The changing pattern of adenocarcinoma of the esophagogastric junction
P.I. Reed (Slough)
In the past 40 years the pattern of esophageal cancer has changed dramatically in Western and Northern Europe, North America and Australasia. In the mid-1950's adenocarcinoma (AC) represented less than 5% of all primary malignant tumours arising in the esophagus, but by the late 1980s this had risen to almost 50% in several countries. By 1987 it actually exceeded the rate of squamous cell carcinoma (SCC) in the United States among white males less than 55 years of age  (Tables I and II). In our own department in Southern England AC accounted for 70% of all esophageal cancers in the 5 years to 1995 (Table II). Data on cancer incidence during the decade to 1987 from population-based cancer registries in the USA indicated that the rate of increase in frequency of AC of the esophagus and gastric cardia exceeded that of any other cancer, including cutaneous melanoma, non-Hodgkin's lymphoma and lung cancer .
By the 1970s the association between AC and Barrett's esophagus (BE) became well established [9, 28] and it is now accepted that almost all esophageal AC develop in the specialized metaplastic columnar epithelial lining of the lower esophagus damaged by extensive and prolonged gastroesophageal reflux.
The true incidence of BE is not known but an autopsy study at the Mayo Clinic  suggested that the frequency could be up to 20 times the endoscopic incidence in symptomatic gastroesophageal reflux patients. Not surprisingly AC are mostly located in the lower third of the esophagus near to or encroaching on the esophagogastric junction (EGJ), unlike SCC which are found throughout the esophagus but with the majority developing in the middle third of the esophagus [30, 31].
The reported average prevalence of adenocarcinoma in BE is almost 10% but with a suggestion that this may be rising (Table III). The majority of the reports are based on small numbers of patients, only five exceeding 100 BE cases. The incidence of AC in BE ranges between 1 in 48 and 1 in 441 patient years in published series, with the risk estimated at between 30 to 125 times greater than that in the general population (Table IV). These data do suggest that AC of the esophagus manifests clinically late in the neoplastic process thus contributing to a generally adverse prognosis for these tumors. As the follow-up time of BE patients has generally been short, rarely exceeding 10 years, the true incidence and estimated risk of AC in BE is probably higher than reported so far.
Concurrently with the changing patterns of esophageal carcinoma important changes also have taken place in gastric cancer (GC) incidence, which has been declining worldwide for over four decades, especially in the developed countries of Europe, North America, Australasia and Japan. Data from various studies have demonstrated that this decline was entirely due to decreasing incidence rates for distal gastric carcinomas [57-59]. The other very significant change has been the steady increase in the incidence of carcinoma of the cardia [22, 24, 25, 60] (Table V) and concurrently also of adenocarcinoma of the EGJ which is about twice as common as esophageal AC [1, 23, 59, 65] (Table VI). However, most of the earlier published studies are misleading because they have not distinguished tumors of the EGJ from cancers of the cardia . Indeed in 1976 a pathologist writing to a medical records clerk responsible for coding in an hospital in our region wrote: "I am surprised to hear that you often receive reports of adenocarcinoma of the esophagus, I would suggest that you change this to cardia of the stomach". He was not unique in this regard.
The observed increasing incidence of AC of the lower third of the esophagus was thought by some workers to be due to the increasingly widespread use of flexible endoscopy, thus ensuring greater diagnostic accuracy which might reflect a change in classification from cardiac tumors to esophageal tumors. However, the latter would probably be cancelled out by the reverse classification changes already mentioned. Analysis of published data [59, 60] confirms that a 4-5 fold increase in age-adjusted incidence rates for adenocarcinoma of the esophagus as well as gastric cardia has taken place over the past three-four decades. Though affecting both sexes, this increase has been more pronounced in males, the recorded sex ratios usually exceeding 3:1. In the USA this trend has only been evident in whites of both sexes, perhaps reflecting their generally higher socio-economic status. A parallel trend has also been shown in a British study by Powell and McConkey  in which 16% more cardiac cancers and 27% fewer distal gastric cancers were recorded in social (professional) classes 1 and 2 than if the distribution had been the same as that for all sites of cancer. In the USA the incidence rates of both SCC of the esophagus and AC of the antrum of the stomach were much higher in blacks of both sexes than in whites [1, 59]. The sex ratios for both these cancers are also very different compared with esophageal and cardiac AC. Numerous studies have revealed an higher male:female ratio in carcinoma of the cardia [24, 25, 58, 69] compared with all other gastric cancers, which have a male:female ratio of 1.9:1 [24, 25].
The higher male:female ratio for adenocarcinoma of the cardia, comparable to adenocarcinoma of the esophagus and EGJ has not only suggested the possibility of a common etiological mechanism but also whether tobacco and alcohol might be risk factors similar to those seen with SCC. Some support for an association with cigarette smoking has come from the studies of Skinner et al. , Wang et al.  MacDonald and MacDonald  but only for carcinoma of the cardia, not AC of the esophagus [70, 71]. Levi et al. [63, 64] have shown that in the Swiss Canton of Vaud carcinoma of cardia was significantly associated with both tobacco and alcohol consumption, as have Kalish et al.  in Michigan, USA. While smoking and drinking are not risk factors for the development of BE in patients with severe esophageal reflux, they become relevant in subjects with a specialised metaplastic columnar epithelium, especially if the intake is heavy as Gray et al.  have noted in their series where 95% of patients who developed AC in BE smoked and over 50% drank >40 units alcohol per week. However, they found the smoking and drinking histories of their AC patients to be similar to those with severe gastroesophageal reflux indicating that development of BE in itself is independent of smoking and alcohol and that a combination of BE and heavy smoking is required for initiation and promotion of AC of the esophagus.
Both population and single centre studies in North America, Western and Northern Europe and Australasia confirm that adenocarcinoma of the esophagus is strongly associated with BE and the majority of the cancers arise at or near the EGJ.
There has been a steadily rising incidence of adenocarcinoma of the esophagus and cardia of stomach over the past 30 years, especially in white males of higher socio-economic status. The average age of diagnosis of AC is lower than squamous cell carcinoma.
Smoking and alcohol are risk factors for carcinoma of cardia and smoking may be for esophageal adenocarcinoma but only in an established specialized metaplastic columnar epithelium.
There has been a steady decrease in distal gastric cancers for over four decades which have a lower sex ratio than cancer of the cardia.
Squamous cell carcinoma of the esophagus is strongly associated with smoking and alcohol and in the USA is much more common in blacks than whites of both sexes. Its incidence has either remained steady or been strongly falling.
I am grateful to Belinda Johnston and Sharon Naylor for typing the manuscript.
18. O'Donoghue J, Waldron R, Gough D, McCabe J, Kerin M, McGuire M, Horgan PG, Given HF. An analysis of the diagnostic accuracy of endoscopic biopsy and cytology in the detection of esophageal malignancy. Eur J Surg Oncol 1992;18:332-334.
32. Naef AP, Savary M, Ozzello L. Columnar-lined lower esophagus: an acquired lesion with malignant predisposition. Report on 140 cases of Barrett's esophagus with 12 adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:826-834.
43. Schnell T, Sontag S, Wanner J, Chirtam R, Chejfec G, O'Connell S, Moroni B. Endoscopic screening for Barrett's esophagus (BE), esophageal adenocarcinoma (AdCa) and other mucosal changes in ambulatory subjects with symptomatic gastroesophageal reflux (GER). Gastroenterology 1985;88:1576.
47. Ollyo JB, Savary M, Monnier P, Wellinger J, Gonvers JJ, Fontolliet C. Is the prevalence in patients with Barrett's esophagus overestimated? In: Giuli R, McCallum RW, eds. Benign lesions of the esophagus and cancer:answers to 210 questions. New York: Springer-Verlag, 1989:765-767.
53. Skinner DB. The incidence of cancer in Barrett esophagus varies according to series. In: Giuli R, McCallum RW, eds. Benign lesions of the esophagus and cancer: answer to 210 questions. New York: Springer Verlag 1989:764-765.
59. Locke III GR, Talley NJ, Carpenter HA, Harmser WS, Zinsmeister AR, Melton III LJ. Changes in the site and histology-specific incidence of gastric cancer during a 50-year period. Gastroenterology 1995;109:1750-1756.
62. Zheng TZ, Mayne ST, Holford TR, Boyle P, Liu WL, Chen YT, Mados M, Flannery J. Time trend and age-period-cohort effects on incidence of esophageal cancer in Connecticut, 1935-89. Cancer Caus Cont 1992;3:481-492.
70. 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;86:461-477.
71. Brown LM, Silverman DT, Pottern CM, Schoenberg JB, Greenberg RS, Swanson EM, Litt JM, Schwartz AG, Hayes RB, Blot WJ. Adenocarcinoma of the esophagus and esophagogastric junction in white men in the United States: alcohol, tobacco and socio-economic factors. Cancer Caus Cont 1994;5:222-340.