Of what value is the determination of DNA ploidy in the classification of epithelial changes in Barrett's mucosa?
A.H. Mulder, H.W. Tilanus, M. van Blankenstein, W.C.J. Hop, M.B.E. Menke-Pluymers (Dordrecht)
The metaplastic columnar change in Barrett's esophagus carries an increased risk of adenocarcinoma development, which is estimated to be 30-125 fold higher than in the general population [1, 2]. The use of endoscopic surveillance for the detection of a developing carcinoma has been challenged and indicates that a more careful delineation of a subgroup of patients, which are at risk, is mandatory . Dysplastic changes in the columnar epithelium haven been shown to be associated with a still further increased risk of malignancy in Barrett's esophagus when compared to all patients with a Barrett's esophagus [4, 5]. Changes in the DNA content in the metaplastic cells have also been implicated to carry a similar increased risk [4, 6]. A discordance between the presence of dysplasia and changes in the DNA content (or DNA ploidy) has been reported . This may suggest that dysplastic changes and DNA ploidy are independent prognostic factors.
In order to evaluate the association of dysplasia and DNA ploidy with malignant change in Barrett's esophagus we compared the presence of both parameters in biopsy material from patients with a Barrett's esophagus (who developed no carcinoma during a follow-up period of at least five years) with patients with an adenocarcinoma in Barrett's esophagus.
Materials and methods
Patients with a Barrett's esophagus are defined as having at least three centimeters of metaplastic columnar epithelium in which at least intestinal changes are present. Patients with an adenocarcinoma in a Barrett's esophagus were only accepted when a segment of Barrett's epithelium was present above the tumour. A more elaborate presentation of the patient data is presented elsewhere .
Three to eight biopsies were removed endoscopically from 38 patients with a Barrett's esophagus. Routine hematoxylin azophloxin slides were prepared and designated as having no, mild, moderate or severe dysplasia using criteria described by Morson . Mild dysplasia overlapping with low grade dysplasia and moderate and severe dysplasia with high grade dysplasia.
Slides from Barrett's mucosa in patients with an adenocarcinoma in a Barrett's esophagus were similarly prepared and analyzed, but were derived from either resection specimens (14 cases) or biopsies taken from the metaplastic, non-tumorous area.
Nuclear suspensions were prepared using the methodology as described by Hedley . The cellular DNA content was measured on a fluorescence activated cell sorter (FACS II, Becton Dickinson). Diploidy was defined as a DNA index of 0.90-1.10. Tetraploidy was defined to be present when more than 10% of the events were having a DNA-index between 1.90 and 2.10. Aneuploidy was defined to be present when the specimen was neither diploid or tetraploid.
For the comparison of percentages in Table I Fisher's exact test was used.
The histologic grade, i.e. dysplasia, is expressed versus DNA-ploidy in Tables I and II.
Ploidy did not significantly correlate with the grade of dysplasia in patients without (Table I) or with an adenocarcinoma present (Table II). Aneuploid and tetraploid cases were significantly more often present in patients with an adenocarcinoma when compared to those without a tumor (p < 0.001).
These results were used in a multivariate analysis and are described elsewhere . A significantly increased risk of adenocarcinoma was associated with dysplasia, both low and high grade (odds ratio resp 9.4 and 33.1) and aneuploidy or tetraploidy below the age of 65 years (odds ratio 44.7). Above 65 years of age aneuploidy/tetraploidy did not result in a significantly increased risk for malignancy in a Barrett's esophagus (odds ratio 2.4). Since the mean age of the patients was 63 years, about half of the patients will be present in this group.
Our results indicate that both histologic grade and DNA-ploidy behave as independent risk factors for malignancy in patients with a Barrett's esophagus. DNA-ploidy is a risk factor which is present only in the group of patients younger than 65 years of age. In previous studies these parameters were described to be correlated [4, 11].
In patients with an adenocarcinoma in a Barrett's esophagus about 31% were shown to be diploid. Virtually all these tumors were moderately or poorly differentiated and also no significant correlation could be established between DNA-ploidy and histologic grade . Thus, in both dysplastic mucosa and in Barrett's adenocarcinoma, histologic grade behaves independent when compared to DNA-ploidy. Both dysplasia and DNA-ploidy are risk factors and can therefore be used to delineate the subgroup of patients at risk more closely. In our series 17/38 = 45% of all patients with a Barrett's esophagus fall in the group which is neither diploid nor histologic normal and carries the increased risk (Table I). The size of this group is much larger than the fraction of patients who will eventually develop an adenocarcinoma (a few percent). It is unrealistic to screen such a large group of patients more frequently and it is perhaps wise to use only the highest risk factors, being moderate/severe dysplasia and DNA ploidy under the age of 65 years. The present risk factors need to be more accurately defined or others should be added.
4. Reid BJ, Hagitt RC, Rubin CE, Rabinovitch PS. Barret's esophagus. Correlation between flow cytometry and histology in the detection of patients at risk for adenocarcinoma. Gastroenterology 1987;93:1-11.
6. Reid BJ, Blount PL, Rubin CE, Levine DS, Haggitt RC, Rabinovitch PS. Flowcytometric and histologic progression to malignancy in Barrett's esophagus: propective endoscopic surveillance of a cohort. Gastroentrology 1992;102:1212-1219.
10. Hedley DW, Friedlander ML, Taylor IW, Rugg CA, Musgrove EA. Method for analysis of cellular DNA content of paraffin embedded pathological material using flow cytometry. J Histochem Cytochem 1983;31:1333-1335.