Can adenocarcinomas of the cardia and short segment of Barrett's mucosa be differentiated morphologically?
H.D. Appelman (Ann Arbor)
There is a group of adenocarcinomas that arise on one side or the other of the cardioesophageal junction. Some straddle and obliterate the junction. Barrett's carcinomas originate in dysplastic Barrett's esophageal mucosa proximal to the junction. Those arising in the gastric cardia have no dysplastic Barrett's mucosa, they are centered below the junction, and in rare cases, there may be adjacent dysplastic cardiac mucosa. In some studies, all proximal gastric carcinomas are equated with cardiac carcinomas, but that may be inappropriate, since carcinomas arising in the proximal body or fundus can invade the cardia, yet they are not cardiac carcinomas [1-3]. To complicate this situation, Barrett's carcinomas sometimes extend below the junction into the cardia, while some cardiac cancers grow upward to involve the distal esophagus.
There is little published data directly comparing Barrett's and cardiac cancers. Most studies analyze one or the other, but not both. The ideal way to know exactly whether a Barrett's carcinoma and a cardiac carcinoma are the same or different is to look at cancers that are totally confined to the cardia and to the Barrett's esophagus (BE). Such small cancers generally invade no deeper than the submucosa, and they are designated as "superficial" Barrett's esophageal cancers and "early" cardiac gastric cancers. For Barrett's cancers, it would seem reasonable to expect there to be considerable published data about the appearance of small cancers, because during surveillance small cancers are occasionally detected, and also because small cancers are found in about one third of esophagi removed for biopsy-proven high-grade dysplasia . However, the data base is actually very scant. We have no such screening program for cardiac cancers, mainly because we don't have a clear notion of the precursors. Thus the finding of a small cardiac carcinoma is pure serendipity. Yet there is about as much published data about the appearances of small cardiac cancers as about small Barrett's cancers. In one study of early Barrett's cancers from Germany, the tumors were tubule-producing and usually highly differentiated . A Japanese study of superficial cancers in the cardia found that tubule-forming carcinomas accounted for about 80% of the total with the rest being the diffuse or signet ring cell type . Furthermore, the greater the degree of intestinal metaplasia, the even greater likelihood the carcinoma will be tubule forming.
The rest of the data emerges from a few studies of these cancers at all stages, usually advanced. Two studies from single institutions, compared the histologic features of both Barrett's and cardiac adenocarcinomas of all stages. In one study, it became obvious that Barrett's and cardiac carcinomas were indistinguishable when it came to individual cases, but there were some minor differences in specific features for each group [6, 7]. Barrett's carcinomas tended to have a villiform surface architecture slightly more often, probably because that is a common surface contour for Barrett's mucosa itself. In contrast, cardiac carcinomas as a group tended to be more poorly differentiated. However, much of this difference resulted from the more frequent finding of signet ring cell carcinomas in the cardia. These hardly ever occurred as Barrett's cancers. Perhaps such cancers in the cardia are not really cardiac cancers, but are cancers arising in the proximal body, invading the cardia secondarily. Another study that separated cancers in this region into 3 primary sites, esophageal (that is, Barrett's cancers), cardiac (those arising at the junction) and subcardiac (those arising in the proximal stomach), much as the first study, found that Barrett's cancers were more highly differentiated than the other two .
In a series of 67 Barrett's cancers of all stages from France, tubular carcinomas of various degrees of differentiation accounted for 86%, with the rest being equally divided between mucinous and signet ring cell types . Only about one out of six of the tubular cancers were felt to be poorly differentiated. The authors also found that decreasing differentiation was related to increasing depth of invasion. In a study from the United States, of 26 Barrett's carcinomas, also of all stages, about 90% were tubular carcinomas, but in this study, the authors concluded that 60% were poorly differentiated, virtually the opposite of the French study .
To summarize, although there is only limited data on the histologic features of Barrett's and cardiac cancers, it seems safe to say that they are virtually identical. Most, by far are tubule-forming with a variety of additional types, such as mucinous, signet ring cell, and adenosquamous.
Although the degree of differentiation differs greatly from study to study, this may be the result of different grading criteria by different pathologists, and probably is of no clinical or epidemiologic consequence.
3. Okamura T, Korenaga D, Saito A, et al. Reactive changes in the esophageal epithelium and predictability of survival for patients with adenocarcinoma of the upper third of the stomach. Cancer. 1989;63:769-773.
6. Appelman HD, Kalish RJ, Clancy PE, Orringer MB. Distinguishing features of adenocarcinoma in Barrett's esophagus and in the gastric cardia. In: Spechler SJ, Goyal RK, eds. Barrett's esophagus: pathophysiology, diagnosis and management. New York: Elsevier Science Publishing Co Inc, 1985 .
7. Kalish RJ, Clancy PE, Orringer MB, Appelman HD. Clinical, epidemiologic, and morphologic comparison between adenocarcinomas arising in Barrett's esophageal mucosa and in the gastric cardia. Gastroenterology 1984;86:461-467.