Gray-Scale Ultrasound Terminology and Artifacts

Lesions that appear dark relative to the background tissue are termed hypoechoic; those that are brighter than background are hyperechoic or echogenic (Fig. 13.1). Spaces that contain fluid do not contain internal interfaces, so no sound is reflected from within them; they appear uniformly black or anechoic. The ultrasound scanner normally compensates for the attenuation of sound that occurs with increasing depth within the tissue by amplifying echoes that return later from the far field. This is called time-gain compensation (TGC) or depth-gain compensation. TGC results in needless amplification of echoes from areas deep to pure fluid structures, because the acoustic pulse is attenuated to a lesser extent by fluid than by tissue. Such artificially bright areas beyond the fluid represent increased through transmission, or acoustic enhancement (Fig. 13.2). Although technically an artifact, acoustic enhancement is useful to characterize structures as cystic if they also lack internal flow on color Doppler.

FIGURE 13.1 Transverse gray-scale ultrasound image of the liver shows the right hepatic vein (hv) as a black (anechoic) fluid-containing structure. Note that the walls of the veins are bright (echogenic). The echogenic curved line posteriorly represents the diaphragm.

FIGURE 13.2 Simple hepatic cyst has no internal echoes and a thin smooth wall. Note the bright posterior acoustic enhancement (arrows) that is a characteristic feature of a cystic lesion.

Several other ultrasound artifacts also are used for diagnosis. At boundaries with large acoustic impedance mismatches, such as between tissue and air or tissue and bone, nearly all the energy in the incident beam is reflected; because none is transmitted, areas beyond the interface will not be visualized. A black acoustic shadow is thereby produced behind the object in the image (Fig. 13.3). Shadowing is characteristic of virtually all gallstones and greatly aids in distinction of gallstones from gallbladder polyps.

FIGURE 13.3 Gallstones (curved arrow) layering in the gallbladder produce an acoustic shadow (straight arrows).

Within fluid-filled structures such as the gallbladder, faint echoes are often seen paralleling the anterior surface. These echoes represent reverberation artifact, which occurs as the sound is repeatedly reflected between the highly echogenic anterior gallbladder wall and the transducer. Reverberation also accounts for comet tail artifact, a streak of tiny, bright, parallel bands tapering behind a punctate crystalline reflector, generated by sound echoing within the crystal itself. Comet tail artifact is a feature of adenomyomatosis of the gallbladder and emanates from cholesterol crystals trapped in dilated Rokitansky-Aschoff sinuses in the gallbladder wall (Fig. 13.4). Fluid-filled structures may also have artifactual echoes along the dependent wall that originate from off-axis portions of the ultrasound beam. This is termed side lobe artifact and may be distinguished from true debris by changing the angle of insonation or repositioning the patient (Feldman et al, 2009; Zagzebski, 1996).

FIGURE 13.4 Bright reflectors in the gallbladder wall (arrows) produce comet tail, or ring-down, artifacts secondary to sound reverberation. This appearance is characteristic of gallbladder adenomyomatosis.

Gray-Scale Ultrasound of Focal Liver Masses

Conspicuity of liver lesions by gray-scale ultrasound is determined by lesion echogenicity relative to the liver. Small, markedly hypoechoic or hyperechoic lesions may be readily detected and characterized by ultrasound (Eberhardt et al, 2003), but masses that have echogenicity similar to adjacent liver may be more difficult to appreciate. On gray-scale ultrasound, liver masses are differentiated by internal architecture and are described as cystic, hypoechoic, and hyperechoic relative to the liver. Table 13.1 shows the common differential diagnoses within each of these categories (Charboneau, 2002; Middleton, 2002). The majority of cystic masses are benign, whereas most hypoechoic masses, with the exception of focal fatty sparing, are suspicious for neoplasm and require further evaluation, as they are clinically significant. Focal sparing often has geographic margins, typically is located in segment IV anterior to the portal bifurcation or adjacent to the gallbladder, and vessels are undisturbed. These distinctive features allow confident diagnosis (Fig. 13.9).

FIGURE 13.9 Focal sparing in an otherwise fatty liver. Transverse image shows a hypoechoic region (arrow) in segment IV anterior to the porta. pv, portal vein; ivc, inferior vena cava.

Hyperechoic masses include benign hemangioma, focal fat, and hepatic neoplasms such as hepatocellular carcinoma (HCC), adenoma, and metastases. Hemangiomas are brightly echogenic in the majority of cases and are sharply circumscribed. This appearance differs from malignant hyperechoic masses, which often have a hypoechoic halo. Any liver mass with a hypoechoic halo is suspicious for malignancy, and this sign has 86% positive predictive value and 89% negative predictive value for diagnosis of malignancy (Wernecke et al, 1992b).

Ultrasound Microbubble Contrast Agents for Evaluation of Liver Masses

Ultrasound microbubble contrast agents improve both detection and characterization of focal liver masses (Albrecht et al, 2003, 2004; Blomley et al, 1999; Bryant et al, 2004; Harvey et al, 2000; Quaia et al, 2004). First-generation contrast agents, such as SHU 508A (Levovist; Schering, Berlin), remain within the Kupffer cells after the vascular phase; when high mechanical index is applied to burst the bubbles, the liver enhances brightly, but metastases and other lesions without Kupffer cells appear dark. A multicenter study showed improved detection of individual metastases from 71% to 87%, and in seven of 80 patients, contrast-enhanced ultrasound showed more metastases than CT (Albrecht et al, 2003).

More recently, second-generation contrast agents and low mechanical index (LMI) scanning techniques have become available. At LMI settings, the microbubbles emit acoustic signals as they expand and contract in response to the ultrasound pulse, but they do not burst. Because the microbubbles stay in circulation, it is possible to image throughout arterial and portal phases. With LMI scanning, hypervascular tumors that enhance in the arterial phase are more easily detected. Lesion detection is also improved in the portal phase, because hepatic malignancies are hypoechoic during this phase (Wilson et al, 2006).

Contrast-enhanced ultrasound with second-generation contrast agents and LMI techniques allow assessment of vascular morphology, arterial enhancement relative to the liver, and duration of enhancement or “washout” during the portal venous phase (Wilson et al, 2006). These patterns of enhancement can be used to differentiate benign from malignant focal hepatic lesions as shown in the algorithm presented in Figure 13.10 (Brannigan et al, 2004; Ding et al, 2005; Wilson et al, 2006). Washout during the portal venous phase is highly correlated with malignant histology (von Herbay et al, 2009). In the portal phase, metastases and HCC are unenhanced relative to the liver, whereas benign hemangiomas and focal nodular hyperplasia (FNH) have sustained enhancement equal to or greater than the liver (Wilson et al, 2009; Quaia et al, 2004; Brannigan et al, 2004).

FIGURE 13.10 Algorithm for diagnosis of focal liver lesions based on ultrasound contrast enhancement patterns. HCC, hepatocellular carcinoma.

(Modified from Wilson SR, et al, 2006: An algorithm for the diagnosis of focal liver masses using microbubble contrast-enhanced pulse-inversion sonography. Am J Roentgenol 186:1401-1412.)

In a study of 452 liver lesions, contrast-enhanced ultrasound improved diagnostic accuracy for two readers from 49% to 85% and from 51% to 88%, and diagnostic confidence was increased from 0.82 to 0.97 and from 0.83 to 0.98, respectively (Quaia et al, 2004). A multiinstitutional study of 1349 patients with hepatic masses found that contrast-enhanced ultrasound had a diagnostic accuracy of 83% for all benign lesions, 82% for hemangiomas, and 87% for FNH. Diagnostic accuracy for malignant lesions was 96% overall, 91% for hepatic metastases, and 85% for HCC (Strobel et al, 2009).

Diagnostic performance of contrast-enhanced ultrasound has been compared to that of CT and MRI in several studies. There is high concordance for appearance in arterial phase imaging but less so in the portal phase, probably due to the fact the ultrasound contrast remains intravascular, but CT and MR contrast agents diffuse into the interstitium (Wilson et al, 2006). A multicenter study of 134 patients found that in comparison to CT and MRI, contrast-enhanced ultrasound was 30% more sensitive in recognition of malignancy, 16% more specific in exclusion of malignancy, and 23% more accurate (Trillaud et al, 2009).

Cystic Masses

Benign hepatic cysts have smooth, thin walls, posterior acoustic enhancement, and no internal echoes (see Fig. 13.2). A partial, thin septum or puckered wall may be present. A confident diagnosis of simple hepatic cyst can be made on gray-scale ultrasound, and no further evaluation is required. Blood vessels and arteriovenous fistulas or aneurysms may appear cystic on gray-scale ultrasound, but these can be differentiated from simple cysts, because they will have internal flow on Doppler imaging. Lymphoma, which is very hypoechoic, may mimic a cyst rarely, but internal vascularity will be present on Doppler imaging. When cysts are innumerable and also present in the kidneys or pancreas, polycystic disease should be considered (see Chapter 69B).

Complex masses that contain fluid have internal echoes within the fluid, irregular or thickened walls, mural nodularity, and solid areas or calcification (Middleton, 2002). Hemorrhage, infection, and neoplasm should be considered in the differential diagnosis. Hemorrhagic cysts may have internal solid components, septations, and or debris layering in the dependent portion of the cyst, but on Doppler imaging, no internal vascularity is evident. Hepatic hematomas from trauma have variable appearance in the acute phase but become hypoechoic and cystic as they evolve (Korner et al, 2008; McGahan et al, 2006).

Abscess may be solid in appearance initially, but it then becomes cystic with debris; the wall is usually vascular (Fig. 13.11). Echinococcal cysts have variable appearance; they may be simple fluid-filled cysts, they may have undulated membranes from a rupture and detached endocyst, they may contain daughter cysts and/or echogenic material, and they may show calcification (Lewall et al, 1985). When hydatid cysts are infected, they lose their characteristic sonographic appearance and become diffusely hyperechoic.

FIGURE 13.11 Hepatic abscess in a patient after embolization for hepatic neuroendocrine metastases (arrowheads). The abscess (asterisk) contains complex fluid and has posterior acoustic enhancement (arrows).

Cystic neoplasms, such as biliary cystadenoma and cystadenocarcinoma, may be multiloculated or may have a nodular wall. Cystic metastases from sarcomas undergo necrosis and have a thick wall as a result of central liquefaction within a solid mass. Ovarian cystic metastases characteristically are located peripherally, secondary to tumor implants on the hepatic surface.

Cavernous Hemangioma

Typical cavernous hemangioma, seen in 80% of cases, is a brightly echogenic homogeneous mass with a sharp margin (Fig. 13.12). Approximately 10% to 20% of hemangiomas are multiple (Charboneau, 2002). Multiple vascular interfaces within hemangiomas cause the increased echogenicity; the margins are well demarcated because histopathologically, only one cell layer is present at the junction with adjacent liver (Charboneau, 2002). Although hemangiomas are vascular, flow is extremely slow, and usually no Doppler signals are evident. Atypical hemangiomas, seen in 20% of cases, may have a thin echogenic rim and a lacy or granular central appearance from foci of collagen (see Fig. 13.12). Hemangiomas are often smaller than 3 cm, but some are giant hemangiomas that become symptomatic as a result of mass effect. Giant hemangiomas, those greater than 5 cm, often lack the characteristic ultrasound features because of central fibrosis, necrosis, and myxomatous degeneration. When the background liver becomes hyperechoic as a result of fatty infiltration, hemangiomas also may appear atypical and less echogenic than the liver. In such cases, contrast-enhanced ultrasound or CT may be useful (Bartolotta et al, 2007; Liu et al, 2009; see Chapter 79A).

FIGURE 13.12 Hemangiomas. A, Longitudinal sonogram of the right lobe shows a brightly echogenic hemangioma (arrow) with a sharp border. B, Atypical hemangioma (arrow) seen in 20% of patients may have areas of internal heterogeneity, a result of fibrosis or myxomatous degeneration, and may demonstrate a thin echogenic rim (arrowhead).

With ultrasound contrast imaging at LMI settings, hemangiomas have sustained enhancement in the portal phase and on the arterial phase, enhancement intensity is usually less than the liver, and the pattern of enhancement is peripheral and nodular with puddling of contrast and centripetal progression, similar to CT (Wilson et al, 2006; Quaia et al, 2004; Brannigan et al, 2004; Fig. 13.13).

FIGURE 13.13 Giant atypical hemangioma. A, Baseline gray-scale image demonstrates a large, mainly isoechoic left lobe mass (arrows) found in a patient with abdominal pain. B, Contrast-enhanced ultrasound in early arterial phase (19 seconds) demonstrates that the mass has globular peripheral enhancement but no central enhancement. C, In the portal venous phase (56 seconds), the mass demonstrates more marked peripheral globular enhancement consistent with hemangioma.

(Courtesy Paul S. Sidhu, King’s College Hospital, London.)

If the ultrasound shows the classic appearance of hemangioma, and the patient has no history of underlying liver disease or malignancy, then no further investigation is required (Leifer et al, 2000). A study of 213 patients with ultrasounds that showed a typical hemangioma appearance and without risk for hepatic malignancy found only one patient with malignancy on long-term follow-up and concluded that typical hemangiomas in low-risk patients do not require follow-up. This rule does not apply to patients with cirrhosis, hepatitis, or chronic liver disease who are at increased risk for HCC, nor does it apply to patients who already have malignancies. In a study of 2000 patients with cirrhosis, 44 had hemangioma-like lesions; on follow-up, half of these proved to be HCCs, and half were hemangiomas (Caturelli et al, 2001). Thus, in patients at risk for HCC, any echogenic lesion merits further evaluation and follow-up.

Focal Nodular Hyperplasia

The second most common benign hepatic neoplasm, FNH may be difficult to detect on gray-scale ultrasound, because it often has similar echogenicity to the liver, but subtle vascular displacement or contour abnormalities may draw attention to the lesion (Buetow et al, 1996; Hussain et al, 2004; Wilson et al, 2005; Fig. 13.14). FNH typically occurs in asymptomatic patients and is more common in women. Approximately 20% are multiple, and they may occur in association with hemangiomas (Nguyen et al, 1999; see Chapter 79A). FNH contains Kupffer cells, hepatocytes, and biliary structures but without portal triads and central veins found in normal hepatic lobules (Hussain et al, 2004). FNH is considered to be secondary to vascular malformation; fibrous septa radiate from a central scar, and tortuous feeding arteries are present.

FIGURE 13.14 Focal nodular hyperplasia (FNH). As shown in this longitudinal gray-scale image of the right lobe, FNH (crosses) has similar echogenicity to the liver, and lesions are often subtle.

On gray-scale ultrasound, FNH has variable echogenicity and lobulated smooth contour (see Fig. 13.14). A central scar may be evident, but it is usually better depicted on CT, MRI, or contrast-enhanced ultrasound. A key to ultrasound diagnosis of FNH is the characteristic power Doppler appearance of the tortuous feeding artery and spoke-wheel vascularity within the lesion. With ultrasound contrast imaging, the central scar is better seen, and the vascular pattern can be diagnostic. FNH is hypervascular on early arterial phase imaging with diffuse enhancement that is sustained into the early portal phase (Wilson et al, 2006; Kamaya et al, 2009; Dietrich et al, 2005; Ungermann et al, 2007; Brannigan et al, 2004; Quaia et al, 2004). Persistent portal phase enhancement differentiates FNH from malignant liver lesions (Kim et al, 2009; von Herbay et al, 2009; Fig. 13.15)

FIGURE 13.15 Focal nodular hyperplasia (FNH) on contrast-enhanced ultrasound. A, An isoechoic mass (arrows) is seen in the left hepatic lobe on gray-scale imaging. B, Twelve seconds after contrast injection, the FNH (arrows) shows enhancement in a central scar and a characteristic spoke-wheel vascular pattern. C, The FNH (arrows) enhances brightly at 18 seconds. D, Five minutes after contrast injection, the FNH (arrows) is still enhanced and is echogenic relative to the liver.

(Courtesy GE Healthcare, Little Chalfont, United Kingdom.)

Hepatic Adenoma

Hepatic adenomas are rare benign tumors that usually occur in young women with a history of oral contraceptive use, but they can also be seen in men using anabolic steroids (see Chapter 79A). They are usually solitary (80%) but can be multiple in patients with type 1 glycogen storage disease or liver adenomatosis. Adenomas lack portal venous supply and bile ductules; they have disorganized plates of cells separated by dilated sinusoids that are perfused by arterial pressure, which predisposes to hemorrhage (Grazioli et al, 2001). On gray-scale ultrasound, adenomas appear echogenic, because adenoma cells have a high lipid content. Internal hemorrhage within adenomas may produce cystic areas, and calcification may also be present (Grazioli et al, 2001). Spectral Doppler flow may have a continuous waveform (Golli et al, 1994; Bartolozzi et al, 1997). Contrast ultrasound of hepatic adenomas more often shows centripetal or mixed filling during the arterial phase; this pattern differs from FNH, which has spoke-wheel vascularity. On portal phase imaging, adenomas are variable, but most commonly they are isoechoic or slightly hypoechoic relative to the liver (Bartolotta et al, 2007; Kim et al, 2006, 2008; Ricci et al, 2008).

Hepatocellular Carcinoma

HCC (see Chapter 80), one of the most common malignancies worldwide, usually arises in the background of cirrhosis and has poor survival (Collier et al, 1998; El-Serag et al, 1999). Chronic hepatitis B and C infection accounts for the high prevalence of HCC in Southeast Asia, sub-Saharan Africa, and the Mediterranean, and the increased prevalence of these viral infections has contributed to rising incidence of HCC in the United States (Caturelli et al, 2003; El-Serag et al, 1999).

Ultrasound in combination with serum α-fetoprotein (AFP) is used to screen for HCC in high-risk patients, and ultrasound-guided biopsy can be used for diagnosis with 90% accuracy (Caturelli et al, 2004; Collier et al, 1998; Sato et al, 2009; Tong et al, 2001). In patients with elevated AFP, ultrasound was found to have 86% sensitivity, 82% specificity, 55% positive predictive value, and a false-negative rate of only 4% (Mok et al, 2004). A study comparing screened versus unscreened patients with cirrhosis showed improved median survival (17 months vs. 12 months) and improved survival rates at 1, 3, and 5 years for HCC arising in the screened Child-Pugh class B patients, but not in the class C patients (Trevisani et al, 2007). Chan and colleagues (2008) compared survival of HCC arising in screened versus unscreened groups during two different 7-year time periods ending in 1997 and 2004. Long-term survival was better in the screened group and, as a result of improved surgical treatment, the proportion of patients undergoing curative resection increased from 51% to 68% in those two time periods.

On gray-scale ultrasound, HCC may be solitary, multiple, or diffuse, and it has a propensity for venous and bile duct involvement (Wilson et al, 2005). Small HCC tumors (<5 cm) are often hypoechoic but may be hyperechoic from fatty metamorphosis (Choi et al, 1993; Caturelli et al, 2001). Larger HCC tumors are often inhomogeneous because of liquefaction necrosis and fibrosis (Tanaka et al, 1983), and sonographic depiction of HCC may be limited in the setting of severe cirrhosis (Bennett et al, 2002). Thrombosis of portal and/or hepatic veins is frequent, seen in 30% to 60% of patients (Wilson et al, 2005; Fig. 13.16). Tumor thrombus usually expands the vein, and Doppler interrogation may reveal hepatofugal arterial flow within the thrombus, a pathognomonic finding (Shah et al, 2007; Lencioni et al, 1995; Giorgio et al, 2004; Tanaka et al, 1993). HCC may also extend into the bile ducts and cause biliary obstruction or hemobilia (Kojiro et al, 1982).

FIGURE 13.16 Hepatocellular carcinoma with portal vein thrombosis. A, Transverse image shows a hypoechoic right hepatic hepatocellular carcinoma (arrows). B, Transverse image of the main portal vein reveals only a trickle of blood flow (arrow), and the lumen is filled with echogenic thrombus. C, Longitudinal image of the main portal vein shows the long area of thrombus (arrow). D, Color Doppler image of the main portal vein confirms arterial flow (arrow) within the hypoechoic thrombus; this is pathognomonic for tumor thrombus. pv, Portal vein.

HCC is hypervascular with dysmorphic arteries and high-velocity flow on Doppler imaging (Koito et al, 1998; Tanaka et al, 1990). Microbubble ultrasound contrast imaging improves HCC characterization and is highly concordant with CT findings, but contrast-enhanced ultrasound is best at identifying lesions greater than 2 cm (Quaia et al, 2007; Giorgio et al, 2004; Inoue et al, 2009; Jang et al, 2008). Moderately differentiated HCC is hypervascular in the arterial phase, whereas well-differentiated and poorly differentiated HCC may have a more variable appearance. Observation during the portal phase should be extended, as many HCCs have delayed washout for up to 300 seconds (Jang et al, 2007; Fig. 13.17)

FIGURE 13.17 Hepatocellular carcinoma. A, Transverse gray-scale image reveals a large hepatocellular carcinoma (HCC); the mass (m) splays the hepatic veins. ivc, inferior vena cava. B, Contrast-enhanced ultrasound, early arterial portal venous phase, shows dysmorphic vessels in the hypervascular hepatocellular carcinoma (HCC) (arrows). C, Peak arterial portal phase image shows brightly enhancing HCC. D, The HCC (arrows) has very slow and weak washout, not shown until more than 4 minutes from the end of the saline flush.

(Courtesy Stephanie R. Wilson.)

Fibrolamellar Carcinoma

Fibrolamellar HCC, a variant of HCC with better prognosis, characteristically presents as a large solitary mass in an adolescent or young adult without liver disease, and AFP levels are usually normal. On ultrasound, fibrolamellar HCC is well defined, lobulated, and has variable echogenicity. Calcification may also be present (Chung et al, 2009). A central scar may be visible on gray-scale ultrasound in a third to two thirds of patients, but scars are better seen with contrast-enhanced ultrasound (Smith et al, 2008; McLarney et al, 1999). Regional lymphadenopathy is common.

Liver Metastases

Liver metastases (see Chapter 81C) are multiple in 98% of patients and are found in up to 50% of cancer patients at autopsy (Wilson et al, 2005; Middleton, 2002). Liver metastases are frequently detected by gray-scale ultrasound, but CT is the preferred modality to determine extent of disease, number of hepatic lesions, and extrahepatic disease. Response to therapy is also better evaluated by CT, because treatment-related fatty infiltration may obscure lesions that have diminished in size or changed echogenicity as a result of chemotherapy.

On ultrasound, hepatic metastases frequently have a hypoechoic halo that corresponds to fibrosis, compressed sinusoids, and tumor neovascularity surrounding the metastatic deposit (Wernecke et al, 1992a; Kruskal et al, 2000; Fig. 13.18). Uniformly hypoechoic lesions are the second most frequent sonographic appearance of metastases (Middleton, 2002; Fig. 13.19) Hypoechoic appearance and hypoechoic halo are highly suggestive of malignancy. In addition to metastases, lymphoma and HCC should be considered in the differential diagnosis. Contrast-enhanced ultrasound improves detection of individual metastases compared with gray-scale ultrasound, and in some cases, contrast-enhanced ultrasound may reveal more lesions than are evident on CT (Albrecht et al, 2003; see Fig. 13.19).

FIGURE 13.18 Liver metastases with hypoechoic halo. Metastatic neuroendocrine tumors (arrows) in the liver have a hypoechoic halo peripherally and a target appearance. A hypoechoic halo usually indicates malignant lesion or, less commonly, an infectious etiology.

FIGURE 13.19 Hepatic metastases evaluated with contrast-enhanced ultrasound. A, Gray-scale image reveals a hypoechoic mass (arrow) consistent with metastasis. B, In the arterial portal phase after contrast injection, the liver metastases (arrows) have peripheral rim enhancement, and an additional lesion is now evident. C, The metastases show complete washout of contrast in less than a minute; image taken at 54 seconds.

(Courtesy Stephanie R. Wilson.)

The ultrasound appearance of metastases may vary according to the primary origin. Echogenic lesions tend to arise from HCC or gastrointestinal primaries. Vascular metastases, such as those from renal cell carcinoma, islet cell tumors, carcinoids, and choriocarcinoma, also are typically hyperechoic (Tanaka et al, 1990; Marchal et al, 1985; Rubaltelli et al, 1980). Hypoechoic metastases are commonly due to untreated metastatic breast, lung, pancreatic, gastric, and esophageal carcinoma. Lymphoma, which is also hypoechoic because of its uniform cellularity, may be diffusely infiltrative or may have multiple masses; this latter pattern is more common in non-Hodgkin lymphoma or AIDS-related lymphoma (Townsend et al, 1989). Calcified metastases may be found with mucinous adenocarcinoma of the colon as well as other primaries. Cystic hepatic metastases are seen in ovarian and pancreatic cystadenocarcinomas and mucinous colon carcinomas. Metastases undergoing central necrosis, such as sarcomas, also may appear cystic, but the thick surrounding tumor wall is usually evident. When metastases are diffusely infiltrative, such as in breast or lung carcinoma, it can be difficult to detect individual lesions, and the liver may have a pseudocirrhotic appearance (Fennessy et al, 2004; Young et al, 1994).

Steatohepatitis

Fatty liver (see Chapter 65) is commonly encountered incidentally on sonograms of the liver and in patients referred for imaging evaluation of abnormal values on liver function tests. The fat deposition within hepatocytes results in increased echogenicity of the liver relative to the kidney and in decreased acoustic penetration (Fig. 13.20). Because echogenicity is not a quantitative measure, one useful objective sign of fatty infiltration is loss of the echogenic borders of the portal vessels. To reduce subjectivity in fatty liver diagnosis, some investigators advocate comparing the relative echogenicity between the liver and right kidney to that between the spleen and left kidney. Because the normal spleen is more echogenic than liver, if the difference in echogenicity between the liver and right kidney is greater than the difference between the left kidney and spleen, then the liver is abnormally echogenic (Tchelepi et al, 2002).

FIGURE 13.20 Diffuse liver disease. A, Hepatic steatosis, longitudinal right lobe. Fatty liver is very echogenic relative to the right kidney. B, Severe fatty infiltration. There is such marked attenuation of sound that the posterior liver is not visualized. gb, gallbladder. C, Cirrhosis. The left lobe is small, with coarsened echogenicity and a nodular surface (arrows). a, aorta.

In fatty liver disease, the liver edge remains smooth. In severe cases, hepatomegaly may be present. The presence of a fatty liver degrades the quality of sonographic evaluation of the hepatic parenchyma because of decreased acoustic penetration, and focal lesions may escape detection. In addition, lesions that typically appear echogenic against a background of normal liver, such as hemangiomas, may appear hypoechoic in the fatty liver.

A fatty liver may be diffuse and uniform, patchy, or focal. Areas of focal fat deposition are commonly wedge-shaped and echogenic and occur in characteristic locations, such as along the ventral surface of segment IV and adjacent to the falciform ligament. Focal fat deposition may also be lobar, rounded, and masslike or perivascular in distribution. In the latter two cases, magnetic resonance imaging (MRI) with in-phase and opposed-phase sequences is highly useful to confirm that the focal abnormality truly represents intracellular fat deposition and not a mass lesion. Focal fat deposition should not have mass effect, and therefore intrahepatic vessels should traverse these areas of increased echogenicity without displacement. However, this finding is not entirely specific, as occasionally some mass lesions may likewise fail to displace vessels. Finally, within the diffusely fatty liver, certain regions may be spared (see Fig. 13.9). These are typically areas of nonportal venous inflow, such as around the gallbladder (cystic veins) and in the region of the porta hepatis.

Acute Hepatitis

Most imaging modalities, including sonography, are neither sensitive nor specific for the diagnosis of acute hepatitis. Definitive diagnosis of hepatitis is usually made based on clinical and laboratory grounds, with imaging useful for excluding other diagnoses, such as biliary obstruction or hepatic mass lesions (Mortele et al, 2001). The most common sonographic abnormality in hepatitis is hepatomegaly. Periportal edema may be appreciated sonographically, but it is better seen on CT or MRI. Occasionally, sonography may demonstrate very low hepatic parenchymal echogenicity, producing an appearance known as the “starry night” liver for the striking prominence of the periportal echoes against the abnormally dark background liver (Mortele et al, 2001; Tchelepi et al, 2002). Hepatitis may also produce secondary changes in the gallbladder, with irregular lamellated thickening of the gallbladder wall noted, especially in hepatitis A (Tchelepi et al, 2002).

Cirrhosis

Cirrhosis (see Chapter 70A) is a diagnosed histologically by the presence of bridging fibrosis and regenerative nodule formation. Although sonographic findings are neither sensitive nor specific for the diagnosis, certain findings may suggest that cirrhosis is present (Tchelepi et al, 2002). Surface nodularity is the most reliable sign of cirrhosis, and nodularity along the deep surfaces of the liver has been found to be a more sensitive sign of cirrhosis than superficial nodularity (Filly et al, 2002; Fig. 13.20). Surface nodularity may be better appreciated in the presence of ascites. Other sonographic features of cirrhosis include coarsening and inhomogeneity of the parenchymal echo texture and volume loss in the right lobe with relative hypertrophy of the left lobe and caudate.

Stigmata of portal hypertension—such as splenomegaly, ascites, reversal of portal venous flow direction, and varices—indicate that severe parenchymal disease is present. The portal vein diameter may exceed 13 mm, although this is not invariably the case. Enlarged, tortuous hepatic arteries are often present in cirrhotic livers and should not be mistaken for dilated bile ducts or venous collaterals; color and spectral Doppler are useful for making this distinction. Although not widely used, a scoring system has been developed for cirrhosis based on four ultrasound criteria: 1) surface appearance, either normal, irregular, undulated; 2) parenchymal echo texture, either normal, heterogeneous, or coarse; 3) intrahepatic vessel caliber, either normal, obscure, or narrowed; and 4) spleen size, which has been proposed with scores above 7 or 8 having relatively high sensitivity and specificity (Lin et al, 1993).

Elastography is a relative new sonographic technique that has shown great promise for noninvasive quantification of hepatic fibrosis in patients with cirrhosis or chronic hepatitis. Two primary methods for performing elastography with ultrasound have been devised. Transient elastography (TE, FibroScan) uses an ultrasound transducer mounted on the axis of a vibrator. A vibration sent into the tissue induces an elastic shear wave, the velocity of which is directly related to tissue stiffness. The harder the tissue, the faster the shear wave propagates. The TE technique is rapid, operator independent, feasible with nonfasting patients, and it yields results that are reproducible and immediately available (Goyal et al, 2009). A recent trial of the technique in 711 patients with chronic liver disease found that a cutoff value of 17.6 kPa for liver stiffness had 90% positive and negative predictive values for the presence of cirrhosis (Foucher et al, 2006). More recently, the technique of real-time elastography was studied in 79 patients with chronic viral hepatitis and known fibrosis and in 20 healthy, normal volunteers. Unlike TE, real-time elastography uses conventional ultrasound probes and can be performed with modified conventional real-time sonography equipment. Results were promising, with areas under the receiver operating characteristic curve 0.75 for significant fibrosis, 0.73 for severe fibrosis, and 0.69 for cirrhosis (Friedrich-Rust et al, 2007).

Portal Vein Thrombosis and Cavernous Transformation

Portal vein thrombosis may be caused by hypercoagulable states; inflammatory diseases; sepsis; myeloproliferative disorders; neoplasms, especially HCC; and most commonly, portal hypertension (Rossi et al, 2006; Wilson, 2005). Gray-scale findings include solid material within the portal vein lumen that may be anechoic, hypoechoic, isoechoic, or hyperechoic. The portal vein diameter may be normal, or it may be expanded. Doppler imaging may show absent color flow or areas in the vessel that do not fill with color completely. Tumor thrombus may be distinguished from bland thrombus by the presence of arterial waveforms within the thrombus, a finding that is specific but not very sensitive (Rossi et al, 2006; see Fig. 13.16). Recent work with microbubble sonographic contrast media suggests that contrast administration improves both detection and characterization of thrombi in the portal system (Rossi et al, 2006). Very slow flow in the portal venous system may result in a false-positive sonographic diagnosis; therefore CT or MRI can be useful for confirmation, especially in the evaluation of patients prior to liver transplantation.

In the setting of long-standing portal vein thrombosis, a collateral hepatopetal venous circulation develops in the hepatic hilum through recanalization of the thrombus and enlargement of collateral veins (de Groen et al, 1999). This phenomenon is known as cavernous transformation of the portal vein, and it manifests on sonography as numerous serpiginous vessels in the periportal region and nonvisualization of the main portal trunk (Wilson, 2005; Fig. 13.22) In patients with cirrhosis, spectral Doppler should be used to confirm the venous character of the varicosities, because enlarged and tortuous hepatic arteries may have a similar appearance on color Doppler. In chronic portal vein thrombosis and cavernous transformation, the morphology of the liver may become distorted, with enlargement of the caudate lobe and segment IV and atrophy of the left lateral segment and/or periphery of the liver (Vilgrain, 2001; Vilgrain et al, 2006; Tublin et al, 2008).

FIGURE 13.22 Vascular disorders of the liver. A, Cavernous transformation of the portal vein. The portal supply has a serpiginous appearance because of numerous collaterals. gb, gallbladder. B, Outflow obstruction in a patient with venoocclusive disease. On transverse view, the intrahepatic inferior vena cava (arrow) is markedly narrowed, and portions of the right hepatic vein (arrowhead) are obliterated. C, The middle hepatic vein (arrows) is narrowed. ivc, inferior vena cava. D, The middle hepatic vein shows flattened Doppler flow consistent with partial venous obstruction.

Transjugular Intrahepatic Portosystemic Shunt Evaluation

Transjugular intrahepatic portosystemic shunts (TIPS; see Chapter 76E) are often placed for management of the complications of portal hypertension, such as variceal bleeding, or as a bridge to liver transplantation. Color and spectral Doppler ultrasound is essential for evaluation of these shunts, which typically extend from the right portal vein to the right hepatic vein. In normally functioning TIPS, flow in the portal veins should be directed toward the TIPS rather than antegrade into the liver parenchyma. Color Doppler should therefore show reversed flow direction in the left and right portal veins. The flow velocity within the mid and distal segments of the shunt should range from 90 to 200 cm/s (Abbitt, 2002; Tchelepi et al, 2002). TIPS evaluation with ultrasound is usually performed immediately after shunt placement to obtain baseline velocity values and at 3-month intervals thereafter to monitor for dysfunction that could indicate developing stenosis. Absent flow on color and spectral Doppler is 100% sensitive and 96% specific for shunt thrombosis (Abbitt, 2002). Doppler sonography can also detect stenosis of the hepatic vein receiving the shunt. An angle-corrected peak velocity below 50 cm/s or reversal of flow in the proximal portion of the hepatic vein suggests stenosis (Abbitt, 2002; Dodd et al, 1995). Other authors have found that an increase or decrease in peak stent velocity of more than 50 cm/s compared with baseline is 93% sensitive and 77% specific for shunt stenosis (Dodd et al, 1995).

Budd-Chiari Syndrome and Venoocclusive Disease (See Chapter 77)

Hepatic venous outflow obstruction may be secondary to obstruction of the inferior vena cava, hepatic veins, or hepatic venules. Etiologies include venous thrombosis, venous webs, venous obstruction by intrahepatic tumor, or injury to hepatic venules from irradiation and chemotherapy (Cura et al, 2009). The liver becomes enlarged and heterogeneous in echogenicity, but the caudate is typically spared and hypertrophied because of its unique venous drainage (Bargallo et al, 2003). Venous obstruction and ascites are key features of Budd-Chiari syndrome in the acute phase; in the more chronic phase, intrahepatic venous collaterals and regenerating nodules are also present (Brancatelli et al, 2002, 2007).

Ultrasound is a primary imaging method for evaluation of Budd-Chiari syndrome (Kane & Eustace, 1995). Gray-scale imaging of the hepatic vein confluence may show narrowed hepatic veins, tumor encasement, or intraluminal thrombus. On color Doppler imaging the hepatic vein flow alterations and venous collaterals are readily evident, and spectral Doppler analysis shows flattened waveforms or areas of high velocity flow because of the more central venous obstruction (Grant et al, 1989; Ralls et al, 1992; see Fig. 13.22).

Cholelithiasis (See Chapter 30)

Ultrasound is the procedure of choice for detection of gallstones, which are highly echogenic and have posterior acoustic shadowing because of the difference in the acoustic impedance between stones and bile (Fig. 13.24; also see Fig. 13.3). It is important to scan patients in different positions, as gallstones are mobile, whereas polyps are fixed. Tumefactive sludge or thick bile may also change with positional variation, but sludge is thick and moves slowly, whereas gallstones roll quickly on real-time ultrasound. Stones in the gallbladder neck may be difficult to visualize, but when the patient is examined in the left lateral decubitus position, they may roll into the gallbladder body or fundus. Fixed stones impacted in the gallbladder neck are frequently associated with cholecystitis. There may be little surrounding bile if large stones or multiple stones fill the entire gallbladder lumen; when this occurs, gallstones are diagnosed by the “wall echo shadow” sign produced by echoes from the anterior gallbladder wall, anterior surface of the stone, and posterior acoustic shadowing.

FIGURE 13.24 Gallstones and gallbladder carcinoma. A, Longitudinal view of the gallbladder shows layering stones (numbered) with acoustic shadow. A small amount of sludge (arrow) is also seen in the dependent portion of the gallbladder. B, Views of the fundus show focal mural thickening (arrows) suspicious for tumor. RK, right kidney. C, Flow is seen within the mass (m) on color Doppler imaging.

Cholecystitis (See Chapter 31)

Acute cholecystitis is associated with cystic duct or gallbladder neck obstruction from cholelithiasis in approximately 90% to 95% of cases (Bennett et al, 2003). Ultrasound has 80% to 100% sensitivity, 60% to 100% specificity, and positive predictive value of 94% for the diagnosis (Harvey et al, 1999; Ralls et al, 1985; van Breda Vriesman et al, 2007). Ultrasound findings include gallstones, gallbladder wall thickening greater than 3 mm, pericholecystic fluid, and a positive sonographic Murphy sign (Smith et al, 2009; Teefey et al, 1991; Fig. 13.25).

FIGURE 13.25 Acute cholecystitis. A, Longitudinal sonogram shows a distended gallbladder with thickened irregular wall (arrows) and layering sludge (arrowheads). B, Computed tomographic scan shows gallbladder wall thickening (arrows).

Gangrenous cholecystitis occurs more often in patients with diabetes mellitus or a white blood cell count greater than 15,000 cells/mL (Fagan et al, 2003). Ultrasound features include floating intraluminal membranes from sloughed mucosa, shadowing foci from air in the gallbladder wall, disrupted gallbladder wall, and pericholecystic abscess formation ( Jeffrey et al, 1983). When emphysematous cholecystitis is present, air in the gallbladder wall produces acoustic shadowing and reverberation artifact. This may be diagnosed sonographically, but CT is more sensitive for the presence of air in the gallbladder wall (Gill et al, 1997).

Gallbladder wall thickening may be seen in acalculous cholecystitis and in other conditions that cause edema in the pericholecystic space such as hypoalbuminemia, congestive heart failure, hepatitis, and pancreatitis.

Gallbladder Polyps and Adenomyomatosis (See Chapter 48)

The spectrum of hyperplastic cholecystosis includes benign cholesterol polyps and adenomyomatosis. Cholesterol polyps are usually numerous, small (2 to 10 mm), nonshadowing, and fixed in position (see Fig. 13.23). These are benign with no malignant potential (Ito et al, 2009; Levy et al, 2002). Adenomyomatosis results from hypertrophy of the smooth muscle in the gallbladder wall with prominent Rokitansky-Aschoff sinuses, which may appear as cystic spaces in the gallbladder wall with a characteristic ring-down or comet tail artifact (see Fig. 13.4). The gallbladder wall thickening may be focal, segmental, or diffuse. Focal thickening is most often seen in the fundus, but when it occurs in the middle region, it produces an hourglass-shaped gallbladder.

Gallbladder Carcinoma

Carcinoma of the gallbladder (see Chapter 49) may appear as a focal mass, thickened gallbladder wall, or infiltrative mass that fills the gallbladder lumen and extends into the adjacent liver and bile ducts (Wibbenmeyer et al, 1995; see Fig. 13.24). When the mass occupies the gallbladder lumen, it can result in a displaced or “trapped” stone that is fixed in position because of an intraluminal tumor (Khalili & Wilson, 2005). Mucosal plaque and wall irregularity of gallbladder carcinoma can be differentiated from other causes of gallbladder wall thickening when tumor vessels are demonstrated on Doppler imaging. Contiguous hepatic involvement in segments IV and V is common. Gallbladder carcinoma also frequently extends into the hepatic hilus and causes biliary obstruction secondary to direct involvement of the common hepatic duct. Adjacent adenopathy may also be present. Ultrasound performs well for detection of the gallbladder primary and local tumor spread, but CT is necessary to more accurately determine resectability, distant disease spread, and peritoneal metastases (Bach et al, 1998).

In the era of laparoscopic surgery for gallstone disease, it is especially important to assess the gallbladder carefully on preoperative ultrasound to exclude occult gallbladder cancer and to plan an appropriate surgical approach. Unfortunately, approximately 47% of gallbladder carcinomas are detected incidentally at laparoscopic cholecystectomy (Duffy et al, 2008). For these patients, reexploration with definitive resection to clear disease and reexcision of laparoscopic port sites is required (D’Angelica et al, 2009; Duffy et al, 2008; Winston et al, 1999).

Congenital Anomalies

Choledochal cysts (see Chapter 46) are classified according to the location of the dilated ductal segment: type 1 is fusiform common duct dilation, type II is diverticulum of the duct, type III is a choledochocele from a dilated terminal portion of the common bile duct, and type IV is multifocal dilation (Todani et al, 1977). The cysts may be large, and the connection with the bile duct is not always evident sonographically (Khalili & Wilson, 2005). Caroli disease is cystic bile duct malformation that results in multiple diverticulum-like outpouchings from the intrahepatic ducts. On imaging, it is important to demonstrate the connection with the bile ducts to differentiate this condition from polycystic disease. Arterial signals from the fibrovascular bundle at the margin of the saccules may also aid in diagnosis (Miller et al, 1995).

Biliary Obstruction

Ultrasound is the preferred initial imaging study to diagnose biliary obstruction and to determine the level of obstruction (Gibson et al, 1986). Signs of intrahepatic biliary obstruction include the “double track” sign and the “too many tubes” sign caused by dilated bile ducts parallel to the portal vein branches (Fig. 13.26). Bile duct dilation is best appreciated in transverse images of the left hepatic lobe. The pattern of bile duct dilation should be assessed to determine whether it is symmetric in both lobes or localized to one portion of the liver. To determine the level of obstruction, the left and right main ducts are followed transversely to their junction, and transverse views of the common hepatic and common bile ducts are obtained. With this approach it is possible to differentiate intrahepatic from extrahepatic obstruction, and usually the etiology of the obstruction can be determined (Fig. 13.27).

FIGURE 13.26 Biliary obstruction from choledocholithiasis. A, Transverse sonogram of the liver reveals “too many tubes” (circled areas), consistent with intrahepatic biliary dilation. B, Longitudinal view of the common bile duct (cbd) shows an echogenic stone (arrow) that produces acoustic shadowing (arrowheads). gb, gallbladder; v, portal vein.

FIGURE 13.27 Biliary obstruction and vascular encasement from adenopathy at the porta hepatis. A, Mass (m) obstructs a mildly dilated common bile duct (arrows) and involves the main portal vein (v). B, Nodal masses (m) encase the portal vein (v) that is markedly narrowed (arrows). ivc, inferior vena cava. C, Color Doppler image shows narrowed hepatic artery (a; arrow) with dilation proximal to the encased segment.

Bile duct calculi (see Chapters 35 and 39) appear as intraluminal filling defects (see Fig. 13.26). Because there is little bile surrounding the ductal calculi, acoustic shadowing may not always be elicited. Debris or thick bile within the ducts may cause internal echoes within ducts and fluid levels, but debris does not shadow, and it shifts with positional variation. Hemobilia (see Chapter 105) may layer or appear echogenic and masslike if the clot is organized. Intraductal tumor from colorectal metastases or HCC may also appear as an intraductal mass, but tumor often expands the duct, and it does not produce acoustic shadowing (Ghittoni et al, 2009; Jhaveri et al, 2009). Obstruction as a result of biliary ascariasis is associated with tubular structures within the bile duct, and movement of the worms is pathognomonic (Lim, 1990; see Chapter 45).

The pattern of biliary obstruction and the appearance of the duct wall are also useful for diagnosis. Recurrent pyogenic cholangitis (see Chapter 44) is evidenced by dilated ducts with intraductal calculi, but dilation is often segmental, frequently in the left lobe. Lobar atrophy may also be present. Primary sclerosing cholangitis (see Chapter 41) causes a beaded appearance of the ducts with wall thickening, strictures, and discontinuous areas of dilation. Mural thickening of the ducts also is seen with AIDS-associated cholangitis. For tumors adjacent to the umbilical portion of the left portal vein, subtle bile duct dilation or wall thickening may be a clue to biliary ductal involvement.

In patients with biliary obstruction who are being considered for surgical resection or palliative biliary drainage, the distribution of biliary ductal dilation should be carefully evaluated to determine management. For example, if surgery is being considered for a mass or vascular or biliary ductal involvement in one lobe, and main biliary duct obstruction is in a contralateral lobe, the disease is not resectable (Jarnagin et al, 2001). Any isolated biliary ductal segments that do not communicate with the main ducts should be noted, as isolated segments may alter surgical approach, and biliary drainage may require placement of multiple catheters.

Primary Bile Duct Neoplasms

Hilar and Distal Cholangiocarcinoma.

Hilar cholangiocarcinoma or Klatskin tumor (see Chapter 50C), occurs at the biliary confluence. It is best evaluated by ultrasound prior to placement of biliary drainage catheters, as pneumobilia or stent artifact may obscure tumor margins, and the pattern of biliary obstruction may not be evident after decompression. Hilar cholangiocarcinoma is classified by extent of tumor along the duct, and some staging systems also include the status of the portal vein, because direct encasement of the adjacent vessels is common; surgical approach is determined by both portal vein status and ductal spread (de Groen et al, 1999; Jarnagin et al, 2001; Fig. 13.28).

FIGURE 13.28 Klatskin tumor. A, Longitudinal view of the common hepatic duct shows a tapered segment (arrow) consistent with tumor stricture. pv, portal vein; ivc, inferior vena cava. B, Transverse image at the biliary confluence reveals an echogenic mass (arrow) at the bifurcation and dilated bile ducts in both lobes. C, The left portal vein is narrowed and encased (arrows), as shown on color Doppler transverse image. rpv, right portal vein.

On ultrasound, extrahepatic cholangiocarcinoma (see Chapter 50B) may be infiltrating, nodular, or papillary (Hann et al, 1997). The papillary form, which is expansile, has a better prognosis and surgical outcome (Jarnagin et al, 2005). With experienced operators, ultrasound can correctly identify 96% of extrahepatic cholangiocarcinomas and can demonstrate tumor infiltration along the duct wall ( Lim, 2003; Robledo et al, 1996). Ultrasound contrast agents aid diagnosis and staging (Lim et al, 2006). Khalili and colleagues (2003) reported that correct prediction of resectability improved from 69% on noncontrast ultrasound to 96% after contrast.

Solid Neoplasms of the Pancreas

Although best staged by CT, adenocarcinoma of the pancreatic head is often detected on sonography, because ultrasound is typically the initial imaging study for evaluation of jaundice, right upper quadrant pain, and epigastric pain (see Chapter 58B). The typical pancreatic adenocarcinoma is an ill-defined hypoechoic mass associated with abrupt dilation of the upstream pancreatic duct and atrophy of the more distal pancreas (Fig. 13.32). Carcinomas of the pancreatic head typically also cause biliary dilation and a “Courvoisier gallbladder.” Dilation of both the pancreatic duct and common bile duct, the “double-duct sign,” is always a suspicious imaging finding for pancreatic adenocarcinoma, with any imaging modality. Adenocarcinoma in the body or tail of the pancreas is inconsistently visualized on sonography because of variables such as patient body habitus and overlying gastric or bowel gas. Color Doppler sonography typically does not show increased vascularity within pancreatic adenocarcinoma. Nevertheless, this mode can be useful to visualize the peripancreatic vasculature, such as the celiac axis and superior mesenteric vessels, which are often encased in locally advanced disease (Atri & Finnegan, 2005).

FIGURE 13.32 Pancreatic adenocarcinoma. A, Longitudinal view reveals a mildly dilated common bile duct that terminates in a pancreatic head mass (M; arrows). v, Portal vein; ivc, inferior vena cava; bd, bile duct. B, On transverse view, the pancreatic duct (arrow) is dilated. m, Pancreatic head mass; v, splenic vein; a, aorta.

In contrast to adenocarcinomas, neuroendocrine tumors of the pancreas are commonly hypervascular masses (see Chapter 61). They are typically more hypoechoic than adenocarcinomas and may be well defined and smoothly marginated. Intraoperative ultrasound is often a useful adjunct to surgery for pancreatic neuroendocrine tumors, because many are localized and can be safely enucleated. In patients with multiple endocrine neoplasia type I (MEN-I), real-time intraoperative scanning can also help identify additional subcentimeter neuroendocrine tumors that may escape CT detection (Shin et al, 2009).

Cystic Neoplasms of the Pancreas

Cystic pancreatic neoplasms (see Chapter 57) fall into two major categories: serous (microcystic) neoplasms and mucinous (macrocystic) neoplasms. The two should be distinguished on imaging, because serous tumors are always benign, but mucinous tumors have malignant potential. Differentiation can be made on imaging studies based on the internal architecture and size of the cysts. Serous tumors have cysts ranging in size from less than 1 mm to 2 cm in size, and the internal architecture is often described as having a honeycomb appearance (Fig. 13.33). However, these tumors can vary considerably in sonographic appearance based on the sizes of the internal cysts they comprise. Serous tumors composed of relatively larger cysts will appear as partially solid cystic masses on sonography. As the cysts in a serous tumor get smaller, the less likely it is that the individual cyst cavities and walls will be resolvable on ultrasound; instead, the numerous interfaces created by the walls of the cysts will confer an overall echogenic appearance to the mass. A central stellate scar is a characteristic feature of many serous cystadenomas and may be visible as an echogenic or frankly calcified focus with shadowing (Atri & Finnegan, 2005; Hutchins et al, 2009).

FIGURE 13.33 Serous (microcystic adenoma) of the pancreas. Transverse intraoperative ultrasound images of the pancreas. A, Mass in the pancreatic tail has a small cyst (arrow). B, On magnified higher frequency images, the mass has honeycomb architecture.

Mucinous neoplasms, by contrast, are typically unilocular or multilocular cystic masses. The cysts within these tumors are larger than 2 cm, and a solid mural nodule within any of the cyst cavities is suspicious for malignancy in these lesions. Intraductal papillary mucinous tumors (IPMTs) arise from the epithelium of the main pancreatic duct or side branches. Side-branch type IPMTs commonly appear as small unilocular cysts or multilocular cystic masses within the pancreatic parenchyma. Main-duct type IPMTs show up as focal or diffuse dilation of the main pancreatic duct. Again, any internal mural nodularity is concerning for malignancy in these tumors.

Endoscopic ultrasound is very useful in evaluation of cystic pancreatic neoplasms, because the internal architecture of the lesions is much better demonstrated. The lesions can also be aspirated under endoscopic ultrasound, and the fluid can be evaluated for the presence of glycogen (serous tumors), mucin (mucinous tumors), and tumor markers such as carcinoembryonic antigen (Hutchins et al, 2009).