Portal de Periódicos da CAPES

The diagnosis of Barrett's esophagus: goblets, goblets, goblets

1996; Elsevier BV; Volume: 44; Issue: 1 Linguagem: Inglês

10.1016/s0016-5107(96)70239-0

1097-6779

Wilfred M. Weinstein, Andrew Ippoliti,

Gastrointestinal Tumor Research and Treatment

In this issue of the Journal, Canto et al.1Canto MI Setrakian S Petras R Blades E Chak A Sivak M. Methylene blue selectively stains intestinal metaplasia in Barrett's esophagus.Gastrointest Endosc. 1996; 44: 1-7Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar describe a technique for highlighting Barrett's esophagus at endoscopy. Barrett's esophagus would be no more than a curiosity except (1) it is associated with an increased risk of adenocarcinoma of the esophagus, (2) adenocarcinoma of the esophagus is increasing, and (3) cardia cancer, which is also increasing dramatically with epidemiologic characteristics similar to Barrett's carcinomas2Chow WH Finkle WD McLaughlin JK Frankl H Ziel HK Fraumeni Jr., JF The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia.JAMA. 1995; 274: 474-477Crossref PubMed Scopus (346) Google Scholar probably arises from short segments of Barrett's esophagus or from intestinal metaplasia at a normally located squamocolumnar junction. Our prerequisite for the diagnosis of Barrett's esophagus is the presence of intestinal-type goblet cells in at least one biopsy from the lower esophagus. The length of the metaplasia can be as short as a few millimeters, i.e., any amount of columnar mucosa in the lower esophagus that has histologic evidence of goblet cells. The goblet cell component of the metaplastic epithelium is believed by virtually all in the field to represent the predisposing cell type for adenocarcinoma. The importance of the goblet cell component is a relatively recent development in Barrett's esophagus. In 1976 Paull et al.3Paull A Trier JS Dalton MD Camp RC Loeb P Goyal RK. The histologic spectrum of Barrett's esophagus.N Engl J Med. 1976; 295: 476-480Crossref PubMed Scopus (695) Google Scholar described the intestinal goblet cell (specialized epithelium) as part of the mosaic of Barrett's along with cardiac and oxyntic (fundic gland) mucosa. Their recognition of specialized epithelium in Barrett's esophagus was very important because many had considered Barrett's to represent gastric tissue in the esophagus. The diagnosis of Barrett's esophagus is obvious when a large segment, e.g., 5 cm of lower esophagus, has the typical pink color with a distinct proximally relocated Z-line. However, histology is absolutely essential to make a diagnosis of Barrett's esophagus when only short pink tongues are present in the lower esophagus. Here the differential diagnosis is between a short segment of Barrett's esophagus and a normal but eccentric Z-line. The combined hematoxylin and eosin-alcian blue pH 2.5 (H & E-AB, pH 2.5) stain is invaluable for the histologic definition of the intestinalized mucosa of Barrett's esophagus. The mucins it stains are seen both in the small and large intestine. The H & E-AB, pH 2.5 stain also helps to avoid overdiagnosis of Barrett's esophagus, which results in patients carrying the worry about increased cancer risk for the rest of their lives. Some otherwise normal gastric cardia cells look bloated with the barrel shape of the goblet cell. The H & E-AB pH 2.5 stain will differentiate these bloated gastric cardia cells from goblet cells. In biopsies from the lower esophagus about half will be identical to gastric mucosa, either the oxyntic (fundic gland) or the cardiac gland type. Above the 2 cm zone in adults it is rare to find mucosa that resembles gastric tissue in all respects.4Lewin KJ Riddell RH Weinstein WM. Inflammatory disorders of the esophagus: reflux and nonreflux types.in: Gastrointestinal pathology and its clinical implications. Igaku-Shoin, New York1992: 401-433Google Scholar However, in the surface (nonpit) epithelium of Barrett's esophagus at all levels there are often smaller numbers of gastric columnar-type cells that do not stain positively with alcian blue (Table 1). Even more common than goblet cells are cells that look just like gastric columnar cells with conventional hematoxylin and eosin stains but, unlike them, stain positively with alcian blue (Table 1). We believe that they represent a transitional type of metaplastic cell intermediate between the gastric columnar cell and the goblet cell. There is no accepted name for them. Our colloquial in-house name, because of its jazzy connotation is the “columnar blues.” Our preference for a more proper name is transitional cells. In the stomach these columnar blues occur commonly as part of the picture of intestinal metaplasia, and have been termed by some incomplete intestinal metaplasia.5Offner FA Weinstein WM Lewin KJ. Columnar non-goblet cell metaplasia: a neglected feature in the diagnosis of Barrett's esophagus.Hum Pathol. 1996; (in press)PubMed Google ScholarTable 1Types of mucous cells in the surface epithelium of patients with Barrett's esophagus*Cell typeAppearance result with H & E-stainStaining result with H & E-alcian blue pH 2.5Goblet cell†Goblet shapedPositiveTransitional (columnar blues)ColumnarPositiveGastricColumnarNegative*In addition to mucous cells, other nonmucin or nucin-depleted cells may be seen along the surface (nonpit) epithelium. They include enterocyte-type cells with poorly developed brush borders, dysplastic, and regenerative epithelium. †The diagnosis of Barrett's esophagus is only made if goblet cells are present in some of the esophageal biopsies. Open table in a new tab At the outset Barrett did not appreciate the significance of what he was describing.6Barrett NR. Chronic peptic ulcer of the oesophagus and oesophagitis.Brit J Surg. 1950; 38: 175-182Crossref PubMed Scopus (702) Google Scholar Even after the description of goblet cells in Barrett's,3Paull A Trier JS Dalton MD Camp RC Loeb P Goyal RK. The histologic spectrum of Barrett's esophagus.N Engl J Med. 1976; 295: 476-480Crossref PubMed Scopus (695) Google Scholar pathologists still diagnosed Barrett's in their absence as “junctional type (cardiac gland) Barrett's” and “fundic type Barrett's.” Even today there are those who would divide Barrett's into a goblet and a nongoblet type.7Spechler SJ Goyal RK. The columnar-lined esophagus, intestinal metaplasia, and Norman Barrett.Gastroenterology. 1996; 110: 614-621Abstract Full Text PDF PubMed Scopus (440) Google Scholar We believe that there has been sufficient progress in our knowledge to make the definition more focused because of (1) better technology, i.e., better endoscopic technology and directed biopsy versus blind biopsy with manometric localization, and (2) data showing that the goblet cell is the predisposing lesion for most if not all Barrett's esophagus–associated adenocarcinomas.8Haggitt RC. Barrett's esophagus, dysplasia, and adenocarcinoma.Hum Pathol. 1994; 25: 982-993Abstract Full Text PDF PubMed Scopus (611) Google Scholar We believe that whenever possible the best names for a disease are those that have a bearing on management and/or prognosis. We also believe that it is justified to change the definition of a disorder based on new knowledge and yet keep the original name. The alternative is to add confusing subnames merely out of historical propriety without necessarily any relevance to the patient or to the mechanism of disease. The precedent has been set in medicine on numerous occasions where a person is immortalized with the name of a disease but where that individual did not appreciate the significance of what was being described. Mr. Barrett was a case in point.6Barrett NR. Chronic peptic ulcer of the oesophagus and oesophagitis.Brit J Surg. 1950; 38: 175-182Crossref PubMed Scopus (702) Google Scholar, 9Allison PR Johnstone AS. The oesophagus lined with gastric mucous membrane.Thorax. 1953; 8: 87-101Crossref PubMed Scopus (343) Google Scholar Landmarks. Traditionally, the diagnosis of Barrett's esophagus is suggested whenever pink columnar epithelium intrudes into the esophagus. This may be in the form of tongues of columnar epithelium or as a circumferential zone of pink mucosa with a proximally relocated squamocolumnar junction (Z-line). The extent of columnar change should be given as the distance between the gastroesophageal junction and the most proximal tip of the pink tongue or the level of the Z-line if there is circumferential involvement. The pathologist also needs these landmarks as well as the biopsy locations as cm from incisors to know the exact sites of the biopsies. The estimates of the location of the gastroesophageal junction and other landmarks at endoscopy have an element of subjectivity, especially in certain circumstances. For example the gastroesophageal junction may be misidentified in the presence of a hiatal hernia (discussed subsequently), and erosions can be confused with columnar islands. The endoscopist's localization of the gastroesophageal junction may vary substantially in some patients, but in our experience usually not more than 2 cm in the same patient. In the presence of marked ulcerations or erosions in continuity with the lower esophageal sphincter (LES) it is often impossible to know if Barrett's esophagus is present. We advise not even trying, but rather making that determination with endoscopy after the lesions have healed. In our referral area the most common source of misdiagnosis arises from the confusion of the lower end of a hiatal hernia with the gastroesophageal junction. The LES in gastroesophageal reflux disease is often lax and the diaphragmatic pinch at the lower end of a hiatal hernia may be confused with the gastroesophageal junction. The tip-off that this is so is to partially deflate the lower esophageal region and the telltale folds of a hiatal hernia will appear. If the sphincter is not evident, then its location is taken to be where the hernia folds end.10McClave SA Boyce Jr, HW Gottfried MR. Early diagnosis of columnar-lined esophagus: a new endoscopic diagnostic criterion.Gastrointest Endosc. 1987; 33: 413-416Abstract Full Text PDF PubMed Scopus (235) Google Scholar The diagnosis of Barrett's esophagus would not be made in adults with this landmark misidentification alone if not for the fact that some pathologists (fortunately fewer each year) still diagnose Barrett's esophagus even if only oxyntic gland (fundic gland) or cardiac gland mucosa are seen. Canto et al.1Canto MI Setrakian S Petras R Blades E Chak A Sivak M. Methylene blue selectively stains intestinal metaplasia in Barrett's esophagus.Gastrointest Endosc. 1996; 44: 1-7Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar have used methylene blue staining to detect specialized epithelium in Barrett's esophagus. Theoretically, the finding of positive staining is the endoscopic equivalent of histology with positive H&E-alcian blue staining. In fact, they found that methylene blue staining had a 95% sensitivity and 97% specificity for the biopsy finding of specialized epithelium. False-positives resulted from staining in a hiatal hernia (see previous paragraph), while false-negatives were due to excessive washing after application of the stain. Methylene blue stains ulcers and erosions, so it will not distinguish those lesions from noninflamed Barrett's mucosa. The usefulness in the evaluation of short segments of intestinalized mucosa or intestinalization right at a normally located Z-line merits future evaluation. The proximal end of Barrett's mucosa is readily identified at endoscopy, but the distal margin is generally more difficult to discern. This is particularly an issue in the short segment (e.g., ≤2 cm) versus eccentric Z-line cases. The notion that the Z-line is straddled by the gastroesophageal junction applies only to healthy subjects. Csendes et al.,11Csendes A Maluenda F Braghetto I Csendes P Henriquez A Quesada MS. Location of the lower oesophageal sphincter and the squamous columnar mucosal junction in 109 healthy controls and 778 patients with different degrees of endoscopic oesophagitis.Gut. 1993; 34: 21-27Crossref PubMed Scopus (123) Google Scholar in an ambitious undertaking, recorded the level of the Z-line at endoscopy and the location of the lower esophageal sphincter by motility, via the mouth, in 778 patients. The Z-line corresponded to the midportion of the lower sphincter, high-pressure zone only in nonrefluxing patients. In patients with reflux symptoms without esophagitis, the Z-line was 2.7 cm above the distal end of the lower esophageal sphincter. In the case of biopsy of either suspected or proven Barrett's, we believe it is essential to take the first biopsy specimens 1 to 2 cm below the lower esophageal sphincter, then at the LES, and at 2 cm intervals proximally. The biopsies below the LES prove (with the finding of oxyntic mucosa) that one's estimate of the location of the LES is correct. If it is not, the biopsies below the apparent LES will contain Barrett's mucosa. In that event, at the next endoscopic biopsy surveillance assessment the reference biopsies should be taken further distally, e.g., at 3, 2, and 1 cm below the LES. One has to acknowledge that when we look at the lower esophagus at endoscopy and see the typical ring-like indentation that we call the lower LES we do not know exactly how this corresponds to the manometric sphincter. As discussed already, for the management of Barrett's esophagus it does not matter. Not often appreciated is that manometrists may be as imprecise with their localization of the LES as endoscopists. One reason is that the distal end of the lower esophageal sphincter is manometrically asymmetrical and the location of the high-pressure zone varies according to the orientation of the recording assembly. When a four-radial port assembly is used, for example, an average has to be taken for both the length of the sphincter and its pressure. Since the mission for the endoscopist in Barrett's esophagus is to rule out dysplasia and carcinoma, the most important issue is to sample all of the intestinalized mucosa. By taking the first biopsy specimens 1 to 2 cm below the LES one bypasses all the vagaries cited above and samples the metaplastic epithelium. When an endoscopist sees short (≤2 cm) pink tongues in the lower esophagus and biopsies them with the question of Barrett's esophagus, three different kinds of mucosa might be encountered. Two of them, i.e., oxyntic gland and cardiac (or a blend of both), prompt us to call that an eccentric Z-line and not Barrett's esophagus. In the endoscopic and biopsy evaluation of more than 250 cases of Barrett's esophagus we have seen nonintestinalized tongues of mucosa extending more than 2 cm into the lower esophagus in less than 1% of cases. Therefore we believe that those who claim to see this regularly are either confusing the lower esophageal sphincter region with the lower end of a hiatal hernia or their pathologists are overlooking isolated goblet cells because they do not use the hematoxylin and eosin-alcian blue pH 2.5 stain. As indicated, we believe that any length of pink mucosa in the lower esophagus that contains intestinalized mucosa on biopsy qualifies for a diagnosis of Barrett's esophagus—even 5 or 6 mm. On the other hand, in research studies of Barrett's esophagus it is often justified to require at least 2 cm of intestinalized mucosa in the lower esophagus. One reason is so others will accept that bona fide Barrett's esophagus was being studied. A second reason, in the case of studies of regression of Barrett's esophagus, is to provide a sufficient baseline for analysis. One 3 cm rule we do not agree with was in a study where tongues of intestinalized pink mucosa up to 3 cm above the LES were still considered intestinalized cardia,12Spechler SJ Zeroogian JM Antonioli DA Wang HH Goyal RK. Prevalence of metaplasia at the gastro-oesophageal junction.Lancet. 1994; 344: 1533-1536Abstract PubMed Scopus (561) Google Scholar whereas we would call those unequivocal Barrett's esophagus with 2 or 3 cm of involvement and would enroll the patient in endoscopic biopsy surveillance. The finding of intestinal metaplasia at a normally located Z-line in patients with gastroesophageal reflux disease was reported more than 10 years ago13Peuchmaur M Potet F Goldfain D. Mucin histochemistry of the columnar epithelium of the oesophagus (Barrett's oesophagus): a prospective biopsy study.J Clin Pathol. 1984; 37: 607-610Crossref PubMed Scopus (76) Google Scholar and should really come as no surprise. However, only recently has it captured the attention of workers in this field because of the likelihood that cardia cancer, which has many similarities to Barrett's-associated cancers,2Chow WH Finkle WD McLaughlin JK Frankl H Ziel HK Fraumeni Jr., JF The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia.JAMA. 1995; 274: 474-477Crossref PubMed Scopus (346) Google Scholar arises from short segments of Barrett's esophagus or from intestinal metaplasia at a normally located Z-line. Also it should come as no surprise (from a recent preliminary study) that inflammation of the gastric cardia (carditis) might be a more sensitive marker of mucosal injury from gastroesophageal reflux disease than biopsies away from the battle lines in the squamous epithelium.14Clark GWB Ireland AP Chandrasoma P DeMeester TR Peters JH Bremner CG. Inflammation and metaplasia in the transitional epithelium of the gastroesophageal junction: a new marker for gastroesophageal reflux disease.Gastroenterology. 1994; 106 ([abstract]): A63Google Scholar A hypothetical scheme for the development of intestinal metaplasia of the gastric cardia is shown in Fig. 1. Gastroesophageal reflux disease is associated with inflammation right at the squamocolumnar junction. The inflammation is highly localized in the cardia and it is this carditis that evolves in some patients to intestinal metaplasia of the cardia. The intestinal metaplasia may develop from the transitional cells (columnar blues) described previously or directly from the carditis (Fig. 1). The carditis is potentially reversible and perhaps even the transitional columnar blues may at times also revert to a normal gastric mucous cell. We believe that until we know more about cancer risk we should not label intestinal metaplasia of the cardia as Barrett's esophagus. One could argue that what we call very-short-segment Barrett's, e.g., 4 or 5 mm of pink intestinalized tongues in the lower esophagus, is really intestinal metaplasia of an eccentric Z-line. That is possible, but one has to pick a definition at some point. In the case of Barrett's esophagus the definition that makes most clinical sense is one where the endoscopist judges that pink mucosa has intruded into the lower esophagus and where the histology reveals goblet cells. From there follows the assumption that this carries an increased cancer risk and indicates a need for periodic endoscopic biopsy surveillance if early neoplasia is to be detected. The biologic significance of the goblet cell itself may not be as clear cut as we formerly thought. New indirect evidence suggests that localized metaplasia at the cardia, and even for short distances higher, may not have the same connotation for cancer risk as in long-segment Barrett's. The basis for this speculation is that in recent preliminary reports of short-segment Barrett's esophagus, women seem to be markedly overrepresented15Cameron AJ Kamath PS Carpenter HC. Barrett's esophagus: the prevalence of short and long segments in reflux patients.Gastroenterology. 1995; 108 ([abstract]): 65Abstract Full Text PDF PubMed Google Scholar, 16Abo SR Stevens PD Abedi M et al.Prevalence of short-segment Barrett's epithelium in patients with gastroesophageal reflux disease.Gastroenterology. 1995; 108 ([abstract]): 43Abstract Full Text PDF Google Scholar relative to their known risk of Barrett's esophagus, esophageal adenocarcinoma, and cardia cancer. Thus, it is likely that the gender and race factors in GE reflux disease for the risk of Barrett's esophagus and cancer seem to occur as postgoblet cell events. The histology of the gastric cardia and its clinical implications raise numerous practical questions that need well-designed studies to answer them. Is carditis a wear and tear phenomenon analogous to the regenerative changes in squamous mucosa in the distal 2.5 cm of the esophagus of normal individuals?17Weinstein WM Bogoch ER Bowes KL. The normal human esophageal mucosa: a histological reappraisal.Gastroenterology. 1975; 68: 40-44PubMed Scopus (162) Google Scholar Should we biopsy the normally located Z-line of all patients with gastroesophageal reflux disease, just white males, or none? When intestinal metaplasia of the gastric cardia is found in gastroesophageal reflux disease, does that justify endoscopic biopsy surveillance at the same frequency as Barrett's esophagus? 1.Provide the patient record and the pathologist with the location of the gastroesophageal junction, the proximally relocated Z-line, and the location of biopsies.2.Take the first biopsy specimens 1 to 2 cm below where you believe the lower esophageal sphincter is located. This helps guarantee that all the metaplasia is sampled in the event that the estimate of the location of the LES is too proximal.3.In suspected Barrett's esophagus (e.g., short tongues), ask the pathologist to do the combined hematoxylin and eosin-alcian blue pH 2.5 stain. This will help prevent overdiagnosis as well as facilitate a positive diagnosis when goblet cells are rare.4.Restrict the diagnosis of Barrett's esophagus only to those cases that have at least one biopsy specimen containing some goblet cells.5.When goblet cells are found at a normally located Z-line, we would diagnose that as intestinal metaplasia of the gastric cardia rather than Barrett's esophagus.