For submucosal lesions, accurate preoperative diagnosis can prevent unnecessary surgical resection of benign lesions. Before EUS, this was difficult because submucosal lesions grow underneath the mucosa and cannot be readily diagnosed by endoscopy and biopsy. EUS appears most useful for differentiating submucosal lesions from extrinsic compression and for characterizing the submucosal lesion. Extrinsic compression from adjacent organs or blood vessels is correctly differentiated from a submucosal lesion in 94% of cases. Using EUS, the size, layer of origin, margins, and echo-pattern of the submucosal lesion are determined; the layer of origin and echotexture are most useful in suggesting a diagnosis (Table 36–2) although imaging alone is only 45.5% accurate for diagnosis. Malignancy can be inferred from larger lesions (>3 cm) with irregular margins. EUS-guided FNA has limited diagnostic yield with definitive diagnosis possible in only 34–43% of cases although a recent study suggests a higher yield up to 68%.
Table 36–2. Endoscopic Ultrasound (EUS) Characteristics of Submucosal Lesions. ||Download (.pdf)
Table 36–2. Endoscopic Ultrasound (EUS) Characteristics of Submucosal Lesions.
|GIST||Hypoechoic, 4th layer|
|Lipoma||Hyperechoic, 3rd layer|
|Carcinoid||Mildly hypoechoic; 1st, 2nd, or 3rd layer|
|Cyst||Anechoic, 2nd or 3rd layer|
|Pancreatic rest||Hypoechoic or heterogeneous; 2nd, 3rd, 4th layer; ductal structures|
|Granular cell tumor||Heterogeneous, 3rd layer|
The most common submucosal lesion is a gastrointestinal stromal tumor (GIST), which has a mesenchymal cell of origin. On EUS, GISTs are hypoechoic and typically arise from the fourth muscularis propria layer (Figure 36–9). Presence of an irregular border, cystic spaces, echogenic foci, and size greater than 3 cm are suggestive of malignancy. Lipoma is the second most common submucosal lesion and may be readily diagnosed during endoscopy by its yellowish hue and soft texture. If the diagnosis is unclear, EUS should be performed; it will reveal a hyperechoic lesion arising from the third submucosal layer (Figure 36–10 and Plate 93). Cysts are also easily diagnosed by EUS as anechoic structures lying within the second or third layer.
EUS image of a gastrointestinal stromal tumor.
EUS image of a lipoma. (For an endoscopic view, see Plate 93.)
Carcinoids are mildly hypoechoic, homogeneous and typically arise from the first, second, or third layer (Figure 36–11 and Plate 94). Size is usually predictive of malignancy, and they are usually benign if less than 2 cm. Other malignant submucosal tumors include metastases, which are rare, and lymphomas, which may appear as hypoechoic heterogeneous masses within the gastrointestinal wall.
EUS image of a rectal carcinoid. (For an endoscopic view, see Plate 94.)
A few other benign submucosal lesions include pancreatic rests and granular cell tumors. Pancreatic rests are submucosal deposits of ectopic pancreatic tissue that typically occur in the antrum and have a characteristic umbilicated appearance on endoscopy. On EUS, they appear hypoechoic or heterogeneous, arise from the second, third, or fourth layer and may contain anechoic ductal structures (Figure 36–12 and Plate 95). Granular cell tumors are believed to arise from neural tissue and appear heterogeneous within the third submucosal layer.
EUS image of a pancreatic rest. (For an endoscopic view, see Plate 95.)
Management of these submucosal lesions is guided by their diagnosis, size, location, and potential for malignancy. GISTs with several EUS features raising concern for malignancy should be removed surgically. Smaller benign-appearing GISTs may be observed, although the appropriate follow-up interval is unknown. Lipomas should only be removed if symptomatic. Carcinoids smaller than 2 cm confined to the first three layers can be removed endoscopically, whereas larger lesions involving the fourth muscularis propria layer should be surgically resected.
Mekky MA, Yamao K, Sawaki A, Mizuno N, Hara K, Nafeh MA, Osman AM, Koshikawa T, Yatabe Y, Bhatia V. Diagnostic utility of EUS-guided FNA in patients with gastric submucosal tumors. Gastrointest Endosc
Philipper M, Hollerbach S, Gabbert HE, et al. Prospective comparison of endoscopic ultrasound-guided fine-needle aspiration and surgical histology in upper gastrointestinal submucosal tumors. Endoscopy
Large gastric folds present a diagnostic dilemma that often necessitates multiple diagnostic studies. The differential diagnosis is broad and includes malignant and benign conditions (Table 36–3). Endoscopic appearance does not usually enable differentiation between malignant and benign conditions, and superficial mucosal biopsies may miss a malignancy. Snare resection or cap-assisted endoscopic mucosal resection increases diagnostic yield from 17% to 87% but carries an increased risk of complications from bleeding and potentially perforation.
Table 36–3. Causes of Thickened Gastric Folds. ||Download (.pdf)
Table 36–3. Causes of Thickened Gastric Folds.
Granulomatous disease (Crohn, sarcoidosis, secondary syphilis)
Acute Helicobacter pylori gastritis
Few studies exist examining the utility of EUS in evaluating a thickened gastric wall (Figure 36–13). The main predictor of malignancy is thickening of the deep gastric layers, including submucosa, muscularis propria, and serosa, whereas enlargement of the superficial mucosal layers is associated with benign conditions or MALT lymphoma, which is usually readily diagnosed with biopsies. Ascites and lymph nodes are present in over 60% of patients with gastric wall thickening from malignancy.
EUS images of the gastric wall. A: Normal gastric wall. B: Thickened gastric wall.
Ginés A, Pellise M, Fernández-Esparrach G, et al. Endoscopic ultrasonography in patients with large gastric folds at endoscopy and biopsies negative for malignancy: predictors of malignant disease and clinical impact. Am J Gastroenterol.
Pancreatic cysts are increasingly discovered, since abdominal imaging studies have improved in quality and have become more frequently utilized. These lesions are often of unclear clinical significance and pose a diagnostic dilemma. Pseudocysts account for the majority of pancreatic cysts while an increasing minority is pancreatic cystic neoplasms, which may be benign serous cystadenomas or premalignant or malignant mucinous lesions that include mucinous cystic neoplasms or intrapapillary mucinous neoplasm (IPMN). Therefore, it is important to differentiate among these different cysts.
Serous cystadenomas occur anywhere throughout the pancreas typically in women over the age of 60 years. On radiology or EUS imaging, they usually appear microcystic and less commonly macrocystic or solid due to the presence of numerous microcysts that give the appearance of a homogeneous hypoechoic mass. A central calcification is pathognomonic but is only seen in about 10% of serous cystadenomas. Malignant transformation is very rare, and these cysts can typically be followed.
Mucinous cystic neoplasms are premalignant lesions that generally occur in women between 40 and 50 years old. They typically appear macrocystic in the body and tail of the pancreas with a rare, but pathognomonic, peripheral eccentric calcification. IPMN is also a mucinous cystic neoplasm that arises from the pancreatic ductal epithelium and communicates with the main duct, side branches, or both (Figure 36–14). It occurs more commonly in men between the ages of 50 and 60 years. About 40% of main-duct IPMNs contain malignancy at the time of diagnosis. Features suggestive of malignancy in side-branch IPMNs include presence of a mass, mural nodules, dilated main pancreatic duct, and cyst size greater than 3 cm, although a recent study demonstrated presence of malignancy in smaller cysts.
EUS image of intraductal papillary mucinous neoplasm.
Accurate preoperative diagnosis of serous cystadenoma, mucinous cystadenoma, and mucinous cystadenocarcinoma using CT scanning occurs in about 20–30% of cases. MRCP may be more helpful than endoscopic retrograde pancreatography (ERP) in diagnosing IPMN by delineating the main pancreatic duct, side branches, and communicating cystic lesions. EUS imaging alone is not sufficient to diagnose pancreatic cystic lesions. EUS-FNA probably improves diagnostic yield for pancreatic cystic lesions; however, the optimal cyst fluid markers remain unknown. Current evidence suggests that carcinoembryonic antigen (CEA) may be most useful while cytology has poor sensitivity and accuracy in differentiating mucinous from nonmucinous lesions as well as benign from malignant or premalignant lesions (Table 36–4). Molecular analysis of pancreatic cyst DNA for mutations, including early k-ras mutation followed by allelic loss, has been similarly disappointing in differentiating mucinous from non-mucinous cystic lesions.
Table 36–4. Diagnostic Markers for Pancreatic Cysts. ||Download (.pdf)
Table 36–4. Diagnostic Markers for Pancreatic Cysts.
|CEA <5 ng/mL (serous vs other lesions)||100%||86–93%|
|CEA >400 ng/mL (mucinous vs other lesions)||57%||100%|
|Cytology (mucinous vs other lesions)||27%||100%|
|Amylase >5000 units/mL (pseudocyst vs other lesions)||61–94%||58–74%|
|k-ras mutation (mucinous vs other lesions)||11–45%||45%|
|Allelic loss amplitude||67%||66%|
Further work is necessary to discover new and more accurate markers of malignancy and mucinous cystic lesions as pancreatic cystic lesions are increasingly uncovered on incidental imaging studies.
Noh KW, Pungpapong S, Raimondo M. Role of endosonography in non-malignant pancreatic diseases. World J Gastroenterol.
Diagnosis of chronic pancreatitis remains challenging, especially because there is no defined gold standard. Histology may be considered the true gold standard; however, it is available in a minority of patients, sampling error may occur if only a core biopsy specimen is obtained, and there is no consensus on a histologic grading scale for severity of chronic pancreatitis. ERP is less attractive as a diagnostic modality due to its potential complications and decreased sensitivity for early-stage chronic pancreatitis. In addition, the Cambridge classification for ERP changes in chronic pancreatitis was based on expert consensus and has not been validated. Only one study has compared secretin-stimulated functional studies with histology; it found accuracy of the functional test to be 81%.
EUS is an attractive alternative diagnostic possibility due to its relatively low morbidity and ability to assess both parenchymal and ductal features. EUS criteria for chronic pancreatitis include the following parenchymal and ductal changes: hyperechoic foci, hyperechoic strands, lobulation, cysts, calcifications, main duct dilation, main duct irregularity, hyperechoic walls of the main duct, and visible side branches (Figure 36–15). Retrospective studies have noted that the presence of four or more EUS criteria had 84–91% sensitivity and 86–100% specificity for diagnosing chronic pancreatitis using histology as the gold standard. The threshold for diagnosing chronic pancreatitis can be varied depending on whether one is trying to establish or exclude the diagnosis. In the previous study, presence of five or more criteria had 100% specificity. In a population at low to moderate risk of chronic pancreatitis, EUS is most accurate when unambiguously normal with two or fewer criteria or abnormal with five or more criteria present. This is especially true in elderly patients, in whom up to three EUS criteria for chronic pancreatitis may be present without actual pancreatic disease. Certain features including calcification and lobulation are more indicative of chronic pancreatitis, and recently a weighted scoring system accounting for these factors has been proposed.
EUS image of chronic pancreatitis demonstrating lobularity.
The clinical significance of the presence of three or four EUS criteria for chronic pancreatitis is unclear. The addition of FNA does not appear to increase diagnostic yield for chronic pancreatitis. A study comparing EUS and MRCP using the gold standards of ERP, histology, or long-term clinical follow-up of median 15 months demonstrated higher sensitivity of 93% for EUS compared with 65% sensitivity for MRCP, and similar specificity of 90–93%. The combination of both studies yielded higher sensitivity and specificity of 98% and 83%. Therefore, the less invasive modalities of EUS and MRCP have largely supplanted ERCP in the diagnosis of chronic pancreatitis. The expert consensus Rosemont EUS criteria for chronic pancreatitis has been proposed, however, has not been validated and interobserver variability for diagnosing chronic pancreatitis was no better using the Rosemont classification than the standard EUS criteria. More recent studies suggest that the addition of endoscopic pancreatic function testing, which can be performed during routine endoscopy or endoscopic ultrasound, may enable earlier diagnosis of chronic pancreatitis.
Wallace MB. Chronic pancreatitis. Gastrointest Endosc
Autoimmune pancreatitis (AIP) can be extremely difficult to diagnose and often mimics pancreatic cancer in presentation with a pancreatic mass. EUS-FNA findings are often nonspecific. Use of a 19-gauge trucut needle has been reported in small case series to yield histologic diagnosis of AIP. In addition, a small Japanese study has proposed novel EUS criteria for AIP, which remains to be validated. Certainly appreciation of a diffusely hypoechoic, enlarged pancreas with chronic inflammatory cells on cytology should raise the concern for AIP.
Mizuno N, Bhatia V, Hosoda W, et al. Histological diagnosis of autoimmune pancreatitis using EUS-guided trucut biopsy: a comparison study with EUS-FNA. J Gastroenterol.
EUS has similarly proven to be a powerfully accurate and safe method for detecting choledocholithiasis. A meta-analysis of randomized controlled blinded trials comparing EUS and MRCP with the gold standard of ERCP or intraoperative cholangiography in patients with suspected bile duct stones demonstrated sensitivity for EUS and MRCP of 93% and 85%, with similar specificities of 96% and 93%, respectively. A cost-analysis study indicated that initial EUS rather than MRCP had the greatest cost-utility by reducing unnecessary ERCP procedures.
Arguedas MR, Dupont AW, Wilcox CM. Where do ERCP, endoscopic ultrasound, magnetic resonance cholangiopancreatography, and intraoperative cholangiography fit in the management of acute biliary pancreatitis? A decision analysis model. Am J Gastroenterol
Verma D, Kapadia A, Eisen G, Adler D. EUS vs. MRCP for detection of choledocholithiasis. Gastrointest Endosc