Is it ever possible to make the unequivocal histological diagnosis of Barrett's mucosa without goblet cells?
R.H. Riddell (Hamilton)
In this question the key words are "unequivocal", " histological", and "diagnosis". There are numerous clinical definitions of Barrett's esophagus, based primarily on endoscopic appearances, but, unless this is itself used as a gold standard, its implications are unclear, for in the absence of intestinal metaplasia there is likely no additional cancer risk. Indeed, without the cancer risk Barrett's esophagus would simply be an esophageal peptic ulcer.
If one looks at the histological subtypes of Barrett's mucosa as originally described by Paull they are described as fundic, junctional and specialized . Bear in mind that this study was carried out using suction biopsies with manometric evaluation of the site of the lower esophageal sphincter (LES). Bear in mind that the LES itself is about two cm in length and that the squamocolumnar junction can be anywhere within the LES, and that because patients with Barrett's esophagus have considerable reflux, their LES pressures are low, and occasionally absent. Hiatal hernias are also invariably present in patients with Barrett's esophagus so that it is easily possible to biopsy gastric fundic mucosa in some patients without going beneath the diaphragm. Finally, in some normal patients the transition from gastric fundic mucosa to esophageal squamous mucosa is very abrupt with very little cardiac mucosa at all.
Bearing this in mind we have to ask whether the fundic mucosa described by Paull et al. as being the most distal type found in Barrett mucosa really could have been from the gastric fundus, and whether the next most proximal (cardiac) could indeed have represented normal cardiac mucosa in some patients. In patients without intestinal metaplasia in the stomach, this leaves only the specialized mucosa that is characterized by the presence of intestinal metaplasia as the only distinctive mucosa that can readily be distinguished from normal gastric mucosa. From this viewpoint then the presence of goblet cells, by which intestinal metaplasia or "specialized" mucosa is identified, becomes the sine qua non for the diagnosis of BE.
Nevertheless, it is important to ask the question the other way round in addition, namely, in quite unequivocal BE, is it possible to see mucosa that is indistinguishable from gastric fundic or junctional (cardiac) mucosa?
Do parietal cells exist in unequivocal BE, and if so are they distinguishable from gastric fundic mucosa? The literature is spectacularly unconvincing over this point, but I can personally vouch for having seen parietal cells in resected specimens of Barrett's esophagus associated with underlying esophageal glands. However these have been in small islands, but have been accompanied by chief cells, thereby resembling gastric fundic mucosa. However, these have not been in the luxuriant thick glands seen in normal fundic glands, but fit comfortably in the thickness of usual Barrett mucosa. They therefore resemble better the occasional specialized cells seen in the cardiac transition zone rather than true fundic gland mucosa.
Does junctional (cardiac type) mucosa exist in Barrett mucosa, and if so can it be distinguished from normal cardiac (junctional) mucosa? There is little doubt that this type of mucosa can exist in unequivocal Barrett mucosa; indeed in patients undergoing multiple biopsies (or resection) in whom it is clear that biopsies are within Barrett's mucosa, this is probably the most common mucosal type after metaplastic mucosa. But can it be reliably distinguished from normal cardiac mucosa? My suspicion is that one often has a hunch that one is really in Barrett's mucosa because of the villiform superficial appearance of the mucosa and the architectural abnormality presumably resulting from previous epithelial destruction and pit regeneration. But no study has yet examined this as a diagnostic criterion.
The final question that needs to be asked is whether, if an alcian blue stain is carried out and alcianophilic mucus cells are found, whether this is diagnostic of BE. Although it is tempting to speculate that at the cardia in particular that these cells might in some patients be the precursor lesion of goblet cells, there have to date been no studies from which this conclusion can be drawn, and one that suggests this is not the case .
While this question asks whether it is possible to make an unequivocal histological diagnosis of Barrett's mucosa in the absence of goblet cells, to which the answer must be NO, it does not ask the sensitivity or specificity of such a finding. Can goblet cells be found in the vicinity of the Z-line in patients with widespread intestinal metaplasia of the stomach as part of that process? Preliminary data suggests that it can .
Does Barrett's esophagus occur in children in the absence of goblet cells? This also seems to be the case. In the latter instance, how does one make the diagnosis of Barrett's esophagus in the absence of goblet cells? In fact, one can only state that the mucosa is glandular, usually of cardiac type, and that if the endoscopist is really convinced that it came from far above the physiological sphincter, that it is consistent with (but not unequivocally diagnostic of) the pediatric (non-metaplastic) form of the disease [4, 5].
Thus , in this situation where it is critical to know whether the patient really has BE in view of the long term cancer risk, in the absence of goblet cells one is unable to offer an unequivocal diagnosis [6, 7]. It is therefore impossible to unequivocally make the diagnosis of Barrett's mucosa in the absence of goblet cells.