**3.2 ABG: Waiting for a pathological definition**

The designation of ABG has been extensively used without a specific definition of its limits. It is possible that the roots of this problem reside in the conceptual vagueness of what is called "atrophic gastritis" or even "gastric atrophy" (Genta, 1997). The criterion of ABG adopted for the present study involves only advanced cases of oxintopeptic mucosal atrophy with the antral mucosa showing no relevant histological changes. From a histological point of view, these patterns overlap those described for gastritis of autoimmune etiology. This definition of ABG is close to the one designated as "metaplastic autoimmune atrophic gastritis" (Park et al., 2010), and far from those defined by other authors (Vannella et al. 2011).

Atrophic Body Gastritis: A Challenge for the

mucosa.

of pseudoantral metaplasia

Presumptive Endoscopic and Histologic Diagnosis of Autoimmune Gastritis 177

easily detected in both the hyperplastic endocrine nodules present in the *lamina propria* and diffusely in the walls of metaplastic glands (Fig. 5). Both intestinal metaplasia and pseudoantral metaplasia were frequent findings in these patients with ABG. Among 60 patients with a diagnosis of ABG, 51 presented areas of pseudoantral metaplasia in the body atrophic mucosa and most of them, i.e., 37 (72.5%) presented ghrelin-immunoreactive endocrine cells in the wall of pseudoantral metaplastic glands (Moreira & Barbosa, 2011). This fact may be of help in the differentiation between antral and body mucosa in cases of ABG. One of the methods currently used in this differentiation is the demonstration the absence of gastrin-producing cells (G cells) in the metaplastic glands of the body (pseudoantral metaplasia) since these cells are only present in the glands of the antral mucosa. In doubtful cases, this differential characteristic between the antral and pseudoantral mucosa has been used to characterize the mucosa of the body with pseudoantral metaplasia, which does not contain G cells, and to differentiate it from the true antral mucosa, which contains G cells (Park et al. 2010).Since this type of characterization of pseudoantral metaplasia is based on a negative fact (absence of G cells), the frequent presence of ghrelinimmunoreactive cells in pseudoantral metaplasia (a positive fact) could be a more reliable marker for this purpose, since these cells are rare in the glands of the normal antropyloric

Fig. 5. A and B - Atrophic body mucosa of the stomach of a patient presenting ABG. Numerous glands exhibit pseudoantral metaplasia and most of them have endocrine cells immunoreactive to the peptide ghrelin (arrows). As ghrelin-immunoreactive cells are usually not present or are rare in antral mucosa this could be used as an reliable indication

This endocrine cell hyperplasia does not seem to occur in multifocal atrophic gastritis. Therefore, the presence of endocrine cell hyperplasia in the body mucosa of patients with ABG is a strong signal indicating the autoimmune etiology of the ABG. These florid changes of the body mucosa contrast with the discrete histological findings of the antral mucosa*.* 

Therefore, when the atrophy of the gastric mucosa is clearly developed, with severe atrophy of the gastric body and fundus, the histological diagnosis of ABG can be relatively easy. Oxintopeptic glands are replaced entirely, or almost entirely, with intestinal glands (intestinal metaplasia), mucous glands (pseudoantral metaplasia) and other poorly differentiated glandular structures (Figs. 2 C, D, and 3). The endocrine cells of the oxyntic mucosa are spared from the process of atrophy, become hyperplastic what is commonly found in the ABG of autoimmune origin. Although these hyperplastic cells are considered to be enterochromaffin-like cells which are strongly reactive to the neuroendocrine marker chromogranin they can also express immunoreactivity to some peptide hormones such as ghrelin (Moreira et al. 2010).

Fig. 4. Body gastric mucosa from a patient with autoimune ABG. A - Immersed in the *lamina propria* there is a nodule (arrow) composed of small cells, overlapping and even suggesting that it is related to endocrine cells (HE staining). B – Histological section of the same paraffin block stained by Grimelius technique showing hyperplasia of endocrine cells (argyrophilic cells) in the wall of the glands and hyperplastic nodules immersed in the *lamina propria* (arrow). These small nodules can be visualized in routine histological preparations, stained by HE and should raise the suspicion for the diagnosis of ABG. The presence of more intense inflammatory infiltrate in the *lamina propria* may mask the visualization of these endocrine nodules. For demonstrating the endocrine nature of these nodules, it is necessary special stainings as Grimelius technique for argyrophilic cells or immunohistochemistry using neuroendocrine markers, eg., chromogranin and neuron specific enolase

According to recent results from our laboratory regarding a large series of patients with ABG, we observed that most of the subjects presented an expressive number of ghrelinimmunoreactive cells in the different types of endocrine hyperplasia that occur in the atrophic body mucosa of these patients. Thus, ghrelin- expressing endocrine cells can be

Therefore, when the atrophy of the gastric mucosa is clearly developed, with severe atrophy of the gastric body and fundus, the histological diagnosis of ABG can be relatively easy. Oxintopeptic glands are replaced entirely, or almost entirely, with intestinal glands (intestinal metaplasia), mucous glands (pseudoantral metaplasia) and other poorly differentiated glandular structures (Figs. 2 C, D, and 3). The endocrine cells of the oxyntic mucosa are spared from the process of atrophy, become hyperplastic what is commonly found in the ABG of autoimmune origin. Although these hyperplastic cells are considered to be enterochromaffin-like cells which are strongly reactive to the neuroendocrine marker chromogranin they can also express immunoreactivity to some peptide hormones such as

Fig. 4. Body gastric mucosa from a patient with autoimune ABG. A - Immersed in the *lamina propria* there is a nodule (arrow) composed of small cells, overlapping and even suggesting that it is related to endocrine cells (HE staining). B – Histological section of the same paraffin block stained by Grimelius technique showing hyperplasia of endocrine cells (argyrophilic cells) in the wall of the glands and hyperplastic nodules immersed in the *lamina propria* (arrow). These small nodules can be visualized in routine histological preparations, stained by HE and should raise the suspicion for the diagnosis of ABG. The presence of more intense inflammatory infiltrate in the *lamina propria* may mask the visualization of these endocrine nodules. For demonstrating the endocrine nature of these nodules, it is necessary special stainings as Grimelius technique for argyrophilic cells or immunohistochemistry using neuroendocrine markers, eg., chromogranin and neuron

According to recent results from our laboratory regarding a large series of patients with ABG, we observed that most of the subjects presented an expressive number of ghrelinimmunoreactive cells in the different types of endocrine hyperplasia that occur in the atrophic body mucosa of these patients. Thus, ghrelin- expressing endocrine cells can be

ghrelin (Moreira et al. 2010).

specific enolase

easily detected in both the hyperplastic endocrine nodules present in the *lamina propria* and diffusely in the walls of metaplastic glands (Fig. 5). Both intestinal metaplasia and pseudoantral metaplasia were frequent findings in these patients with ABG. Among 60 patients with a diagnosis of ABG, 51 presented areas of pseudoantral metaplasia in the body atrophic mucosa and most of them, i.e., 37 (72.5%) presented ghrelin-immunoreactive endocrine cells in the wall of pseudoantral metaplastic glands (Moreira & Barbosa, 2011). This fact may be of help in the differentiation between antral and body mucosa in cases of ABG. One of the methods currently used in this differentiation is the demonstration the absence of gastrin-producing cells (G cells) in the metaplastic glands of the body (pseudoantral metaplasia) since these cells are only present in the glands of the antral mucosa. In doubtful cases, this differential characteristic between the antral and pseudoantral mucosa has been used to characterize the mucosa of the body with pseudoantral metaplasia, which does not contain G cells, and to differentiate it from the true antral mucosa, which contains G cells (Park et al. 2010).Since this type of characterization of pseudoantral metaplasia is based on a negative fact (absence of G cells), the frequent presence of ghrelinimmunoreactive cells in pseudoantral metaplasia (a positive fact) could be a more reliable marker for this purpose, since these cells are rare in the glands of the normal antropyloric mucosa.

Fig. 5. A and B - Atrophic body mucosa of the stomach of a patient presenting ABG. Numerous glands exhibit pseudoantral metaplasia and most of them have endocrine cells immunoreactive to the peptide ghrelin (arrows). As ghrelin-immunoreactive cells are usually not present or are rare in antral mucosa this could be used as an reliable indication of pseudoantral metaplasia

This endocrine cell hyperplasia does not seem to occur in multifocal atrophic gastritis. Therefore, the presence of endocrine cell hyperplasia in the body mucosa of patients with ABG is a strong signal indicating the autoimmune etiology of the ABG. These florid changes of the body mucosa contrast with the discrete histological findings of the antral mucosa*.* 

Atrophic Body Gastritis: A Challenge for the

for chromogranin

Presumptive Endoscopic and Histologic Diagnosis of Autoimmune Gastritis 179

The gastric hyperplastic polyps are the result of reactive hyperplasia of the foveolar epithelium in response to injury or to primary disease of the gastric mucosa, including chronic gastritis with glandular atrophy, as previously mentioned (Abraham et al, 2001). Apparently, these types of polyps almost never occur in normal gastric mucosa (Fig. 6 A).

Fig. 6. Body gastric mucosa of a patient with the diagnosis of atrophic body gastritis. Intense hyperplasia of the foveolar epithelium (A), and of endocrine cells in the *lamina propria* (B) which could origin small gastric polyps. A – H.E. staining; B- Immunoperoxidase staining

Other endoscopic abnormalities often associated with ABG are the "gastric polyps" that occur in the body and fundus of the stomach and that derive from nodular proliferation of endocrine cells in the connective tissue of the *lamina propria (*Figure 6 B). They are usually multiple lesions located only in the body and fundus of the stomach and their presence should raise the suspicion of endoscopic ABG. Fusion of the endocrine hyperplastic nodules frequently occurs and the differential diagnosis of a neuroendocrine tumor should be considered. Since these nodules are usually multiple and not uniform, sampling of some of them does not always provide sufficient information for a final conclusion about the presence of endocrine neoplasia. In conclusion, there is a clinical and pathological significance of gastric polyps in relation to ABG because of its frequent association with both hyperplastic polyps and the polyps derived from endocrine cells proliferation (endocrine hyperplastic nodules or carcinoid tumors). However, the removal of these polyps must be accompanied by biopsies from the gastric antrum and body in order to characterize the primary process responsible for the appearance of polypoid lesions. The presence of these polyps may serve as a signal to the endoscopist of the presence of underlying disease or

injury of the gastric mucosa, which is not always perceived by standard endoscopy.

The disagreement between endoscopic and histologic diagnosis for the presence of glandular atrophy of the gastric mucosa is not new and is relatively frequent (Torbenso et al, 2002). Standard endoscopy, although it can employ technical methods to study the thickness of the gastric mucosa such as stretching of the stomach wall by inflating air into the gastric

**3.4 Diagnosis of gastric mucosa atrophy: Endoscopic perspectives** 

*H. pylori* is usually negative in these patients and, regardless of its presence, the atrophic gastritis with the morphological characteristics described above should preferably be regarded as autoimmune (Capella et al, 1999; Park et al, 2010). If necessary, serum antiparietal cell and anti-intrinsic factor antibodies can be investigated to confirm this possibility. It should be noted that in a small percentage of patients with the pattern of morphological changes of the gastric mucosa typical of ABG, and clinically supposed to have autoimmune gastritis with pernicious anemia, the search for anti-parietal cells and/or anti-intrinsic factor antibodies may give negative results.

Some patients with pernicious anemia have been reported to have severe histological changes of the antral mucosa indistinguishable from those seen in the gastric body (Lewin et al, 1976). Recently, severe atrophic gastritis of the antrum and body mucosa has been reported in association with systemic autoimmune diseases. This pathological condition was not associated with *H. pylori* or endocrine cell hyperplasia and may affect a especific group of patients with a particular type of autoimmune gastritis. Hypothetically, these patients would develop the production of antibodies directed against multiple cell lines of the gastric mucosa (Jevremovic, 2006).

#### **3.3 Gastric polyps and atrophic body gastritis**

As seen in this series, the main endoscopic diagnosis of "gastric polyps" was relatively common among patients undergoing endoscopy who had a final diagnosis of ABG as the most important disease. We do not know how many of the 196 cases of ABG studied also had gastric polypoid lesions that were not described in the findings of endoscopic diagnostic by being considered irrelevant or by having been omitted. Likewise we are not aware of the frequency of cases of gastric polyps which, considered being the major and only injury, were removed by polypectomy without adequate sampling for histology of the mucosa of the gastric body and antrum. We believe that many pathologists face this reality in the laboratory routine, which may result in the omission of the diagnosis of the most important underlying gastric disease often responsible for the presence of polyps, such as ABG (Jain & R Chetty, R, 2009; Haruma et al, 1993).

Endoscopically, the term "gastric polyp" is applied to any bulge or swelling of the gastric mucosa, whether of epithelial origin or from underlying tissues. From a structural viewpoint, however, the term is being used by pathologists to designate lesions mainly consisting of epithelial proliferation that project into the lumen of the organ. However, the nomenclature of these lesions has not been fully defined. Moreover, the terminology used for polyps of the stomach and the description of their morphological structure are very similar to those used to describe polypoid lesions of the colon, although the biological behavior of gastric and colonic polyps does not always show the same evolutionary pattern. Gastric polyps can be divided into several types, many of which are still poorly understood in terms of their etiopathogenesis. The types for whose definition there is a more general consensus are: hyperplastic polyps, adenomatous polyps (adenomas), mixed polyps, fundic gland polyps, hamartomatous polyps, and retention polyps (juvenile). These different types of polyps may be associated with clinical syndromes (syndromic polyps) including the Peutz-Jeghers, Gardner, Cronkhite-Canada and Crowden syndromes. Among these various types of gastric polyps, the most common are hyperplastic non-syndromic polyps, which are also those most often associated with ABG. It should be added that the histological pattern of gastric hyperplastic polyps may be indistinguishable from that of polyps of syndromic origin (Lam-Himlinet al, 2010).

*H. pylori* is usually negative in these patients and, regardless of its presence, the atrophic gastritis with the morphological characteristics described above should preferably be regarded as autoimmune (Capella et al, 1999; Park et al, 2010). If necessary, serum antiparietal cell and anti-intrinsic factor antibodies can be investigated to confirm this possibility. It should be noted that in a small percentage of patients with the pattern of morphological changes of the gastric mucosa typical of ABG, and clinically supposed to have autoimmune gastritis with pernicious anemia, the search for anti-parietal cells and/or anti-intrinsic factor

Some patients with pernicious anemia have been reported to have severe histological changes of the antral mucosa indistinguishable from those seen in the gastric body (Lewin et al, 1976). Recently, severe atrophic gastritis of the antrum and body mucosa has been reported in association with systemic autoimmune diseases. This pathological condition was not associated with *H. pylori* or endocrine cell hyperplasia and may affect a especific group of patients with a particular type of autoimmune gastritis. Hypothetically, these patients would develop the production of antibodies directed against multiple cell lines of the gastric

As seen in this series, the main endoscopic diagnosis of "gastric polyps" was relatively common among patients undergoing endoscopy who had a final diagnosis of ABG as the most important disease. We do not know how many of the 196 cases of ABG studied also had gastric polypoid lesions that were not described in the findings of endoscopic diagnostic by being considered irrelevant or by having been omitted. Likewise we are not aware of the frequency of cases of gastric polyps which, considered being the major and only injury, were removed by polypectomy without adequate sampling for histology of the mucosa of the gastric body and antrum. We believe that many pathologists face this reality in the laboratory routine, which may result in the omission of the diagnosis of the most important underlying gastric disease often responsible for the presence of polyps, such as

Endoscopically, the term "gastric polyp" is applied to any bulge or swelling of the gastric mucosa, whether of epithelial origin or from underlying tissues. From a structural viewpoint, however, the term is being used by pathologists to designate lesions mainly consisting of epithelial proliferation that project into the lumen of the organ. However, the nomenclature of these lesions has not been fully defined. Moreover, the terminology used for polyps of the stomach and the description of their morphological structure are very similar to those used to describe polypoid lesions of the colon, although the biological behavior of gastric and colonic polyps does not always show the same evolutionary pattern. Gastric polyps can be divided into several types, many of which are still poorly understood in terms of their etiopathogenesis. The types for whose definition there is a more general consensus are: hyperplastic polyps, adenomatous polyps (adenomas), mixed polyps, fundic gland polyps, hamartomatous polyps, and retention polyps (juvenile). These different types of polyps may be associated with clinical syndromes (syndromic polyps) including the Peutz-Jeghers, Gardner, Cronkhite-Canada and Crowden syndromes. Among these various types of gastric polyps, the most common are hyperplastic non-syndromic polyps, which are also those most often associated with ABG. It should be added that the histological pattern of gastric hyperplastic polyps may be indistinguishable from that of polyps of syndromic

antibodies may give negative results.

**3.3 Gastric polyps and atrophic body gastritis** 

ABG (Jain & R Chetty, R, 2009; Haruma et al, 1993).

mucosa (Jevremovic, 2006).

origin (Lam-Himlinet al, 2010).

The gastric hyperplastic polyps are the result of reactive hyperplasia of the foveolar epithelium in response to injury or to primary disease of the gastric mucosa, including chronic gastritis with glandular atrophy, as previously mentioned (Abraham et al, 2001). Apparently, these types of polyps almost never occur in normal gastric mucosa (Fig. 6 A).

Fig. 6. Body gastric mucosa of a patient with the diagnosis of atrophic body gastritis. Intense hyperplasia of the foveolar epithelium (A), and of endocrine cells in the *lamina propria* (B) which could origin small gastric polyps. A – H.E. staining; B- Immunoperoxidase staining for chromogranin

Other endoscopic abnormalities often associated with ABG are the "gastric polyps" that occur in the body and fundus of the stomach and that derive from nodular proliferation of endocrine cells in the connective tissue of the *lamina propria (*Figure 6 B). They are usually multiple lesions located only in the body and fundus of the stomach and their presence should raise the suspicion of endoscopic ABG. Fusion of the endocrine hyperplastic nodules frequently occurs and the differential diagnosis of a neuroendocrine tumor should be considered. Since these nodules are usually multiple and not uniform, sampling of some of them does not always provide sufficient information for a final conclusion about the presence of endocrine neoplasia. In conclusion, there is a clinical and pathological significance of gastric polyps in relation to ABG because of its frequent association with both hyperplastic polyps and the polyps derived from endocrine cells proliferation (endocrine hyperplastic nodules or carcinoid tumors). However, the removal of these polyps must be accompanied by biopsies from the gastric antrum and body in order to characterize the primary process responsible for the appearance of polypoid lesions. The presence of these polyps may serve as a signal to the endoscopist of the presence of underlying disease or injury of the gastric mucosa, which is not always perceived by standard endoscopy.

#### **3.4 Diagnosis of gastric mucosa atrophy: Endoscopic perspectives**

The disagreement between endoscopic and histologic diagnosis for the presence of glandular atrophy of the gastric mucosa is not new and is relatively frequent (Torbenso et al, 2002). Standard endoscopy, although it can employ technical methods to study the thickness of the gastric mucosa such as stretching of the stomach wall by inflating air into the gastric

Atrophic Body Gastritis: A Challenge for the

**5. Acknowledgment** 

technical assistance.

**6. References** 

Presumptive Endoscopic and Histologic Diagnosis of Autoimmune Gastritis 181

with histology for the evaluation of patients with different degrees of gastric mucosa

This work was supported by grants from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Brazil. The authors thank Ms. Luciene S. P. Faria for

Abraham, S.C., Singh, V.K., Yardley, J.K. & Wu, T.T. (2001). Hyperplastic polyps of the

Anagnostoupoulos, G.K., Yao, K. et al. (2007). High-resolution magnification endoscopy can

Canto, M.I. (2010). Endomicroscopy of Barrett´s Esophagus. Gastroenterol Clinic North Am,

Capella, C., Fiocca, R., Cornaggia, M., Rindi, G., Moratti, R. & Solcia, E. (1999). Autoimmune

Carmel, R. (1996). Prevalence of undiagnosed pernicious anaemia in the elderly. Arch Intern

Correa, P. (1984). Chronic gastritis as a cancer precursor. Scand J Gastroenterol, Vol.19, pp

Correa, P. (1992). Human gastric carcinogenesis: a multistep and multifactorial process –

Dixon, M.F., Genta, R.M, Yardley, J.H. et al. (1996). Classification and grading of gastritis.

Dixon, M.F., Genta, R.M., Yardley, J.H. & Correa, P. (1999). Classification and grading of

Eshmuratov, A., Nah, J.C. et al. (2010). The correlation of endoscopic and histological

Genta, R.M. (1996). Recognizing atrophy: another step toward a classification of gastritis.

Genta, R.M. (1997). Helicobacter pylori, inflammation, mucosal damage and apoptosis:

Goetz, M. & Kiesslich, R. (2010). Advances of endomicroscopy for gastrointestinal

Gastritis, Houston 1994. Am J Surg Pathol, Vol.20, pp 1161-1181.

diagnosis of gastric atrophy. Dig Dis Sci, Vol.55, pp 1364-1375.

stomach: associations with histologic patterns of gastritis and gastric atrophy. *Am J* 

reliably identify normal gastric mucosa, Helicobacter pylori associated gastritis and

gastritis, In: Gastritis, Graham, D.Y, Genta, R.M. & Dixon, M.F., pp 79-96, Lippincott Willians & Wilkins, ISBN 0-397-51675-4, Philadelphia, New York,

First American Society Award Lecture on Cancer Epidemiology and Prevention.

The updated Sydney System. International Workshop on the Histopathology of

gastritis. The updated Sydney System. In: Gastritis, Graham, D.Y, Genta, R.M. & Dixon, M.F., pp 35-49, Lippincott Willians & Wilkins, ISBN 0-397-51675-4,

pathogenesis and definition of gastric atrophy. Gastroenterology, Vol.113, No.6, pp

physiology and diseases. Am J Physiol Gastrointest Liver Physiol, Vol.298, No.6, pp

atrophy and consequently with chronic gastritis of autoimmune etiology.

*Surg Pathol*, Vol.25, No.4, pp 500-507.

Med, Vol.156. No.10, pp 1097-1100.

Cancer Res, Vol.52, pp 6735-6740.

Philadelphia, New York, London.

Am J Surg Pathol, Vol.20, No.1, pp 23-30.

Vol.39, No.4, pp 759-769.

London.

131-136.

51-55.

G797-806.

gastric atrophy. Endoscopy, Vol.39, pp 202-207.

cavity, continues to have a low accuracy index. The degree of distension of the stomach wall and visualization of the submucosal vascular network depends primarily on the professional skill and experience of the examiner, as well as on the availability of good quality equipment. Even at centers specialized in gastrointestinal endoscopy, the endoscopic-histologic correlation is weak, with sensitivity and specificity of about 40 to 60% (Eshmuratov, et al. 2010).

Therefore, since the histological method is also not reliable, at least for cases of mild or moderate atrophy, its use in combination with endoscopy continues to rest on quicksand. Opening good perspectives for the near future, the endoscopic method has progressed with the description of new visualization techniques to amplify the resolution and definition of gastrointestinal mucosa details. Thus, endoscopic techniques involving magnification with high resolution have been reported to be considerably more reliable than standard endoscopy to identify normal gastric mucosa, chronic gastritis and gastric atrophy (Anagnostopoulos et al. 2007). The progress of endoscopic techniques and the availability of high-resolution confocal laser endomicroscopy are now starting to gain firmer ground in the detection of minute lesions of the gastrointestinal mucosa, among them the different degrees of gastric mucosa atrophy (Li, CQ and Li, YQ, 2010; Goetz, M. and Kiesslich, R. 2010; Canto, 2010).
