Pyloric gland adenomas (PGAs) are uncommon to rare and are found mostly in the stomach but in a number of other anatomical sites as well. They were recognised years ago by European and Japanese colleagues whereas North American pathologists learned to diagnose them more recently. They are associated with conditions that result in pyloric metaplasia, the prototype of which is autoimmune gastritis. Since the latter is more common in women and men, gastric PGA has a striking female predominance. There appear to be differences in the distribution of PGA in various populations. Herein we review PGA and note similarities and differences in their distribution in our institutions in Germany and the USA and review their morphological appearance, immunolabelling profile, and preliminary genetic data on these unusual lesions.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Pyloric gland adenoma (PGA) is a previously and currently underdiagnosed entity that was initially recognised by Japanese and European colleagues1–8 and more slowly by North American pathologists.9 ,10 As such, most series and case reports that gave insight into the clinical behaviour and molecular findings were published in European and Japanese journals. The delayed recognition of PGA in North America lies in the fact that dysplasia with gastric differentiation appears completely different from the ordinary intestinal type of dysplasia with which most pathologists are quite familiar. The first description of PGA dates back to 1976. Kurt Elster11 from Bayreuth described an adenoma-like lesion of pyloric glands in a book chapter, although the images accompanying the description may simply be of glandular hyperplasia. The resemblance of these lesions to gastric pyloric glands accounts to the current terminology. Elster further noted that, especially in the prepyloric antrum, a proliferation of mucoid glands with lobule formation can be seen occasionally, resulting in an adenoma-like lesion. He further noted that, when applying critical standards, it was difficult to exclude that the lesions were simply hyperplastic since encapsulation was lacking. The reader was cautioned about diagnosing a highly differentiated adenoma of the antral or pyloric region and stated that the same processes can be found at the cardia.11
It was not until 1986 and 1987 when Franz Borchard1 showed a case of a PGA with transition to a moderately differentiated adenocarcinoma during two International Academy of Pathology (IAP) slide seminars in Germany. A first immunohistochemical analysis was published by Borchard et al in 1990.12 Borchard et al used ‘home brewed’ antibodies against m1 (today: Apomucin 5AC) and m2 (today: Mucin 6). Apomucin 5AC is known to stain foveolar epithelium whereas Mucin 6 stains the deep parts of the gastric mucosa (figures 1 and 2).
In PGA, Apomucin 5AC was classically described as expressed more or less within a pre-existing rind of surface foveolar epithelium whereas Mucin 6 stains the mucoid proliferations all the way up to the surface but, in fact, both mucins can display marked coexpression throughout a lesion (figure 3). Rarely, the layer of Apomucin 5AC positive foveolar epithelium is overrun (figure 4). For example, on re-review of data published in a series by one of us in 2010,13 in 1.7% cases, no Apomucin 5AC was expressed, whereas in just over a quarter of cases (26.7%) there was focal coexpression with Mucin 6 and in nearly half of cases (45%) there was a diffuse display of the classically described apical single layer of foveolar epithelium expressing Apomucin 5AC.
In 1990, independently from Borchard, Hidenobu Watanabe from Nijgata, Japan14 described PGA as a unique novel entity among other gastric epithelial tumours within the WHO classification. The German literature has also used the term ‘gastric differentiated adenoma’, 12 which pays more tribute to differentiation than to the cell of origin idea that typifies English terminology. The English term PGA is more often used in the literature than the German term ‘gastric differentiated adenoma’ which adds to further confusion. Furthermore, in cases with transition to an adenocarcinoma the terminology ‘gastric differentiated adenoma with transition to well differentiated adenocarcinoma’ is difficult to understand since the term ‘differentiated’ has two different meanings.
Clinical features and diagnostic criteria for PGA
Though classically described in the stomach, PGA can also be encountered elsewhere in the gastrointestinal tract. More precisely, they can be found in any site in which pyloric metaplasia or tissue (including gastric heterotopia or gastric metaplasia) can arise and thus PGAs have been encountered in the cervix, bile ducts, gallbladder, in rectal gastric heterotopia, and in the duodenum. These lesions were fully characterised in 2003 by Vieth et al in Germany7 and finally confirmed by Chen et al9 in North America much later. These polyps show a preference for the gastric body, and gastric examples show a remarkable female predominance. This female predominance occurs because many PGAs arise in patients with autoimmune (metaplastic) atrophic gastritis, which has a striking female predominance and characteristically features ample pyloric metaplasia. In one study of gastric polyps in patients with autoimmune gastritis, among 461 patients with autoimmune (metaplastic) atrophic gastritis, 143 had endoscopically identifiable lesions. The lesions (n=240) consisted of 179 polyps (138 hyperplastic polyps, 20 oxyntic mucosa pseudopolyps, 18 intestinal-type gastric adenomas and only 3 PGAs). As such, even in the setting of autoimmune gastritis, PGAs are relatively rare.10 More recently, similar lesions have been observed in patients with familial adenomatous polyposis (FAP) by Wood et al15 and by Lee et al16 in patients with Lynch syndrome. Histologically, PGAs are composed of closely packed pyloric gland-like cuboidal to low columnar epithelium showing pale or eosinophilic, ‘ground glass’ cytoplasm (figures 5 and 6). Nuclei are round without prominent nucleoli.17 Foci of dysplasia/carcinoma are fairly commonly encountered. High grade dysplasia is seen in some cases and invasive carcinoma is associated with 12%–47% of the lesions, depending on the authors’ criteria for carcinoma.13 PGAs show coexpression of MUC6 (marker of pyloric gland mucin) and some MUC5AC (marker of foveolar mucin), and, in their pure form, they lack expression of MUC2 (marker of intestinal mucin) and CDX2. While gastric foveolar-type adenomas (figure 7) show MUC5AC expression, they lack expression of MUC6 and MUC2.13 Some PGAs, however, show areas of transition from gastric to intestinal differentiation and these foci may show immunolabelling with MUC2 and CD10.13 As with other types of adenomas, complete excision of PGA with biopsy of the background flat mucosa (especially in the stomach) is appropriate in these patients, as PGAs often arise in the setting of chronic mucosal injury. Although the morphology is very characteristic and does not require further immunohistochemical analysis, immunohistochemistry can be helpful in identifying small foci, especially in unusual locations outside the stomach. Normally PGAs are strongly positive for deep gastric MUC6 with a small overlying variable expression of MUC5AC. Sometimes this small, often single, epithelial layer of MUC5AC can be absent, also. Notably Concanavalin A or MUC3 (stains pyloric glands, gastric mucous neck cells and Brunner's glands) shows a similar expression to MUC6.13 ,18 In gastric lesions, these immunohistochemical labelling patterns can help to differentiate PGA from other gastric polyps such as hyerplastic polyps, foveolar hyperplasia and other tubular adenomas such as the common intestinal type tubular adenoma (figure 8), and also from the above already mentioned foveolar tubular adenoma.19 ,20 Of interest, in patients with FAP, their PGAs arose in pristine rather than damaged oxyntic mucosa.15
The differential diagnosis of PGA includes all lesions with hyperplastic epithelium such as hyperplastic polyps and reactive foveolar hyperplasia that result from proliferation of foveolar epithelium rather from deep (pyloric) gastric glands.20 Epithelia are tall columnar compared with the epithelium in PGA and have paler cytoplasm on H&E. Each cell also displays a crisply delineated apical cap of mucin rather than the more bubbly ground glass like cytoplasm in PGA; essentially PGA cytoplasm is more eosinophilic.17 Notably, the rare Brunner's adenoma cannot be differentiated from PGA morphologically or immunohistochemically. Site, location and origin from Brunner's glands are the only features that allow differentiation. Whether Brunner's adenoma is really a distinct entity or just another extragastric example of PGA is the topic of debate. Brunner's gland adenoma is readily confused with Brunner's gland hyperplasia, including in published reports.21 The latter can be encountered in association with chronic pancreatitis.22 ,23 It has been clearly shown using characteristics of mucin expression that most such lesions are reactive with a component of gastric metaplasia rather than true and rare neoplasia.24 Sometimes these reactive lesions are also referred erroneously as hamartomas.25
Combinations or differential diagnosis of PanIN or BilIN with PGA can hamper diagnosis as well but can be sorted out immunohistochemically.26
Immature regenerative epithelium is a differential diagnosis as well but can be differentiated in and outside the stomach with mucin analysis and on morphological grounds.
Foci of malignant transformation in PGA show a higher ki67 proliferation index than the remaining lesion with abrupt transition towards low grade dysplasia. P53 is usually weakly staining and is not really helpful in identifying foci of malignant transformation showing either high-grade dysplasia (HGD) or carcinoma,3 ,13 but the proliferation index by Mib1 can help to identify areas with higher proliferation and thus suspicious for higher grade of neoplasia.17 Cystic dilations are commonly seen, especially in larger lesions, but they are not a characteristic feature of PGA. The diagnosis of HGD and carcinoma is solely based on morphology, including cytological features for non-invasive lesions as to be expected in HGD and in addition, in carcinomas, intertubular fusion or lateral expansion of the glands should be present.27 The criteria to differentiate HGD from early invasive mucosal carcinoma may differ and are not accepted worldwide. Regardless, the discussion is somewhat academic since all observers would agree that these lesions should be completely removed and metastases are very rare in such early lesions.7
Where do PGAs arise?
The majority of PGAs can be expected in the gastric body (corpus). Moreover, most PGAs should not be expected in the normal gastric mucosa. This has been shown in the past7 and is still valid in a more recent analysis of a PGA cohort of 373 patients, during 1990–2014, in Bayreuth (Europe; see table 1) and in a smaller cohort from Baltimore (North America, table 2) since PGAs were recognised much later by one of us (EAM) than the other (MV); table 2 is derived from 189 cases diagnosed between 2006 and 2014 at Johns Hopkins Hospital. Interestingly, in the European cohort, more than 1% of the cases are found in rectal gastric heterotopia, thus supporting the idea of genesis on the basis of gastric/pyloric metaplasia and heterotopia as the precursor lesion. Thus, as noted above, PGA can be found outside the stomach including the cervix, bile duct, gall bladder, pancreatic duct, corpus heterotopia in the rectum and especially the small bowel4 ,5 In the current Bayreuth analysis (see table 1), three further PGAs were identified in rectal gastric heterotopias after the initial description of Bayreuth cases.28 These lesions are rare, but not as rare as anticipated.
In a first series from 2003,7 30% of PGAs showed transition to adenocarcinoma at the time of first diagnosis. The actual numbers are in Bayreuth (n=373): 28% with stable frequency in almost 25 years, and in Baltimore (n=189): 11%. Differences in percentage of malignant transition can be explained by a partly different population but also by a different threshold for early invasive carcinoma since criteria for diagnosing the first sign of invasion are not uniformly used worldwide. One of us (MV) has a lower threshold for diagnosing early carcinomas than the other (EAM).
Most of the PGAs are found in elderly female patients (73.7±11.3 years; Bayreuth) in autoimmune gastritis (tables 3 and 4). Male patients are a bit younger on average (70.3±11.3 years; Bayreuth) and less likely to have autoimmune gastritis (table 5). When selecting those with PGA and autoimmune gastritis, age and gender distribution still differ between men and women (age 72.4±13.2 years (m) vs 75.3±9.1 years (f); gender m:f=17:69). Differences between the North American population in the patient cohort from Baltimore and the middle European population in Bayreuth can be explained by differences in the population itself. It is known that Hispanics tend to present with autoimmune gastritis more commonly and at a younger age than non-Hispanics10 and autoimmune gastritis is a precursor for PGA. In Baltimore there was a 2:1 female to male ratio for gastric PGA and most (77%) were associated with autoimmune gastritis. The median age of men with PGA and autoimmune gastritis was 76 years whereas it was 73 years for women in the Baltimore population (women were younger). Additionally, cholecystectomies are frequent operations in the USA and account for finding many gall bladder PGAs in Baltimore, and Johns Hopkins is a centre for Whipple's procedures, so pancreatobiliary lesions may be over-represented. Lastly, Johns Hopkins is a referral centre for patients with FAP, and PGA seems to be a feature of FAP.
Molecular findings in PGA
Not much is known about the molecular pathology of PGA. Moreover, a recent report indicates that gastric adenocarcinoma of the fundic gland type/gastric adenocarcinoma with chief cell differentiation (termed oxyntic gland adenoma by some North American observers, figure 929) shares common genetic and phenotypical features with PGA and thus is more closely related than previously anticipated based on a mucous neck cell/chief cell lineage phenotype.19 It is known that GNAS and KRAS mutations occur at high frequencies in gastric and duodenal PGAs and thus they are considered to represent a characteristic feature.30 Comparative genomic hybridization analysis in PGA demonstrated chromosomal abnormalities commonly found in gastric adenocarcinoma such as gains of 17pq and 20q, and losses of 4q,5q and 6q.31 These findings can differ with location. For example, a reported PGA associated with Barrett's oesophagus showed losses at 2p24–24.2, 2q14.1-ter, 5q31.3–32, 6q23–24, 8q23–24, 8q23–24.2, 11q22.3–24 and 18q21.1–22.32 These findings suggest an unstable and precancerous nature of PGA. Indeed, carcinomas in PGA have shown that gains and losses can be identified in 13 chromosomes. Chromosomes 15–17 and 19–22 show gains in various degrees whereas 2–8, 10–13 and 18 show losses and 1, 7, 9 and X show losses and gains.31 Six per cent of patients with FAP developed upper gstrointestinal-tract PGA in a recent series.15 Thus, the identification of a PGA in normal gastric mucosa should lead to a consideration of FAP or attenuated forms of FAP (see tables 3⇑–5). Interestingly it is speculated that PGA may represent a precursor lesion to gastric adenocarcinoma in patients with Lynch syndrome based on the finding that all PGAs in Lynch syndrome were microsatellite instability (MSI)-high and none of the analysed sporadic PGAs were MSI-high. The conclusion was drawn that the mismatch-repair deficient pathway in gastric carcinogenesis is an early event in patients with Lynch syndrome with PGA.16 In patients with juvenile polyposis, dysplastic lesions that arise in gastric juvenile polyps can also show pyloric gland differentiation in a background on non-damaged gastric mucosa, but the mechanism for this remains unstudied.33
In summary, PGA is an uncommon type of mucosal polyp that is most often detected in the stomachs of middle-aged to elderly women that has some proclivity to undergo malignant degeneration and that is not well recognised. As the recognition increases and the population ages, we suspect these interesting lesions will become better studied.
Take home messages
Pyloric gland adenomas are precursor lesions that classically arise in the gastric body of patients with autoimmune gastritis.
Pyloric gland adenomas often arise in areas where pyloric metaplasia is common—autoimmune gastritis features pyloric metaplasia as does the gall bladder.
A subset of pyloric gland adenomas progresses to adenocarcinoma.
Some syndromic gastric polyps can differentiate along the lines of pyloric gland adenoma.
Contributors This is for a theme issue that we were asked to submit.
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.