Evaluation and Management of Mucosal and Submucosal Lesions in the Foregut



Fig. 12.1
Splenic artery. Endoscopy shows a pulsatile prominence in the posterior gastric wall. Varying views on EUS shows a tubular, also round anechoic structure. Doppler examination shows pulsatile flow





Incidence and Clinical Presentation


Subepithelial lesions are encountered in 0.3–3.5% of upper endoscopies [5, 6]. The majority of these lesions are in the stomach (51.6–75.5%), followed by esophagus (18.5–41.3%), and duodenum (6–7.1%) [5, 7]. Because a biopsy is not performed or warranted for most SELs, histologically confirmed diagnosis is not available in a majority of cases [1]. A presumptive diagnosis is made based on appearance on endoscopy, endoscopic ultrasound (EUS), and other imaging modalities [1, 2, 4, 8].

Among cases with a histologic diagnosis, 41.8–63.8% are confirmed to be gastrointestinal stromal tumors (GISTs), 21.9–34.3% are leiomyomas, 1.4–10.4% are lipomas, 2.9–8.7% are granular cell tumors, 7.5% are heterotopic pancreas, 3.0–5.5% are cysts, including duplication cysts, and 2.6% are carcinoid tumors [1, 4, 7, 8]. The proportion of the different types of foregut SELs vary based on location. For example, in a study which included only gastric SELs, 30.5% were GISTs, 30.1% were heterotopic pancreas, 15.5% were leiomyomas, 5.8% were lipomas, and 2.2% were schwannomas [9]. Other diagnoses include heterotopic liver, leiomyosarcoma, lymphoma, bronchial carcinoma, and metastasis (renal cell primary, ovarian primary). Of note, in the studies on SEL with histologic confirmation, reasons for tissue acquisition included hypoechogenicity on imaging, size >10 mm, increase in size by >25%, and worrisome endosonographic changes. This indicates that the above incidence rates for the different types of SELs cannot be applied to SELs as a whole.

The majority of small SELs do not increase in size or develop worrisome endosonographic features. In a retrospective study of gastric SELs ≤30 mm followed for up to 13 years (median 7 years), only 8.5% (84 of 989) showed a significant increase in size, developed ulceration, and/or had their appearance on EUS change over a median period of 24 months [10]. Twenty-five lesions underwent resection or enucleation; 19 of the 25 (76%) were found to be GISTs [10]. Of 65 hypoechoic foregut SELs (size 7–29 mm) prospectively followed for up to 9 years (median 2.5 years), 16 (25%) increased in size. Nine had a tissue diagnosis, 6 of which were GISTs [7].

Most foregut SELs are asymptomatic and incidentally discovered during upper endoscopy or abdominal surgery for an unrelated indication. If present, symptoms are often related to the location and size of the lesion. Dysphagia is mostly reported in cases of esophageal SELs when the size of the lesion is >1 cm. Vomiting or other obstructive symptoms can be the presenting symptom for large gastric SELs near the pylorus (Fig. 12.2). Duodenal SELs can be complicated by jaundice and, rarely, pancreatitis [2, 4, 11]. Other presenting symptoms of foregut SELs include GI bleeding (overt, with hematemesis or melena, or occult resulting in anemia), abdominal pain, and chest discomfort.

A352733_1_En_12_Fig2_HTML.jpg


Fig. 12.2
Giant antral lipoma causing intermittent obstruction of the pylorus in a 61-year-old woman with intermittent vomiting. Image courtesy Douglas G. Adler MD


Diagnosis


A diagnosis cannot be made for all SELs solely based on endoscopic visualization. However, upper endoscopy is 95% accurate in identifying a subepithelial lesion as solid, cystic, or vascular [12]. Also, for lipomas, indentation with a closed biopsy forceps during endoscopy (the “pillow sign”) is 98.8% specific [1]. Upper endoscopy can be used to biopsy suspected lipomas in a tunneled manner—if this reveals adipose tissue this is termed the “naked fat” sign. Also upper endoscopy has a sensitivity of 87–98%, but a low specificity of 29–64% in determining whether the lesion is intramural or extramural [1, 2].

Endoscopic ultrasound (EUS) plays an important role in the diagnosis and management of SELs. It is superior to endoscopy in determining whether a lesion is intramural or extramural, with a sensitivity of 92% sensitive and specificity of 100% [2]. EUS helps narrow the differential diagnoses of an SEL based on the echogenicity and layer of origin (Table 12.1). EUS imaging alone is sufficient for diagnosing lipomas, simple cysts, and varices [13, 14]. However, a presumptive diagnosis by EUS without performing biopsy is correct in only 45.5–48% of cases [1, 15]. This is because several SELs (GIST, leiomyoma, schwannoma, heterotopic pancreas) have overlapping endosonographic features. EUS is also helpful in predicting risk of malignancy. Large size (>3 cm) and irregular margins in hypoechoic SELs [16], as well as the presence of perilesional lymphadenopathy, correlate with malignancy or indeterminate malignant potential. EUS allows evaluation for candidacy for endoscopic resectability of an SEL. Findings which indicate that a tumor is generally not amenable to endoscopic resection include (1) extension to the muscularis propria, (2) the presence of lymphadenopathy, and (3) size >2 cm [17, 18].


Table 12.1
EUS features of different SELs































































 
Echogenicity

Layer of origin

Homogeneity

Margin

Lipoma

Intensely hyperechoic

SM

Homogenous

Distinct

Heterotopic pancreas

Hypoechoic or mixed

Mainly SM, also MM and MP

Heterogenous

Mostly indistinct

Carcinoid

Isoechoic or slightly hypoechoic

Mainly SM, also MP

Homogenous

Distinct

Duplication cyst

Anechoic

Any

Homogenous

Distinct

GIST

Hypoechoic or mixed

Mainly MP, rarely MM, SM

Homogenous or heterogenous

Mostly distinct (irregular margins suggest intermediate or high risk of malignancy)

Leiomyoma

Hypoechoic

Mainly MP, also MM

Mostly homogenous

Distinct

Schwannoma

Hypoechoic

Mucosa, SM, MP, serosa

Homogenous

Distinct

Granular cell tumor

Hypoechoic

Mucosa, MM, SM, MP

Homogenous or mildly inhomogenous

Distinct


SM submucosa, MM muscularis mucosa, MP muscularis propria

EUS is frequently used to perform tissue acquisition, usually by EUS-guided fine-needle aspiration (FNA). Diagnostic rate for EUS-guided FNA or biopsy ranges from 34 to 91% [14]. In a recent meta-analysis which included 17 studies from 2004 to 2014, the overall diagnostic rate was found to be 59.9% in 978 cases. There was no difference in the diagnostic rate among fine-needle aspiration, needle biopsy, and trucut biopsy, as well as 19-gauge, 22-gauge, and 25-gauge needles. Only studies which used ability to perform immunohistochemical staining as a parameter for specimen adequacy were included [14]. Other techniques (mucosal incision or unroofing, resection, and dissection) are available which allow for acquisition of more tissue.

Endoscopic mucosal resection (EMR) is a technique which allows not only deeper sampling but also complete removal of a small, superficial lesion via endoscopy. EMR is essentially a variation of the technique of standard polypectomy as performed during colonoscopy and involves submucosal injection of a fluid, usually saline or saline diluted with epinephrine to create a submucosal cushion, followed by band ligation and removal of the lesion via electrocautery with a polypectomy snare (Fig. 12.3). A lesion is considered amenable to EMR if it is less than 2 cm, not invading the musclaris propria, and with no surrounding lymphadenopathy [17, 18]. Endoscopic submucosal dissection (ESD) is another technique that allows en bloc lesion removal, for lesions >2 cm. Margin-negative (curative) resection can be achieved and surgery avoided via ESD, allowing preservation of the native organ. Compared to EMR, ESD has a higher risk of bleeding and perforation a longer procedure duration and is limited to referral centers [19]. Endoscopic full thickness resection can also be performed for deeper lesions (involving the muscularis propria), but this procedure also limited to only a few tertiary centers [20, 21].

A352733_1_En_12_Fig3_HTML.jpg


Fig. 12.3
Endoscopic mucosal resection of a gastric carcinoid. Lesion is suctioned with a cap-fitted endoscope and band ligation is performed, followed by snare resection

Other imaging modalities have been used to characterize SELs. These include computed topography (CT) scan, magnetic resonance imaging (MRI), and magnetic resonance pancreatography (MRCP). Both CT and MRI take advantage of fat content and are able to specifically diagnosis lipomas [22, 23]. Overall accuracy of multidetector CT in detecting and classifying SELs is 85.3% and 78.8%, respectively [24]. While EUS is generally better than CT in detecting foregut SELs, rarely, CT can show a lesion not visualized on EUS if it is too far from the gut wall to be seen clearly [24].


Management


Management of foregut SELs depends on the presence of symptoms attributable to the lesion, as well as concern for a malignant or a potentially malignant condition. Symptomatic lesions should, in general, be treated by endoscopic or surgical resection or enucleation. For duplication cysts, marsupialization is also an option but is rarely required. For asymptomatic lesions, the decision of whether or not to biopsy can be individualized.

If a lesion appears yellow in hue and the “pillow sign” or the “naked fat” signs are positive, the lesion is likely a lipoma and further biopsy is not typically warranted. If a lesion appears vascular (such as a varix, which is often serpiginous and has a bluish hue) or cystic, endoscopic biopsies should not be performed until the target lesion is further evaluated by EUS. On EUS, a varix appears as a round or tubular anechoic structure, usually deep to the submucosa, and Doppler examination shows flow in the structure [25]. If an SEL appears solid, forceps biopsy should be considered. Forceps biopsy is low yield [4, 5, 26] but is often sufficient for carcinoid tumors since these lesions tend to invade the mucosa [27]. If the biopsy of a solid-appearing lesion is non-diagnostic, EUS should be pursued. Alternatively, repeat endoscopy in 1 year can be considered if the lesion is <1 cm and the patient is not enthusiastic about further evaluation. If EUS shows concerning features such as a hypoechoic mass >3 cm in size, referral for surgery should be considered. For hypoechoic lesions less than <3 cm, deeper biopsy (endoscopic mucosal resection, EUS FNA or biopsy, mucosal dissection) may be performed [28].


Foregut SELs



Lipomas


Gastrointestinal lipomas are benign, slowly growing tumors [29]. They can occur in any part of the GI tract but are most commonly found in the colon (65–75%). Four to eight percent of GI lipomas are found in the stomach [23, 29]. Microscopically, they are made up of a fibrous capsule containing mature adipose tissue lobules and thin regular fibrovascular septa [22].

Lipomas are often asymptomatic and incidentally found during endoscopy. Rarely, large lipomas can present with intussusception, GI obstruction, acute, and chronic bleeding, manifesting with anemia.

Prior to the advent of CT and endoscopy, the diagnosis of a lipoma was usually made after surgical intervention [23, 30]. Gastrointestinal lipomas are seen as a fatty density with distinct margins on CT. On MRI, they are characterized by hyperintensity on T1-weighted images [23]. Endoscopy usually shows a yellow to red-orange smooth-surfaced mass. Indentation with a closed biopsy, the “pillow sign,” is 98.8% specific. However, this sign is only 40% sensitive for lipomas [1]. In such cases, EUS can be used for further evaluation. On EUS, GI lipomas usually appear as a homogenous hyperechoic lesion with distinct margins (Fig. 12.4). In some cases, they can have a more atypical and heterogeneous appearance. Approximately 90–95% arise from the submucosal layer, the rest from the subserosa [31]. EUS features are characteristic, and no further diagnostic intervention is usually needed [12]. Biopsies are generally unnecessary and should be limited only to cases with atypical imaging [29].

A352733_1_En_12_Fig4_HTML.jpg


Fig. 12.4
Lipoma

GI lipomas have essentially no malignant potential. Surveillance and treatment are not indicated for incidental lipomas. For symptomatic lipomas, surgical or endoscopic resection can be performed. Removal by endoscopy is contraindicated if EUS shows layer of origin is the serosa or if there is infiltration of the muscularis propria [31].


Pancreatic Rests


A pancreatic rest AKA heterotopic pancreas is pancreatic tissue outside of its normal location. It can be seen anywhere in the GI tract but can be found outside it as well (pelvis, thorax, mesentery, spleen). These lesions are found in up to 2% of the population [32]. Pancreatic rests have been reported in 0.2% of surgical operations involving the upper GI tract and 0.6–13% of autopsies [33]. Heterotopic pancreas can be found in patients at any age but is most often discovered in the fifth and sixth decades of life, more commonly in males. In the GI tract, it is most frequently seen in the gastric antrum, along the greater curve, in the prepyloric space [33].

Pancreatic rests are often asymptomatic. The most common complaint in symptomatic patients is abdominal pain. Other symptoms include GI bleeding, gastric outlet obstruction, biliary obstruction, and weight loss [31, 32]. Pancreatitis, pseudocyst, and cystic neoplasms can arise in heterotopic pancreas rarely [32, 33].

The classic appearance of heterotopic pancreas on endoscopy is a smooth subepithelial lesion with central umbilication, which corresponds with the opening of a partial or complete ductal structure. This umbilication is found in approximately half of cases. Size ranges from 0.1 to 5 cm, usually 0.6–3 cm [34]. Heterotopic pancreas is usually found in the antrum, 2–6 cm from the pylorus, in the 3–7 o’ clock position [35]. Compared with gastric mesenchymal tumors (GISTs, leiomyomas, schwannomas), size <3 cm and location in the middle or lower third of the stomach is 92% sensitive and 78% for heterotopic pancreas [36].

On EUS, a pancreatic rest frequently has a hypoechoic or mixed echogenicity [13, 26, 34, 37]. Margins are indistinct in 72–80% of cases, owing to the lobular structure of the acinous tissue at the margin [24, 26, 37]. The layer of origin is mainly the submucosa, although these lesions can also (less commonly) originate from the muscularis propria and muscularis mucosa [13, 37]. Anechoic areas within a lesion can be seen corresponding to ductal or cystic structures [26, 37].

Histologic diagnosis of heterotopic pancreas is often difficult using endoscopic biopsy forceps [26]. Up to 12.5% of heterotopic pancreas can be diagnosed by endoscopic biopsy alone [32], 36% with biopsy or snare [35]. EUS FNA of pancreatic rests is rarely indicated.

Management of pancreatic rests remains controversial. Malignant transformation has been reported but is exceptionally rare, and an invasive approach is generally not recommended in asymptomatic patients [37]. Tissue acquisition may not always be required, particularly for lesions with a typical appearance on endoscopy or EUS [35]. It is proposed that a pancreatic rest should be removed when it is enlarging or the diagnosis is in doubt, besides lesions causing symptoms [34]. Endoscopic resection appears safe and effective in small-sized lesions (~1 cm) [35].


Spindle Cell Lesions: GIST and Leiomyoma


GISTs are the most common soft tissue sarcoma of the GI tract. They can arise anywhere along the digestive tract, the stomach being the most common primary site (60%) followed by jejunum and ileum (30%). Duodenal GISTs account for 4–5% and rectal GISTs 4% [38, 39].

GISTs have an incidence of 14–20 per million. They are typically found in older adults (median age of 60–65 years). There is a slight male preponderance. Cases are extremely rare in patients younger than 21 years [40].

Close to a third of GISTs are asymptomatic and found incidentally during surgery, imaging, or endoscopy. Of those with symptoms, most present with ulcer-like symptoms or GI bleeding [40].

All GISTs are potentially malignant. Risk of metastases depends on tumor site, size, and mitotic index. For GISTs with malignant behavior, the most common sites of metastases are the liver and peritoneal cavity [40]. Lymph node metastases are extremely rare [39]. It can take up to several years before a primary GIST metastasizes [40].

Most gastrointestinal smooth muscle tumors referred to in the older literature as leiomyomas, leiomyosarcomas, and leiomyoblastomas would now be considered GISTs [38]. However, there remains a subset of GI smooth muscle tumors referred to as leiomyomas (or “true” leiomyomas), leiomyosarcomas, and leiomyoblastomas which are distinct from GISTs. Whereas all GISTs are considered to have malignant potential, leiomyomas are essentially invariably benign [41]. They are more common than GISTs in the esophagus [38]. They stain positive for desmin and negative for c-KIT (CD117) antigenic stain, in contrast to GISTs which stain positive for c-KIT and negative for desmin [42, 43].

On EUS, GISTs and leiomyomas appear as hypoechoic, well-marginated tumors, mostly arising from the muscularis propria, but they can also originate from the muscularis mucosa (Fig. 12.5, Video 12.1) [9, 44]. Compared to leiomyomas, GISTs are sometimes more hyperechoic relative to the surrounding muscle echo. Marginal halo and hyperechogenic foci are also found more commonly in GISTs [41].

A352733_1_En_12_Fig5_HTML.jpg


Fig. 12.5
Gastrointestinal stromal tumor

Biopsy of a primary GIST is needed before preoperative therapy at most institutions [39]. Needle aspiration or biopsy using EUS is preferred over the percutaneous route due to risk of intra-abdominal tumor dissemination, although percutaneous biopsy can be used to confirm metastasis [39]. Ninety-five percent of GISTs express the tyrosine kinase c-KIT (CD117). Immunohistochemical staining for CD 117 confirms the diagnosis. For c-KIT-negative GISTs, immunostaining for DOG1, a chloride channel protein, can be used [39]. Tissue obtained via needle aspiration or biopsy is not always adequate for immunostaining. Other techniques (such as incision or unroofing before biopsy) can be employed to increase tissue acquisition [45]. EUS core needle sampling is often superior to routine cytology with regard to obtaining adequate tissue for immunostaining.

CT with intravenous contrast is the initial imaging modality of choice for biopsy-proven GISTs. It allows for evaluation of extent of disease and the presence of metastases, thus conveying overall staging. A patient with a GIST can subsequently undergo surgery if there are no metastases and resection does not carry a significant risk of morbidity. If there is increased surgical morbidity (such as with pancreaticoduodenectomy for duodenal GISTs or esophagectomy for esophageal GISTs), preoperative imatinib or sunitinib scan be administered to allow for potential downstaging prior to surgery [39]. If there is response, resection should be considered. For progressive disease, options include radiofrequency ablation, embolization or chemoembolization, dose escalation of imatinib, and palliative radiotherapy [39]. Endosonographic surveillance every 6–12 months is an option for gastric GISTs <2 cm without high-risk EUS features such as irregular border, cystic spaces, ulceration, echogenic foci, and heterogeneity, and many patients select observation, especially if they have other comorbidities that would make surgery high risk [39].


Carcinoid Tumors


Neuroendocrine aka carcinoid tumors are slow-growing tumors that arise from neuroendocrine cells and can occur anywhere in the body. The majority of these tumors (86%) are found in the GI tract [46]. Based on SEER registry data, the incidence of neuroendocrine tumors has increased from 1.09 per 100,000 in 1973 to 5.25 per 100,000 as of 2004, likely as a result of the explosive growth and use of endoscopy which identifies many of these lesions [46]. Gastric and duodenal carcinoids are most often diagnosed in the seventh decade of life [46, 47].

The most common site for digestive tract carcinoids is the rectum (29.8% of GI carcinoids), followed by jejunum and ileum (23.2%), stomach (10.4%), colon (6.9%), and duodenum (6.6%) [46]. This is a shift from earlier studies which showed the appendix as the most common site (42.7–49.4%) [47].

Most GI carcinoid tumors are asymptomatic. If present, symptoms include abdominal pain, vomiting, and GI bleeding [48]. Carcinoid syndrome (diarrhea, flushing, abdominal pain) is more common with extra-gastrointestinal primary (such as lung or bronchi) but can occur with GI carcinoids in the setting of liver metastases [49]. Primary GI carcinoids alone do not cause carcinoid syndrome as the biologically active metabolites are destroyed on first pass metabolism through the liver.

On endoscopy, carcinoid tumors can have varying appearances. In some cases, a carcinoid tumor appears as a smooth, round, subepithelial mass with yellowish hue, sometimes with a central depression. In other cases, the lesion can appear to be mucosal in nature. Most carcinoid lesions are firm when probed with a closed biopsy forceps. Most GI carcinoids are small (75% <1 cm) [48], but a tumor size of 23 cm has been reported [50]. EUS often shows a homogenous isoechoic or slightly hypoechoic mass, arising mostly from the submucosa layer (Fig. 12.6) [12]. It can also originate from the muscularis propria or even the mucosa [51]. Forceps biopsy is usually diagnostic [27].
Jul 25, 2017 | Posted by in GASTROENTEROLOGY | Comments Off on Evaluation and Management of Mucosal and Submucosal Lesions in the Foregut

Full access? Get Clinical Tree

Get Clinical Tree app for offline access