The body and tail of the pancreas, including the pancreatic duct, along with the abdominal aorta, celiac axis and its branches, superior mesenteric artery, splenic vein, superior mesenteric vein, and portal vein are often best imaged through the body and fundus of the stomach, as well as along the greater curvature. The left kidney and central portions of the liver and spleen are also seen from the stomach. 8
The head, body, and uncinate process of the pancreas, as well as the gall bladder, common bile duct, portal vein, superior mesenteric vein, inferior vena cava, right kidney, and central portions of the liver are imaged with the EUS scope placed in the duodenum. The ampulla of Vater is recognized as a raised area of the wall of the duodenum, in which are seen the common bile duct and the pancreatic ducts.
Advancing the relatively stiff scope into the duodenum is probably the biggest technical challenge the endoscopist encounters when performing EUS. The first, second, and third portions of the duodenum can be evaluated with EUS, but the fourth portion cannot be seen.
TNM STAGING OF CANCERS
The staging of cancers involving the upper GI tract is the most common application of EUS. With its ability to image the wall layers and to detect adjacent lymph nodes, EUS is more accurate than a CT scan for both the T and N-staging of cancers of the esophagus. 9 T-stage accuracy is 88% with EUS versus 59% with CT scanning, and N-stage accuracy with EUS is 74% versus 54% with CT scanning.
Endoscopic ultrasound has also been shown to be highly accurate when compared with surgery in the T and N-classifications of esophageal tumors. 10 Similar to its role in the esophagus, EUS also provides accurate staging of gastric adenocarcinoma and gastric lymphoma. Whereas it may offer important information in differentiating benign lesions of the stomach from infiltrative ones, biopsy is still necessary to establish a diagnosis.
The majority of endoscopic ultrasounds of the upper GI tract are performed to stage tumors that have previously undergone biopsy. When malignant mucosal tumors enlarge, they penetrate deeper into the wall of the organ as they invade into surrounding structures. While growing, they progressively destroy the layers of the viscus normally seen with EUS. Sonographic staging is performed according to the TNM system. 11 The T-classification is the most important contribution from EUS. The N-classification is not as clear. M-staging (distant metastases) is not applicable to EUS because of its limited views.
Errors result in overstaging or understaging. Oblique scanning can result in overstaging of a T1 or T2 lesion. Microscopic infiltration of the adventitia may not be detectable with EUS, resulting in understaging of a T3 lesion. The sensitivity of EUS in evaluating GI tumors does not remove the need for CT, which is superior to EUS for evaluating distant metastases and the degree of involvement of tumor spread into adjacent organs. The two modalities are complimentary to each other and the information from both is critical in management.
T-CLASSIFICATION OF CANCERS OF THE UPPER GASTROINTESTINAL TRACT
- T1: There is wall thickening involving the mucosa and submucosa, but the muscularis propria remains intact. The submucosa has not been breached by these tumors (Fig. 10).
- T2: The tumor has broken through the submucosa and it has entered the muscularis propria. There is complete loss of the sonographically visualized layers in the wall; however, the outer wall of the muscularis propria remains smooth (Figs. 11A.–C)
- T3: The tumor has crossed through the muscularis and is invading into the adventitia (esophagus) or serosa (stomach), causing irregularity of the outer margin of the tumor. Its outer layer develops small irregular microlobulations or pseudopods. It can acquire a somewhat serrated appearance (Figs. 12A–C).
- T4: The tumor has invaded into other adjacent structures or organs (Figs. 13A,B).
N-CLASSIFICATION OF CANCERS
There is no good standard yet for the criteria for N-classification of lymph nodes seen with EUS (Fig. 14). Normal, small, nontumorous lymph nodes will be seen. Deciding whether a tumor is N1 (malignant lymph nodes near the tumor) or N0 (no lymph nodes or only benign lymph nodes) can be difficult. Generally, small oval nodes smaller than 5 mm are considered benign (N0). Nodes larger than 5 mm. that are rounded and markedly hypoechoic are considered malignant (N1). 12,13 The chance of malignant lymph nodes is more likely with increasing levels of T-classification, and lymph nodes next to the primary tumor that have a similar appearance to the primary are more likely to be malignant. 14
M-CLASSIFICATION OF CANCERS
No attempt is made to evaluate tumors of the upper GI tract for their M-staging, in which distant metastases are detected. For practical purposes, EUS has a limited field of view, and information obtained is generally focused on a relatively small area. This is because of the weak penetration of the ultrasound beam itself, the relatively few scanning windows available for good imaging, and the commonly encountered air-filled structures that absolutely block the transmission of ultrasound. On a rare occasion, liver metastases may be seen and commented on, but the M-staging of cancers of the upper GI tract is reserved for CT or magnetic resonance imaging. It is not a part of a EUS examination.
PITFALLS WITH ENDOSCOPIC ULTRASOUND
The success of an endoscopic ultrasound relies on the skills of the endoscopist and the technical expertise and cross-sectional anatomic knowledge of the operator of the ultrasound component of the equipment. Endoscopic ultrasound is subject to the same physical problems encountered using ultrasound elsewhere in the body. It is restricted by available scanning windows, depth from the probe, gas, nonperpendicular or oblique scanning planes, and equipment operation.
Endoscopic ultrasound examinations can be plagued by uncontrollable artifacts. Several imaging artifacts will be seen in many of the images in this article. We have had annoying problems with electronic interference from cellular phones being used by patients sitting in the waiting area adjacent to the endoscopy suite. It has been necessary to make requests to shut-off these devices when a EUS examination is in progress (Fig. 19B).
Interobserver variability must also be considered with follow-up studies and measurements. Wall thickness can vary with inexact placement of calipers, semi-tangential views in which the layers are stretched out in the image (Fig. 15), and with balloon overdistension. This can cause errors in overestimating or underestimating lesions and in creating pseudolesions.
With EUS, a well-demarcated hypoechoic area can be seen normally in the lateral part of the pancreatic head relative to the more echogenic tissue throughout the remainder of the pancreas. This normal appearance can mimic a tumor, and careful real-time observation is important here. Because of oblique scanning angles, size measurements in the pancreas are not reliable in its evaluation. Bowel loops adjacent to the pancreas are possible sources of errors in diagnosis by simulating masses related to the margins of the pancreas. Careful scanning, maintaining objectivity, and respect for the strengths and limitations of ultrasound are all critical factors in interpreting EUS properly.
DIAGNOSTIC APPLICATIONS OF ENDOSCOPIC ULTRASOUND
In addition to its important role in cancer staging, endoscopic ultrasound of the upper GI tract has other significant applications. Endoscopic ultrasound easily evaluates submucosal lesions within the gut wall, and it also investigates masses that are located outside, but adjacent to, the walls of the upper GI tract. Common submucosal tumors such as leiomyomas and lipomas have characteristic sonographic patterns and typical locations in the wall layers of a viscus. Not all submucosal masses seen endoscopically are tumors. An apparent submucosal mass of the upper GI tract may actually be only a visible bulge caused by extrinsic compression on one of the walls of the upper GI tract by varices and other blood vessels, or by adjacent normal organs such as the trachea, spleen, gallbladder, or bowel loops.
Endoscopic ultrasound is being used more frequently in pancreaticobiliary disease. Changes of chronic pancreatitis (calcifications) can be detected. The pancreatic ducts, small pseudocysts, and small tumors can be observed. Endoscopic ultrasound has been used to stage early pancreatic cancer and it is more reliable than angiography for detecting vascular invasion by pancreatic tumors. It also can localize neuroendocrine tumors with better accuracy (82%) than angiography, transabdominal ultrasound, or CT scan. Endoscopic ultrasound can be more sensitive than endoscopic retrograde cholangiopancreatography (ERCP) in the detection of choledocholithiasis before laparoscopic cholecystectomy, and it is without the risks of ERCP-induced pancreatitis. However, the sensitivity of EUS in these situations is operator-dependent, and it requires considerable experience in the performance of EUS examinations.
Biopsies of submucosal lesions, mediastinal nodes, and pancreatic masses can be performed through any dedicated EUS scope. Directly visualized real-time EUS guidance is also possible with the curvilinear endosonographic scope.
CANCER STAGING OF THE UPPER GASTROINTESTINAL TRACT
The TNM staging of cancer of the esophagus and stomach has been thoroughly discussed, and the sonographic assessment is essentially identical in both organs. Nevertheless, there are certain EUS aspects, peculiar to each organ, that deserve individual emphasis.
Better in this role than CT, EUS is currently the most accurate diagnostic tool for tumor staging in the esophagus. It has a critical role in the therapy, decision making, and prognostic assessment of patients with this disease. 15,16
The esophagus is the easiest part of the upper GI tract to evaluate with ultrasound. Tumors with a T4 classification can be seen invading almost any mediastinal structure, including paraesophageal fat, aortic wall, azygos vein, left atrium, or trachea. Because esophageal tumors can metastasize to the area around the celiac axis, this area should be scanned to detect adenopathy with EUS whenever possible.
Any difficulty in examining esophageal tumors is primarily a technical issue based on tumor size. Some cancers are just too large to allow passage of the scope through a lumen severely narrowed by the tumor itself. In these situations, only the proximal part of the tumor is evaluated, and the staging obtained represents only the lowest score possible. In an incomplete observation of a tumor, because of its size, it must be recognized that more aggressive findings are likely, but just hidden from view.
The resolution of the equipment cannot detect microscopic spread or inflammation around a tumor, and a few tumors can be incorrectly staged by underestimating their extent. As previously mentioned, distortion of layers because of oblique scanning angles across a tumor can lead to incorrect staging by overestimating tumor involvement.
The sensitivity and specificity of EUS in detecting lymph node metastases is high, with both being approximately 90%. Whereas lymph nodes seen with EUS have been analyzed for size, shape, borders, and echotexture, these features by themselves have little role in the N-staging of malignant lesions of the esophagus. With a known esophageal cancer, whenever lymph nodes simply are imaged with EUS, the incidence of N1 disease is 86%. When no lymph nodes are seen, the chance of N0 disease is 79%. 17
Features suggestive of metastatic lymph nodes are a markedly hypoechoic echotexture, sharp borders, a rounded contour, and a size larger than 1 cm. When all of these EUS features are present in lymph nodes in patients with esophageal cancer, malignancy is almost certain. Distant metastases are beyond the limited penetration of EUS transducers, and complimentary cross-sectional imaging studies, such as conventional ultrasound, CT, or magnetic resonance imaging are required for the detection of distant metastases. Whereas EUS has shown accuracy in outcome studies evaluating curative or palliative resectability, and in determining nonresectability of esophageal cancers, 18 it has not offered any help in predicting response to chemotherapy and radiation before surgery. 19
Endoscopic ultrasound is also valuable in examining a surgical anastomosis for tumor recurrence. The normal anastomosis has layered walls. Tumor recurrence behaves like the original tumor and it destroys wall layers. However, there can be false-positives from inflammatory changes. As in almost any focal lesion seen anywhere with EUS, a biopsy is necessary to confirm the diagnosis of recurrence. 20
When cancers of the stomach are imaged with EUS, the accuracy of the observations is similar to what is encountered in the esophagus. These lesions, which originate in the mucosa, are basically hypoechoic masses, with variable heterogeneity, that are evaluated for the presence or absence of the expected layers of the wall of the viscus. The imaging crossover between the sonographic appearance of inflammatory and malignant lesions, where an ulcer is present, must be considered when there is no histologic proof of tumor.
Large tumors present technical ultrasound challenges that cannot be overcome by even the most talented endoscopists. Large tumors have ulcers, recesses, and crypts that contain air (Fig. 16). These pockets of trapped air are an absolute barrier to ultrasound, and any structure located deep to a pocket of air against the probe will be invisible to ultrasound imaging. Ulcerated areas of the mucosa are seen as echogenic foci adjacent to the probe that are accompanied by a dirty shadow, typical of air, in the area deep to the ulcer. Large tumors also have the same limited penetration problems of all high-frequency scanners, and sometimes they cannot be evaluated completely. Their depth is greater than the relatively limited penetration of the high-frequency ultrasound used in EUS. In effect, these areas are not examined and no sonographic information is available about them.
Blood clots within the lumen of the gut can be encountered on occasion (Fig. 17). They are mobile, echogenic, irregularly shaped masses that are obvious intraluminal findings during real-time examination. Clots can be secondary to nonspecific upper GI bleeding, or they may develop after an endoscopic biopsy that was performed just before the EUS examination. Fragile tumors can bleed easily.
Inability to cross a tumor has the same connotations as in the esophagus, and this problem is common in tumors involving the cardia. Questions regarding these invisible areas are unanswered. The staging by EUS is only the maximal seen and most likely, the tumor is more aggressive than the portion imaged. If the apex of the tumor cannot be seen, then the maximal extent is not evaluated. Tumors usually have their most aggressive characteristics at their thickest area, closest to their center of origin.
Endoscopic ultrasound has some limitations in evaluating gastric tumors. Early, small, flat or depressed gastric cancers that can be seen directly via the endoscope may not be able to be found with EUS. Although these small tumors themselves may not be imaged sonographically, the intactness of visualized layers can be considered sonographic evidence of a T1 lesion. As in the esophagus, difficulty is sometimes encountered in the stomach in distinguishing between T2 and T3 lesions.
Linitis plastica is usually marked both by diffuse hypoechoic wall thickening and loss of layers, and it is limited to the gastric wall. 21 However, there are also times when linitis plastica may still have visible layers. The gastric wall is rigid and it may be difficult to differentiate this cancer from lymphoma. 22
The N-classification of lymph node evaluation with EUS in the stomach is slightly different from that in the esophagus. With gastric carcinoma, N0 means that no lymph nodes are detected. N1 refers to the detection of perigastric nodes within 3 cm from the tumor. N2 is reserved for perigastric nodes that are found further than 3 cm from the tumor or if nodes are seen around the celiac axis or its branches.
Similar to what is encountered in the esophagus, the N-staging of gastric cancer is somewhat nebulous. Although quite accurate in predicting the T-stage, EUS is less accurate in predicting the N-stage of cancers of the stomach. Nevertheless, EUS remains superior to CT in predicting both the T and N-stages of gastric cancer. 23 Other conventional cross-sectional imaging studies are still needed for the M-staging of these tumors.
Like esophageal cancers, the N-staging of gastric cancers shows a correlation between the T-staging and the N-staging. A lymph node seen with a T3 lesion is more likely malignant than one with a T1 lesion.
There is variability to the appearance of gastric lymphoma when seen with EUS. It can be a focal or a diffuse hypoechoic process resulting in obliteration of wall layers. It can also extend across the gastric wall and infiltrate into other organs. In addition, multiple lymph nodes, that appear similar in echotexture to the primary, may be seen. Some lesions may be ulcerated. As mentioned earlier, lymphoma can have a sonographic appearance similar to gastric carcinoma. Endoscopic ultrasound is sensitive both in defining tumor extent and in detecting perigastric lymph nodes. It is similar to other upper GI tract tumors in its inability to detect more distant lymphadenopathy because of its limited sonographic windows and penetration. Gastric lymphomas can be accurately reevaluated with EUS during and after treatment. 24
TUMORS OF THE PANCREAS
With its high-resolution imaging and without the interference from intestinal gas, EUS is also a good tool to detect occult lesions of the pancreas. 25 The majority of patients undergoing EUS of the pancreas are those in whom tumor is suspected, and the sensitivity of EUS in detecting all types of pancreatic carcinoma is high. 26 Besides pancreatic carcinoma, endocrine tumors of the pancreas and tumors of the ampulla of Vater can also be evaluated.
The sonographic appearance of pancreatic tumors is somewhat variable. Most pancreatic tumors are seen as hypoechoic focal abnormalities that are easily found against the relatively hyperechoic tissue of the normal pancreas (Fig. 18). Tumors smaller than 3 cm are fairly homogeneous, and they have smoother margins than larger tumors. Larger tumors can be more heterogeneous, with more irregular and indistinct borders. 27 As tumors enlarge, areas of necrosis and calcifications may be seen. Typical of EUS, large tumors are not seen well because of the limited penetration of the probe.
The pancreatic duct is commonly dilated distal to a pancreatic tumor. When pancreatic tumors are in the head of the pancreas, both the common bile duct and the pancreatic duct will usually be dilated.
Most cancers of the pancreas have a relatively similar sonographic appearance, whether they be adenocarcinoma, malignant gastrinoma, squamous cell carcinoma, undifferentiated malignancy, or lymphoma. These tumors cannot be differentiated from each other sonographically.
Cystic tumors of the pancreas, such as cystadenomas and cystadenocarcinomas, are remarkable for the presence of cysts in the neoplastic process. These tumors appear as focal collections of smaller cysts or as a single larger cyst with surrounding hypoechoic tissue. Endoscopic ultrasound cannot differentiate between these cystic tumors.
Benign tumors of the pancreas are rare and tend to be small, well-marginated, hypoechoic masses. However, small pancreatic cancers can have a similar appearance. Typical of ultrasound everywhere, the findings here are sensitive but not specific.
Though conceptually similar, the TNM staging of pancreatic cancer is technically different from the system used for the layered hollow viscera of the upper GI tract. 11 The T-stage is:
- T1: The tumor remains confined to the pancreas.
- T2: The tumor has infiltrated peripancreatic tissue, including the common bile duct and duodenum.
- T3: The tumor has infiltrated adjacent major blood vessels (the portal venous system as well as the celiac axis and its branches), or it has infiltrated into adjacent organs such as the stomach, colon, liver or spleen.
The N-staging of pancreatic tumors has the same nonspecificity as it does wherever lymphadenopathy is seen with ultrasound. As elsewhere, size and appearance have only a subjective role in evaluating adenopathy seen with EUS. Nodes greater than 5 mm that are rounded, hypoechoic, and well-demarcated have a nonspecific appearance suspicious for malignancy.
In addition to observing the characteristics of the pancreatic tumor itself, EUS is used to evaluate the local extent of the disease process, although much of this information can be obtained with other conventional imaging studies. The organs and the major blood vessels surrounding the tumor are investigated for evidence of direct invasion. Tumor infiltration can be seen into the walls of the stomach or duodenum, the common bile duct, liver, spleen, or colon. The most critical areas to evaluate with a known pancreatic cancer are the major peripancreatic blood vessels, specifically the portal venous system (main portal vein, superior mesenteric vein, and splenic vein) as well as the celiac axis and its major branches (hepatic artery and splenic artery). Involvement or lack of involvement of these vessels is important information in detecting resectability of pancreatic cancers.
The celiac axis and its branches can become encased by tumor, which indicates a nonresectable lesion. Rarely are the lumens of these arteries actually invaded by tumor. However, tumor invasion and tumor thrombus can be seen involving the portal venous system. Portal venous involvement should be suspected when the sharp echogenic tumor-vessel interface is lost, or when the tumor-vessel interface is irregular. When frank tumor thrombus is identified, it may cause either complete or partial occlusion of the vessel in question. Meticulous technical care must be taken in avoiding oblique scanning that can create the artifactual impression of a tumor thrombus. 28
Whether EUS has a significant role in the outcome of patients with pancreatic carcinoma is uncertain. Whereas it does not have widespread use for the detection of pancreatic cancer, EUS offers important information in a select groups of patients with this disease, and it can evaluate the pancreas for a small tumor when ductal signs are seen on ERCP.
Endocrine Tumors of the Pancreas
These tumors are often small when detected because they have a tendency to be symptomatic early. Like all small pancreatic tumors, they are well-marginated, hypoechoic, solid masses with a nonspecific appearance. They may be difficult to differentiate from lymph nodes, 29 and they can also be missed. Endoscopic ultrasound seems to be the most accurate imaging modality in detecting endocrine tumors, even when other imaging modalities are negative. However, EUS should not be used routinely, and it should be reserved for proven situations when there is solid clinical or laboratory evidence of an endocrine tumor of the pancreas. Conventional imaging modalities still must be used to search for distant metastases.
TUMORS OF THE DUODENUM AND BILIARY TRACT
Tumors involving the walls of the duodenum itself are evaluated with EUS in a fashion similar to the esophagus and stomach. This involves simply observing the integrity of the visualized wall layers.
Tumors of the Papilla of Vater
These tumors require special attention. They demand endoscopic skill and they must be examined carefully. Small lesions can be missed and balloon overdistention can cause problems in staging these tumors correctly.
When seen, a tumor of the papilla of Vater appears as a focal area of hypoechoic thickening of the wall of the duodenum at the level of the papilla. A dilated common bile duct can be found extending down to the tumor, and sometimes tumor growth within the lumen of the bile duct itself can be observed. With their unique anatomic relationship between duodenum, pancreas, and biliary tree, tumors of the papilla of Vater are staged in the TNM system according to layers seen, tumor size, and other organs involved.
- T1: The tumor is limited to the papilla or ampulla of Vater, and the wall layers of the duodenum, including the submucosa and muscularis propria, are maintained.
- T2: The tumor breaches the submucosal layer, and involvement of the muscularis propria must be assumed.
- T3: The tumor invades less than 2 cm into the pancreas.
- T4: The tumor invades more than 2 cm into the pancreas or other adjacent organs or blood vessels.
The N-staging of these tumors is similar to the same vague criteria in classifying nodes seen in the presence of other tumors. N0 means that no nodes are seen. N1 indicates the presence of nodes. The size and echotexture of the nodes detected are used as subjective criteria for suspicion of malignancy.
Endoscopic ultrasound is sensitive in detecting tumors of the papilla of Vater, 30, but differentiating between a benign adenoma and a small carcinoma is not possible. Endoscopic ultrasound cannot distinguish between a small tumor and focal inflammatory changes in the area of the ampulla, which are common after the passage of a stone. It is also not helpful in the presence of papillitis. Endoscopic ultrasound should only be used for staging a histologically proven tumor at this site. The demonstration of tumor extent is critical information for a surgeon in management of treatment decisions. Whereas EUS is superior to CT and ultrasound in detecting these small tumors of the papilla of Vater, CT and ultrasound are able to detect distant metastases where EUS cannot. 31
Bile Duct Carcinoma
The extent of bile duct carcinoma can be evaluated by EUS. 32 The TNM staging of bile duct tumors is tailored to relate to the particular anatomy involved here. There are three layers in the wall of the bile duct: mucosa, muscularis propria, and serosa. The T-staging is:
- T1: The tumor is limited to the bile duct and the outer margin is smooth. The mucosa and muscularis can be involved.
- T2: The tumor invades into the connective tissue surrounding the duct. The outer margin of the tumor is broken and irregular.
- T3: The tumor invades into other structures in the area: major blood vessels or organs such as pancreas, gallbladder, liver, stomach, duodenum, or colon.
In the nonspecific N-staging of these tumors, N0 indicates that nodes are not seen, and N1 indicates simply that nodes have been found. T-staging and resectability do not correlate well with bile duct cancers, especially T3 lesions. Resectable tumors invading pancreas and duodenum can be confused with nonresectable ones invading portal venous structures.
Whereas EUS is sensitive in detecting small bile duct tumors, it cannot differentiate inflammatory processes from malignancy. Inflammatory stenoses can mimic small bile duct tumors. 33
Gallbladder carcinoma has a spectrum of appearances from a polypoid mass within the lumen of the gallbladder to varying degrees of wall infiltration or extension into surrounding structures such as the liver. The gallbladder, like the bile duct, has three layers, and the T-staging of gallbladder carcinoma is:
- T1: The tumor invades the muscularis propria, but the outer margin is smooth.
- T2: The tumor invades the echogenic serosal layer but does not extend into the liver.
- T3: The tumor invades less than 2 cm into the liver or beyond the serosa on the nonhepatic surface of the gallbladder.
- T4: The tumor invades more than 2 cm into the liver, or it invades two or more adjacent organs.
In the N-staging of gallbladder carcinoma, the N0 and N1 classifications refer simply to the absence or presence of lymph nodes.
SUBMUCOSAL MASSES OF THE UPPER GASTROINTESTINAL TRACT
Examining submucosal masses is the second most common application of endoscopic ultrasound. The differential diagnosis of submucosal tumors of the upper GI tract is greatly facilitated by EUS. 34 It is exactly the same process as tumor staging, and it focuses on observation of the wall layers of the upper GI tract. A submucosal lesion of the upper GI tract is basically an endoscopically visualized mass on one wall of the esophagus, stomach, or duodenum, with normal mucosa overlying the apparent mass. Sometimes, a possible submucosal mass is questioned on a CT scan and a EUS is performed to evaluate the area in question. The characteristics of such a mass are unknown. Endoscopic ultrasound can provide critical information in these cases and it should be performed before a biopsy of a submucosal mass.
Endoscopic ultrasound is able to differentiate intrinsic mural abnormality from extrinsic compression. Submucosal masses may be either real tumors of the gut wall or pseudotumors secondary to a compression effect on the inner wall of the gut by an adjacent normal structure or blood vessel. Extrinsic compression can be caused by a host of things including normal liver, spleen, gallbladder, spine, bronchus, aorta, kidneys, as well as cysts or tumors of these structures that are located within the limited viewing range of the EUS probe. Identifying a blood vessel as the cause of a submucosal mass seen endoscopically prevents an unnecessary biopsy with potentially disastrous complications.
Submucosal lesions can originate in the muscularis mucosa, the submucosa, or the muscularis propria. The echotexture of these lesions helps in their differentiation with EUS. Submucosal lesions are either somewhat hyperechoic, hypoechoic, or anechoic (Figs. 19A–C).
All true submucosal tumors anywhere in the upper gastrointestinal tract can be evaluated in an identical fashion. Endoscopic ultrasound is able to detect these lesions down to a size as small as 3 mm, and it has been proven to be superior to CT in this application. 35 Endoscopic ultrasound is helpful in providing important information about these frequently encountered lesions. It can establish size, margins, echo characteristics, and layer of involvement. Observation by EUS is also an option for small, clearly demarcated, focal abnormalities when surgery may not be indicated. However, obtaining tissue is always the best choice.
The list of different types of submucosal tumors is long, but leiomyomas are the most frequently encountered. There can be an overlap in the appearance of benign submucosal leiomyomas and malignant submucosal tumors (leiomyosarcomas and leiomyoblastoma). Large size, heterogeneity, calcifications, areas of necrosis, and irregular tumor margins are findings that are more suspicious for malignancy. Whereas a biopsy to establish histology is critical, EUS can provide important information as to the degree of involvement.
Leiomyomas originate in smooth muscle layers, mostly in the muscularis propria, but occasionally within the muscularis mucosa. They are usually homogeneously hypoechoic lesions with smooth margins, but occasionally they can be heterogeneous. 36
Lipomas are rare tumors that are well-marginated, somewhat hyperechoic masses located in the submucosal layer. 37 They are less echogenic than the submucosa itself and their margins are smooth.
Other unusual submucosal tumors will be seen anecdotally. These include aberrant pancreas, simple cysts, lymphoma, carcinoid, ectopic spleen, metastases, and other rare entities.
OTHER ABNORMALITIES RELATED TO THE UPPER GASTROINTESTINAL TRACT
A problem area for EUS is in its relative inability to distinguish between malignant and inflammatory changes in the layers of the wall of the GI tract. Diffuse wall thickening, loss of layers, and outer wall irregularity, which simulate tumor, can be seen in patients with Barrett's esophagus on occasion. If a focal abnormality is seen, the appearance would be suspicious for tumor and a biopsy would be necessary. At this time EUS, has little role in the management of patients with Barrett's esophagus until a tumor has been found and has undergone biopsy with conventional endoscopy.
Problems With Motility
Endoscopic ultrasound using dedicated EUS scopes has considerable limitations in evaluating the esophagus in patients with achalasia and other motility disorders. Resistance to passing the scope, inability to inflate the water balloon around the probe, and oblique scanning planes limit accurate detection, locating, and measuring of wall structures. Above the stenosis, all of the layers can appear thickened, and food may be found in a distended lumen. The normal thickness of the esophageal wall is 3 to 4 mm. There is controversy about whether wall thickness measurements are useful. Work is being performed with small catheter-based probes in motility disorders.
A benign gastric polyp is a fairly echogenic mucosal lesion with an intact submucosa on its deeper surface. A large polyp could present technical challenges in an EUS examination.
Thick Gastric Folds
Prominent folds can be seen with both benign and malignant disease, and they can present a diagnostic dilemma. In these situations, EUS has a complimentary role in conventional endoscopy and biopsy. It can offer help in management. Intramural varices are a cause of apparent wall thickening. This situation provides an ideal opportunity for a simple EUS observation to prevent a biopsy.
In hyperplasia and in Menetrier disease, the layered structure of the gastric wall shows marked mucosal enlargement while often retaining a normal thickness of the submucosa and the muscularis. 38 Patients who have inflammatory processes, such as from Helicobacter pylori infection, may show varying degrees of thickening of gastric wall layers (Fig. 20). Patients with anisakiasis have thickening of only the submucosal layer.
Malignancy is also a cause of thickened folds, and linitis plastica and lymphoma are considerations (Fig. 21). Whereas nonspecific in establishing a tissue diagnosis, EUS is helpful in evaluating the extent of the disease process. When the muscularis propria is thickened, usually, malignancy is the cause. In a patient with giant gastric folds and no mucosal ulceration, the EUS finding of a thickened muscularis propria layer is suspicious for malignancy, even if endoscopic biopsy is negative, and further workup is warranted. 39
Patients with linitis plastica show enlargement of both the submucosa and muscularis propria layers. Patients with gastric lymphoma have enlargement of both the mucosal and submucosal layers. These layers can also be thickened in healthy individuals with giant gastric folds. 40
Endoscopic ultrasound offers no help in the evaluation of gastric ulcers. All stages of gastric ulcers have been observed, even after treatment; the findings are nonspecific and often confusing. Ulcers produce an inflammatory response that causes tissue changes that can be seen with EUS. There is a broad spectrum of involvement ranging from a focally contained hypoechoic area to a larger area with loss of layers and sometimes an irregular outer margin. Air trapped in any kind of ulcer can be seen with EUS as an elongated, fuzzy, hyperechoic, reverberating area, originating at the ulcer surface, gradually fading into a sonographic shadow with increasing depth (Fig. 22). A large ulcer could have sonographic features seen with malignant lesions.
Duodenal disorders such as ulcers and tumors are approached in a fashion similar to that of the stomach. Air will be seen within the ulcer centers, and the layers of the wall are evaluated for their degreee of involvmennt.
Whereas EUS is not used to evaluate portal hypertension itself, patients with portal hypertension who undergo an EUS examination for another reason will display a plethora of signs associated with the disease (Fig. 23). As mentioned earlier, EUS can detect varices as a cause of large gastric folds seen with conventional endoscopy, and a biopsy performed on these sites will be prevented.
A dilated portal vein and its tributaries can be found. The azygos vein can also become enlarged. Intramural varices in the submucosa of the esophagus and stomach will be seen in most patients. 41 The endoscopist must be careful not to overfill the water balloon surrounding the EUS probe so that it will not compress and obscure the varices. Collateral vessels outside the walls of the esophagus and stomach will be encountered in the chest and in the abdomen, and they will be especially prominent in the areas of the splenic hilum and adjacent to the left kidney, where splenorenal shunts are not uncommon.
Patients with varices who have been treated with sclerotherapy can have esophageal wall thickening and extramural collaterals after treatment. 42 In some patients, intramural varices can still be seen with EUS after sclerotherapy. 43
Endoscopic ultrasound is able to detect findings associated with pancreatitis, but it cannot differentiate these findings from tumor in many cases. The pancreas can appear normal sonographically in mild acute pancreatitis. A patient with chronic pancreatitis may have a somewhat hypoechoic lobular pancreas with scattered calcifications (Figs. 24A,B).
The sonographic appearance of pancreatitis varies according to the stage of the disease. Its size can range from focally or diffusely enlarged through normal to shrunken, and there is no clear knowledge of what is a normal size. The texture may become less echogenic, and mottled, with areas of calcification. The ducts can be dilated and irregular. If the amount of calcification is severe enough, it will hamper visualization of the gland with EUS and the information obtained will be limited.
In patients with pancreatitis, it is common to find lymph nodes in the areas around the pancreas (Fig. 25). With a variety of pancreatic textures, the goal of EUS is to identify a focal abnormality in whatever texture the pancreas has. Natural tissue contrast makes it easier to identify a hypoechoic mass (tumor) against a hyperechoic background (normal pancreas) than a hypoechoic background (chronic pancreatitis).
Simple pancreatic cysts can be seen against all texture backgrounds (Fig. 26). The cyst walls can be irregular, and sometimes they will contain debris. Fine wrinkles in the walls of cyst are ordinarily caused by external compression from thin bordering structures stronger than the cyst wall itself. This finding is common in high-resolution ultrasound throughout the body. Differentiation from cystic neoplasm and pseudocyst may be difficult at times, but thin wall wrinkles in cyst walls should be recognized as a common normal finding.
In the past, the nonspecificity of EUS limited its role in the management of pancreatitis. However, it was recognized that it often added another layer of important information to the management of specific cases. 44 Endoscopic ultrasound now can be a complement to ERCP. It can offer information if ERCP fails, or it can clarify equivocal findings that are uncovered during an ERCP. Current experience suggests that it may be more sensitive than ERCP for detecting stones in the common bile duct, and its use may prevent ERCP-induced pancreatitis. 45
Endoscopic ultrasound can accurately differentiate between normal and diseased pancreas, and it is unlikely to fail to detect a pancreatic carcinoma. Whereas it is good in excluding the presence of a pancreatic tumor, it is unreliable in differentiating between tumor and inflammation as the cause of a focal pancreatic mass.
Benign Biliary Processes
Benign processes can also be seen in the gallbladder, even though these may not be the indication for the EUS study. Gallstones are common observations seen from the antrum and duodenum, and they are no different from what is seen with conventional ultrasound scanning (Fig. 27). Stones also will be encountered in the common bile duct at times.
Cholesterol polyps with their comet-tail artifacts in the Rokitansky-Aschoff sinuses can be observed. Polyps are small, nonspecific, solid masses. Gallbladder sludge may also be seen. Tumefactive sludge, small polyps, and small tumors may offer difficulty in EUS, and all may appear similar. Differentiation of these small abnormalities may not be possible with EUS alone.
In the hands of talented endoscopists, EUS also may be more sensitive than ERCP for detecting stones in the common bile duct (CBD), 46 and when used in this area, the risks of ERCP-induced pancreatitis are avoided.
Because the aorta is one of the most important anatomic landmarks sought by EUS, abnormalities of the aorta are frequently encountered. Aortic atherosclerotic plaques and aneurysms are common incidental observations (Figs. 28A, B).
THERAPEUTIC APPLICATIONS OF ENDOSCOPIC ULTRASOUND
Endoscopic ultrasound can be used for interventional techniques. Endoscopic ultrasound-guided biopsies of intramural or extramural lesions of the upper GI tract are performed safely, easily, and accurately.
New therapeutic uses of EUS are evolving, including transgastric alcohol, steroid, or Bupivacaine injection into the celiac plexus for pain management in patients with intraabdominal malignancy. 47 Endoscopic ultrasound can be used to select a safe site for transgastric drainage of a pancreatic pseudocyst, ensuring that no critical structure, such as a large blood vessel or a bowel loop, is in the way. 48 It may have a role in the management of patients with varices, and it has also been used in the management of achalasia. The penetrating vessel that causes Dieulafoy lesions in the stomach can be found with EUS, and targeted hemostasis can be performed either with electrocautery or with sclerotic injection. Botox has been injected into the lower esophageal sphincter in patients with achalasia.
Whereas the modality is clearly established, EUS is relatively young. The rapidly expanding technology will only improve the use of endoscopic ultrasound for future new clinical applications. Somewhat time-consuming and with a limited field of view, endoscopic ultrasound has no role in screening the GI tract for nonspecific and vague indications.
CATHETER-BASED ULTRASOUND PROBES
Whereas most EUS work is performed with dedicated flexible endoscopic ultrasound scopes, the development of smaller catheter-based ultrasound probes has offered another way of imaging the GI tract. These smaller probes can pass through smaller strictures, and they generally use higher-frequency transducers for better resolution of wall layers, but with far less penetration than conventional EUS scopes. They can fit through the biopsy channel of conventional endoscopes, and their small size increases the likelihood of being able to traverse a large tumor that narrows the lumen of the gut. This may allow more accurate tumor staging because all of the tumor can be examined sonographically.
Catheter-based probes have no water balloon, and this feature results in less compression and distortion of the walls of viscera. Smaller lesions and varices are seen more clearly. However, the lack of a balloon also has its own problems with poor acoustic coupling because of intestinal gas or convex surfaces of tumors. With these probes, portions of the wall may be unable to be visualized.
Small catheter-based probes have been used experimentally and with success, and clinical applications have evolved. 8 Applications of catheter-based probes are outside the scope of this article. Having higher resolution than dedicated EUS scopes, and causing relatively little distortion of the gut, catheter-based probes have been used in esophageal motility disorders to measure muscle thickness at the lower esophageal sphincter in patients with achalasia and to evaluate the muscularis propria in patients with scleroderma. In patients with portal hypertension, catheter-based probes seem to be more sensitive than dedicated EUS scopes to detect and to measure more accurately the size of varices of the esophagus and stomach.
Endoscopic ultrasound has firmly established itself as an important diagnostic tool that provides critical information to surgeons. It offers valuable information in the T and N-staging of cancers of the upper GI tract and also in the pancreas and biliary system. The finely detailed focused information offered by EUS is available by no other imaging modality today.
In addition to cancer staging, EUS also has a major role in the evaluation of submucosal tumors of the upper GI tract. Other applications of EUS are continuing to evolve. It has been proven to be a safe way for biopsy, and it is now being used interventionally for therapeutic purposes. Experience is also expanding in the use of EUS in the management of nonmalignant diseases of the pancreas and common bile duct.
This article focused on some of the main uses of endoscopic ultrasound in the upper GI tract. Our experience has been with dedicated EUS scopes, and other applications, such as in the lower GI tract or with miniature catheter-based probes, were not discussed.
It is important to stress the nonspecificity of EUS. Whereas it is sensitive in the detection and evaluation of focal abnormalities, it is generally incapable of differentiating between a benign and a malignant process. Endoscopic ultrasound does not substitute for a biopsy. Tissue is almost always needed to establish a diagnosis.
The performance of dependable endoscopic ultrasound is not an easy skill to master. The approach at our institution was to combine the skills of radiologists experienced in diagnostic ultrasound and cross-sectional imaging with the skills of gastroenterologists experienced in endoscopy of the upper GI tract. The result has been a friendly and cooperative team effort that has provided accurate information reliably, objectively, and with confidence. For far more detailed information, readers are referred to excellent textbooks devoted to this topic. 6,8,49,50
1. Rifkin MD, Gordon SJ, Goldberg BB. Sonographic examination of the mediastinum and upper abdomen by fiberoptic gastroscope. Radiology 1984; 151:175–80.
2. Wiersema MJ, Kochman ML, Cramer HM, et al. Endosonography-guided real-time fine needle aspiration biopsy. Gastrointest Endosc 1994; 40:700–7.
3. Vilmann P, Hancke S, Henriksen FW, et al. Endoscopic ultrasonography-guided fine-needle aspiration biopsy of lesions in the upper gastrointestinal tract
. Gastrointest Endosc 1995; 41:230–5.
4. Rubein LD. Endoscopic ultrasonography and the malignant esophageal stricture: implications and complications. Gastrointest Endosc 1995; 41:613–5.
5. Kallimanis GE, Gupta PK, Al-Kawas FH, et al. Endoscopic ultrasound
for staging esophageal cancer, with or without dilation, is clinically important and safe. Gastrointest Endosc 1995; 41:540–6.
6. Rosch T, Classen M. Gastroenterologic Endosonography
New York: Thieme Medical Publishers, 1992:36.
7. Dancygier H. Endosonographic evaluation of biliary tract disease. Gastrointest Endosc Clin N Am 1992; 2:697–713.
8. Liu JB, Scgiano TD, Miller LS. Upper gastrointestinal tract
. In: Liu JB, Goldberg BB, eds. Endoluminal Ultrasound. London: Martin Dunitz, 1998:150–2.
9. Tio TL, Coene PPLO, Schouwink MH, et al. Esophagogastric carcinoma: preoperative TNM classification with endosonography. Radiology 1989; 173:411–7.
10. Murata Y, Suzuki S, Hashimoto H. Endoscopic ultrasonography of the upper gastrointestinal tract
. Surg Endosc 1988; 2:180–3.
11. Sobin LH, Hermanek P, Hutter RP. TNM classification of malignant tumors. Cancer 1988; 61:2310–4.
12. Dancygier H, Classen M. Endosopic ultrasonography in esophageal diseases. Gastrointest Endosc 1989; 35:220–5.
13. Tio TL, Tytgat GNJ. Endosopic ultrasonography in analyzing peri-intestinal lymph node abnormality. Preliminary results of studies in vitro and in vivo. Scand J Gastroenterol 1986; 21(S123):158–63.
14. Tio TL, Cohen P, Coene PP, et al. Endosonography and computed tomography of esophageal carcinoma. Gastroenterology 1989; 96:1478–86.
15. Vilgrain V, Mompoint D, Palazzo L, et al. Staging of oesophageal carcinoma: Comparison of results with endoscopic sonography
and CT. AJR Am J Roentgenol 1990; 155:277–81.
16. Botet JF, Lightdale CJ, Zauber G, et al. Preoperative staging of esophageal cancer: comparison of endoscopic US and dynamic CT. Radiology 1991; 181:419–25.
17. Catalano MF, Sivak Jr, MV Rice T, et al. Endosonographic features predictive of lymph node metastasis. Gastrointest Endosc 1994; 40:442–6.
18. Rosch T, Lorenz R, Zenker K, et al. Local staging and assessment of resectability in carcinoma of the esophagus, stomach and duodenum by endoscopic ultrasonography. Gastrointest Endosc 1992; 38:460–7.
19. Souquet JC, Napoleon B, Descos F, et al. Contribution of endosonography to the follow-up of esophageal squamous cell carcinoma treated by multimodal protocol (abstract). Gastrointest Endosc 1991; 37:243.
20. Lightdale CJ, Botet JF, Kelsen DP, et al. Diagnosis of recurrent upper gastrointestinal cancer at the surgical anastomosis by endoscopic ultrasound
. Gastrintest Endosc 1989; 35:407–12.
21. Andriulli A, Recchia S, DeAngelis C et al. Endoscopic ultrasonographic evaluation of patients with biopsy negative linitis plastica. Gastointest Endosc 1990; 36;611–5.
22. Bolondi L, Casanova P, Caletti G, et al. Primary gastric lymphoma versus gastric carcinoma: Endoscopic US evaluation. Radiology 1987; 165:821–6.
23. Botet JF, Ligthdale CJ, Zauber AG, et al. Preoperative staging of gastric cancer; comparison of endoscopic US and dynamic CT. Radiology 1991; 181:426–32.
24. Fujishima H, Misawa T, Maruoka A, et al. Staging and follow-up of primary gastric lymphoma by endoscopic ultrasonography. Am J Gastroenterol 1991; 86:719–24.
25. Zerbey AL, Lee MJ, Brugge WR, et al. Endoscopic sonography
of the upper gastrointestinal tract
and pancreas. AJR Am J Roentgenol 1996; 166:45–50.
26. Rosch T, Lorenz R, Braig C, et al. Endoscopic ultrasound
in pancreatic tumor diagnosis. Gastrointest Endosc 1991; 37:347–52.
27. Yasuda K, Mukai H, Fujimoto S, et al. The diagnosis of pancreatic cancer by endoscopic ultrasonography. Gastrointest Endosc 1988; 34:1–8.
28. Rosch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by endoscopic ultrasonography. Gastroenterology 1992; 102:188–99.
29. Lightdale CJ, Botet JF, Woodruff JM, et al. Localization of endocrine tumors of the pancreas with endoscopic ultrasound
. Cancer 1991; 68:1815–20.
30. Tio TL, Tytgat GNJ, Cikot RJLM, et al. Ampullopancreatic carcinoma: Preoperative TNM classification with endosonography. Radiology 1990; 175:455–61.
31. Rosch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by endoscopic ultrasonography. Gastroenterology 1992; 102:188–99.
32. Tio TL, Cheng J, Wijers OB, et al. Endosonographic TNM staging of extrahepatic bile duct cancer: Comparison with pathological staging. Gastroenterology 1991; 100:1351–61.
33. Rosch T, Lorenz R, Hagenmuller F, et al. Is endosonography useful in the diagnosis and staging of biliary tumors? (Abstract). Gastrointest Endosc 1991; 27:254.
34. Kawamoto K, Ymada Y, Utsunomiya T, et.al. Gastrointestinal submucosal tumors: Evaluation with endoscopic ultrasound
. Radiology 1997; 205:733–40
35. Yasuda K, Nakajima M, Yoshida S, et al. The diagnosis of submucosal tumors of the stomach by endoscopic ultrasonography. Gastrointest Endosc 1989; 35:10–5.
36. Rosch T, Lorenz R, Dancygier H, et al. Endosonographic diagnosis of submucosal upper GI tract tumors. Scand J Gastoenterol 1992; 27:1–8.
37. Nakamura S, Idia M, Suekane H, et al. Endoscopic removal of gastric lipoma: Diagnostic value of endoscopic ultrasonography. Am J Gastroenterol 1991; 86:619–21.
38. Rosch T, Classen M. Gastroenterologic Endosonography
. Stuttgart, Germany and New York: Thieme Medical Publishers, 1992;85.
39. Mendis RE, Gerdes H, Lightdale CJ, et al. Large gastric folds: a diagnostic approach using endoscopic ultrasonography. Gastrointest Endosc 1994; 40:437–41.
40. Songur Y, Okai T, Watanabe H, et al. Endosonographic evaluation of giant gastric folds. Gastrointest Endosc 1995; 41:468–74.
41. Caletti GC, Brocchi E, Baraldini M, et al. Assessment of portal hypertension by endoscopic ultrasonography. Gastointest Endosc 1990; 36: S21–S27.
42. Tio TL, Kimmings AN, Rauws EAJ, et al. Endosonography of gastroesophageal varices (abstract). Gastrointest Endosc 1991; 37:241–2.
43. Ziegler K, Gregor M, Zimmer T, et al. Evaluation of endosonography in sclerotherapy of esophageal varices. Endoscopy 1991; 23:247–50.
44. Sahai AV, Zimmerman MD, Aabakken L, et al. Prospective assessment of the ability of endoscopic ultrasound
to diagnose, exclude, or establish the severity of chronic pancreatitis found by endoscopic retrograde cholangiopancreatography. Gastrointest Endosc 1998; 48:18–25.
45. Canto MI, Chak A, Stellato T, et al. Endoscopic ultrasonography versus cholangiography for the diagnosis of choledocholithiasis. Gastrointest Endosc 1998; 47:439–48.
46. de Ledinghen V, Lecesne R, Raymond JM, et al. Diagnosis of choledocholithiasis: US or magnetic resonance cholangiography? A prospective controlled study. Gastrointest Endosc 1999; 49:26–31.
47. Wiersema MJ, Wiersema LM. Endosonography-guided celiac plexus neurolysis. Gastrointest Endosc 1996; 44:656–62.
48. Knox TA. Endoscopic ultrasound
: diagnostic and therapeutic uses. Surg Endosc 1998; 12:1088–90.
49. Van Dam J, Sivak MV. Gastrointestinal Endosonography
. Philadelphia: WB Saunders, 1999.
50. Dancygier H, Lightdale CJ. Endosonography in Gastroenterology: Principles, Techniques, Findings
. Stuttgart, Germany and New York: Thieme Medical Publishers, 1999.
Keywords:© 2001 Lippincott Williams & Wilkins, Inc.
Endoscopic ultrasound; Sonography; Gastrointestinal tract; Cancer staging; Submucosal masses; Pancreaticobiliary disease; Therapeutic endoscopy