Are there histological criteria that are capable of distinguishing high-grade dysplasia with biologic potential of invasion from high-grade dysplasia without invasive potential?
R.H. Riddell (Hamilton)
What this question really asks is whether there is any evidence that the traditional gold standard of histology can be replaced using other techniques. The short answer to this question is, at present, NO. However, as data accumulate which looks more closely at, and uses as its gold standard, the morphological spectrum of changes from nondysplastic through to high-grade dysplasia, a series of changes occurs in the following:
1. An increase in the amount of epithelial nuclear DNA (aneuploidy).
2. An increase in the proportion of cells entering the G1, S and G2 part of the cell cycle, resulting in an increased mitotic rate.
3. The expression of oncogenes and their products.
4. Chromosomal abnormalities.
5. Mucin changes.
6. Nuclear organizer regions.
All of these are discussed elsewhere in this book and what follows is a brief summary. The bottom line is that while it is clear that a sequence of changes tends to occur in any patient as this spectrum is travelled, and that there are overall broad correlations, the single best predictor of invasion remains high-grade dysplasia. Let us examine some of these.
There is evidence that aneuploidy precedes or accompanies both dysplasia and invasion , although the total number of cancers and background mucosa examined to date is relatively small. However, measurement of aneuploidy is dependent on a flow cytometer, or equivalent, to measure the amount of nuclear DNA and the ability to take biopsies of adequate size with sufficient tissue to carry out the procedure reliably. The easiest way to achieve this is using large (e.g., 9 mm open) biopsy forceps for both histology and flow cytometry. However, cell image analysis can be carried out on thick paraffin embedded sections, but counts many fewer cells. If available, this procedure has several implications. The first is that provided multiple biopsies are evaluated, then patients with no aneuploidy are very unlikely to have dysplasia or invasive carcinoma. The real question is whether, because of this, aneuploidy can be used to stratify patients into those without aneuploidy, who could be screened much less frequently than those with aneuploidy. For example, every 2 or even 3 years compared with those with aneuploidy who might be endoscoped annually or even more frequently if aneuploidy is widespread. To date no studies are known that have actually used aneuploidy in this way. Interestingly not all tumors in Barrett's esophagus (BE) are aneuploid; there has been a modest number of cancers not arising on a background of aneuploidy ; if confirmed this policy will need to be revised. There appears to be no role for aneuploidy in the decision for surgical resection.
Proportion of cells entering the G1, S and G2 part of the cell cycle
While it is difficult to measure each part of the cell cycle independently, the availability of flow cytometry, which allows measurement of the proportion of cells in S-phase and G2M, and antibodies such as Ki-67 which allows assessment of cells not in the G0 part of the cell cycle. One study assessing all of these parameters in the same patient cohort found that, compared with normal controls from the gastric fundus and cardia, patients with Barrett's esophagus had an increased proportion of cells in the non-G0 part of the cell cycle . Increased S-phase was particularly found in those patients with higher grades of dysplasia or carcinoma and aneuploidy. They concluded that neoplastic progression in BE was associated with at least three types of cell cycle abnormalities, including mobilization from G0 to G1, loss of control of the G1/S phase transition and accumulation of cells in G2.
A variety of oncogenes or their products have been examined in BE. The question is whether any of these are markers of increased risk of having or developing invasive carcinoma, independent of other markers such as dysplasia or even aneuploidy. Many of these are found in nondysplastic BE with a variable increase as dysplasia and cancer are sampled. The frequency varies markedly depending on the method used. Follow-up studies have shown the following changes:
p53. This is a chromosome 17 gene product that is not detectable in the normal or if there is an allelic deletion, however, mis-sense mutations are associated with a product having an increased half-life which is therefore detectable immunocytochemi-cally. It is variably (5-75%) present in BE without dysplasia and in about the same number of patients with dysplasia or carcinoma [4,5]. For example, it was found in 50% of carcinomas in one series  and 67% in another , In some studies there was a correlation with the degree of dysplasia .
Ras. This was found in 40% of nondysplastic mucosa and 56% of high-grade dysplasia .
c-erb 2. This tends to be present in intestinal cancers and, in the small proportion of patients examined, tends to portend a poor prognosis [10-12],
There currently appears to be little role for molecular markers in the decision tree of patients with BE at risk for carcinoma.
A variety of deletions have been described in Barrett's cancers including p53 (v.s.), the adenomatous polyposis gene (ARC), the MCC, Retinoblastoma and DCC genes [4,13-17]. Deletions of the Y chromosome have also been reported .
BE frequently stains with acidic (intestinal) type mucins and sulfated (colonic) mucins are frequently present, particularly in patients with carcinoma [19-21]. Consequently, related lectins, including peanut and sialosyl-Tn, are also frequently present . However, these parameters have neither the sensitivity nor specificity to be of value in determining patients at high risk of developing carcinoma.
Nuclear organizer regions
These are genes that code for ribosomal RNA and are located on the achrosomal chromosomes 13-15, 21 and 22, and are demonstrated by a silver staining technique. In Barrett's mucosa, their number is increased in dysplastic mucosa, but there is no distinction between high- and low-grade dysplasia, while they are also seen in reactive changes. This technique is therefore unlikely to be of value .
1. Reid BJ, Blount PL, Rubin CE, Levine DS, Haggitt RC, Rabinovilch PS. Row cytometric and histological progression to malignancy in Barrett's esophagus: prospective endoscopic surveillance of a cohort. Gastroenterology 1992; 102:1212-1219.
2. Menke-Pluymers MBE, Hop WCJ, Mulder AH, Tilanus HW. The role of DNA-ploidy, stage and grade as prognostic factors in the survival of patients with Barrett's adenocarcinoma. Gastroenterology 1993;104:A428.
18. Krishadath KK, Tilanus HW, Mulder AH, van Dekken H. Detection of numerical chromosomal aberrations in Barrett's adenocarcinoma and Barrett's esophagus by non-isotopic in situ hybridization to tumour sections. Gastroenterology 1993;104:A1045.
21. Itzkowitz SH, Kahn E, Auerbach M, Stiel L, Gerardi F, McKinley M. Sialosyl-Tn (STn) antigen: a potential marker of malignant transformation in Barrett's esophagus. Gastroenterology 1993;104: A412.
S.R. Hamilton (Baltimore)
Are there histological criteria which are capable of distinguishing high-grade dysplasia with the biologic potential of invasion from high-grade dysplasia without invasive potential?. The answer to this question in a word, is "NO". Dysplasia is unequivocably neoplastic epithelial proliferation , i.e., intraepithelial neoplasia, and is recognized in Barrett's mucosa on the basis of histopathologic abnormalities of mucosal architecture, epithelial morphology and cytology [2,3]. The continuous spectrum of dysplasia is categorized artificially into discrete grades to permit the pathologist to communicate the severity of the abnormalities. Grading is important in the management of patients because the more severe the dysplasia the higher the likelihood of synchronous and metachromous invasive adenocarcinoma. However, because sampling plays a role in the apparent relationship between dysplasia and adenocarcinoma, especially in biopsy specimens, the precise relationship in an individual patient is often subject to uncertainty. For example, endoscopic biopsies may be too superficial to include areas of adenocarcinoma in the submucosa and even in the mucosa itself an area of cancer may not be sampled. Columnar epithelial dysplasia in Barrett's mucosa may differ morphologically from adenocarcinoma only in lacking identifiable invasion. Furthermore, the neoplastic process does not need to proceed through the phase identifiable histopathologically as high-grade dysplasia in order to have invasion. In other words, invasion can occur in areas of dysplasia less severe than that categorized as high-grade dysplasia. The consequence of the lack of histopathologic criteria capable of distinguishing dysplasia with malignant potential, or more importantly dysplasia without invasive potential, is the necessity to consider prophylactic esophagectomy. Whether or not markers capable of predicting invasive behavior can ultimately be identified remains to be seen.