several bile ductules, a hepatic artery branch, a portal vein branch, and lymphatic channels embedded in connective tissue. The size of the portal tracts is approximately three to four times the diameter of the hepatic artery branch. The amount of connective tissue and the size of intraportal structures depend on the size of the portal tracts. Portal tracts of different sizes may be seen in biopsy specimens. Pathologic processes do not necessarily affect large and small portal tracts to the same extent. Portal tracts normally contain a few lymphocytes, macrophages, and mast cells but do not contain neutrophils or plasma cells. The amount of inflammatory cells increases with age and typically varies from one portal tract to another.

The larger intrahepatic or septal bile ducts are lined by tall columnar epithelial cells, and the smaller or interlobular bile ducts are lined by cuboidal or low columnar epithelium. One or more interlobular bile ducts may be present in any particular portal tract. Bile ducts are always accompanied by a hepatic artery, which has approximately the same diameter as the bile duct. Larger bile ducts have more periductal fibrous tissue than smaller ones. Sizeable bile ducts are often seen in subcapsular liver parenchyma in liver biopsies. Bile ductules are located at the peripheral zone of the portal tracts, and they are smaller than interlobular bile ducts. They have a basement membrane and are lined by cuboidal cholangiocytes. Bile canaliculi are not readily recognized microscopically unless distended, as in parenchymal cholestasis with hepatocyte rosette formation. Bile canaliculi are connected to bile ductules through the canals of Hering. Bile ductules and canals of Hering have received more attention recently because they may represent the site of progenitor cells.²⁴

There are several changes in the liver related to aging, particularly in individuals older than 60 years (Fig. 43.7). There is more variation in the size of hepatocytes and the number of their nuclei (i.e., polyploidy) and an increase in lipofuscin pigment deposition. There may be apparent dilation of sinusoids because of hepatocyte cord atrophy. The portal tracts may contain denser collagen and may contain an increased quantity of mononuclear inflammatory cells. The hepatic arteries may have thickened walls, even in normotensive individuals. These changes are accompanied by alterations in the metabolic function of the liver, including the metabolism of various toxins and drugs.²⁵

Interpretation of Liver Biopsies

Histologic examination of liver biopsies should conform to a specific routine and include all tissue fragments and structures of the liver (i.e., architecture, portal triads, limiting plate, hepatocytes, sinusoidal cells, and terminal hepatic venules). A systematic approach ensures that important diagnostic findings are not overlooked. A sensible approach is to begin the examination with a low scan magnification to appreciate the lobular architecture, the presence and quantity of the various anatomic structures of the liver, and the presence or absence of normal structures or focal changes. At low magnification, the type and location of inflammation and steatosis can be well appreciated. This should be followed by a careful examination of the zone 3 acinus, where many changes, such as congestion, steatosis, necrosis, cholestasis, pigments, and endophlebitis, are often found. The remainder of the parenchyma and portal tracts may be examined individually. Specific histopathologic changes may be easily recognizable, but it is often the topographic and functional relationships of the structures of the liver that contribute to a clinically meaningful diagnosis.

The initial histologic examination of a liver biopsy specimen is often better conducted without knowledge of the clinical and laboratory information. After the morphologic changes are appreciated and a generalized pattern of injury has been ascertained, a differential diagnosis can be rendered in combination with the clinical and laboratory information. Clinical information and laboratory data should always be reviewed before submitting a final diagnosis, because more often than not, this information is essential in narrowing the differential diagnosis to a specific cause.

The liver tends to react similarly to a broad range of injuries, and in most instances, the needle biopsy is fairly representative. However, a sampling error, particularly in focal or irregularly distributed disease processes, must always be taken into consideration. For example, in primary sclerosing cholangitis, the characteristic bile duct injury and periductal concentric fibrosis may not be uniformly seen throughout the liver or may not be present at all. Cytomegalovirus inclusions may not be seen in every section of a liver biopsy, and it may be necessary to cut several serial sections before the typical microabscesses are identified. Biopsy of space-occupying lesions may not show neoplastic cells in every section. In cirrhosis, a small-caliber needle biopsy may not obtain septal fibrosis and may yield only fragments of parenchyma without fibrosis.

A study of needle biopsy specimens by Crawford and colleagues²⁶ showed that the number of portal tracts in a biopsy specimen was proportional to the total length of the specimen obtained and that not all of the portal tracts contained all portal triad structures. Portal triads may not contain at least one of the three key structures (usually the portal vein); 38% of portal tracts do not contain a portal vein, 7% do not have a bile duct, and 9% do not contain a hepatic artery. The most important finding of this study was that portal tracts almost always contained a paired bile duct and hepatic artery of approximately equal diameters. Fragmentation of the tissue into several smaller fragments and partial tears in the region of the terminal hepatic veins at the edges of the specimen are not uncommon by the time histologic sections are placed on a glass slide, even when the original piece of liver tissue obtained from the biopsy needle was not fragmented.

Innocent variations that should be considered to avoid erroneous interpretation are subcapsular liver parenchyma and surgery-associated changes. Needle biopsy of the immediate 2-mm subcapsular space often shows parenchyma that mimics cirrhosis (Fig. 43.8) and that may give a false impression of the status of the liver as a whole.²⁷ The liver capsule may be present in percutaneous liver biopsies at one end of the specimen or in the form of separate pieces of connective tissue. The capsule can be distinguished from most pathologic fibrous tissue by its density and maturity, and it often contains blood vessels and bile ducts. Similar artifacts can be found in biopsy samples when the needle enters the liver at an angle close to the capsule or in a wedge biopsy that contains capsular fibrosis tissue on two surfaces.

In liver biopsy specimens obtained at the end of a long surgical procedure, surgery-associated changes are usually seen as small, tight clusters of neutrophils within or under the hepatic capsule, sinusoids around central venules, portal tracts, and hepatic plates, and they resemble microabscesses (Fig. 43.9