What is the value of brush cytology in the diagnosis of non-dysplastic Barrett's esophagus?
H.D. Appelman (Ann Arbor)
Several published studies comparing biopsy to brush cytology in the diagnosis of dysplasia and carcinoma in Barrett's mucosa suggest that these two diagnostic technics are complementary [1-4]. The diagnostic yield is higher when both are employed than when only one is used. In contrast, there is very little published data on the use of brush cytology in the diagnosis of non-dysplastic, non-carcinomatous Barrett's mucosa. Therefore, since the literature is so limited, the answer to this question clearly requires input from highly sophisticated esophageal cytopathologists who can offer some insights based upon personal experiences. I consulted with several such cytopathologists, including Drs. Kim Geisinger from the Bowman Gray School of Medicine, Claire Michael from the University of Michigan and Barbara McKenna from Albany Medical College, and all of them told me much the same thing with minor variations [5-7]
The diagnosis of Barrett's mucosa, whether by biopsy or by brushing, depends upon two conditions: 1) an abnormal columnar mucosa, and 2) its presence within the tubular esophagus. To satisfy these requirements, it is essential that Barrett's mucosa be clearly defined cytologically. Brush cytology strips off superficial epithelium. In Barrett's mucosa, this is columnar epithelium, most of which looks like surface epithelium in the stomach. The only reasonably specific cytologic feature of Barrett's epithelium is the goblet cell, a component of the incomplete intestinal metaplasia that characterizes the specialized or distinctive epithelium . We depend upon goblet cells to make a confident biopsy diagnosis of Barrett's mucosa in most cases. However, not all Barrett's mucosa has goblet cells. Some is modified, atrophic-appearing gastric type epithelium, covered by typical gastric-type surface cells. This is particularly true in small children who often don't have goblet cells in their Barrett's mucosa. Thus if goblet cells are considered to be the only cells diagnostic of Barrett's mucosa, then some Barrett's will be missed by brush cytology, because cytology will not be able to detect the architectural abnormalities that will appear in the biopsy. It is also important that the endoscopist brush only mucosa that is clearly within the tubular esophagus, since there may be an irregular squamocolumnar interface at the cardioesophageal junction, that is an irregular Z-line. As a result, it has been suggested that no brushings be obtained from the distal inch or about 2.5 cm in order to avoid the irregular cardioesophageal junction . However, this immediately prevents the diagnosis of short-segment Barrett's by brush cytology. In addition, the mucosa that is brushed should look like Barrett's endoscopically.
Nevertheless, there are some benefits to brush cytology. It can sample a larger surface area than the biopsy. Furthermore, it is cheaper than a biopsy, but only if it is used alone. It adds expense, if it is used to complement the biopsy.
The standard against which brush cytology is measured remains the biopsy. There is very limited data suggesting that brushings can be complementary to biopsy in the diagnosis of Barrett's mucosa [3, 4]. In the largest series, the authors reported that about 1/3 of Barrett's mucosa of the distinctive type was detected by brushings when the biopsies were not diagnostic . However, the proof that these cases really had Barrett's was not always convincing. About half of the discrepancies occurred in patients who had had earlier biopsies with distinctive Barrett's, so it was assumed that the positive cytologic exams were true positives and the negative biopsies were false negatives. In the other half, the biopsies contained ulcers, so that there was no biopsy proof that the brushings actually were diagnosing Barrett's.
Finally, there have been a few studies evaluating balloon cytology in Barrett's esophagus [8, 9]. The value of balloon cytology is that it does not require endoscopy, and so it is much less expensive than either biopsy or brushing . The balloon is passed into the stomach, and gradually withdrawn, scraping the esophageal mucosa as it is pulled upward. The scant data indicate that this technic is not as sensitive as either brush cytology or biopsy in detecting Barrett's mucosa, especially short segment Barrett's [8, 9].