| Japanese Journal of Clinical Oncology | Pages |
Editorial
EARLY DETECTION OF GASTRIC ADENOCARCINOMA: THE KEY TO REDUCE MORTALITY OR AN ILLUSION ?
A recent paper by Everett and Axon in The Lancet arguing the significance of early gastric cancer (EGC) might cause dangerous misunderstandings for Japanese readers (1). As the reduction in gastric cancer mortality remains vitally important in our country, other relevant information should be considered before we reach such conclusions.The authors suggested several reasons why some types of EGC might be a `pseudo-disease'. First, the diagnosis of intramucosal well differentiated adenocarcinoma is sometimes based on structural and cellular atypia, without any invasion into the lamina propria. Such cases would be diagnosed as severe dysplasia by many Western pathologists. Second, the reported doubling time of EGC is much longer than that of advanced cancer. Third, the few reports available regarding the natural history of EGC showed that one third of cases remained `early' for an average of 29 months, and 50% and 10% of patients survived more than 6 and 10 years, respectively (2,3). Fourth, no randomized trials show a benefit of a screening program for gastric carcinoma. Moreover there is `evidence' that increased detection of early cancers is not associated with reduced deaths from gastric cancer (4). Fifth, there is a Japanese series interpreted as suggesting that the increase in cases of EGC has not been associated with a reduction in the absolute number of cases of advanced gastric cancer. While the article concluded that a reasonable approach for early lesions is an endoscopic mucosal resection (EMR), the arguments give an overall impression that increased detection of EGC is more likely to be detection of a `pseudo-disease' and does not contribute to the reduction of mortality.
These arguments raise several issues common to the secondary prevention of any cancer and some specific arguments related to gastric adenocarcinoma. `Secondary prevention' refers to identification of curable lesions which would become lethal if left untreated. If most curable lesions are undetectable, or if most of the lesions detected are incurable, secondary prevention does not work. In the 1950s, EGC was first detected. Most lesions detected then had prominent ulcerative change or were clearly protruding (5). Since then, many advanced gastric cancers were discovered between regular annual endoscopic surveillance, which made Japanese endoscopists feel guilty for overlooking early lesions and led to a search for `unrecognizable' initial lesions. Fortunately, due to the practice of recording endoscopic observation on 16 mm film in all cases, review of previous endoscopy records was possible. This allowed identification of the precursor lesions. Thus, it was shown that many advanced gastric cancers originate as subtle mucosal change, with discoloration, hyperemia or slight unevenness (5). These lesions keep such features until they eventually (and often suddenly) show the typical appearance of advanced gastric adenocarcinoma. However, it is impossible to predict how long it will take for a particular lesion to progress into advanced cancer and what is the turning point. Some remain early for more than 5 years, sometimes even 10 years, but others for just a few months (6). These lesions with minimal mucosal change (LMMC) can show poor or well differentiated histology: one third the former and two thirds the latter (5). Through detection of LMMCs the total number of EGC has increased year by year. It was recognized that well differentiated LMMCs that are confined to the mucosa do not have lymph node metastasis (7). Thus many can be treated safely by EMR.
There are no prospective randomized trials proving the benefit of screening for gastric cancer, melanoma, prostate cancer or colonoscopic screening for colorectal cancer. However, the lack of proof that screening reduces mortality does not prove the reverse. It certainly does not clarify the significance of EGC. Screening for any highly prevalent and lethal disease detects some lesions that were destined to remain occult. This is an issue in prostate, breast and colorectal cancer and is a cost of secondary prevention programs. The experience with breast cancer is relevant to this discussion. While many invasive cancers arise from ductal carcinoma in situ (DCIS), not all DCIS results in life-threatening disease. Population-based, mammographic screening has led to the situation where about 15% of all breast cancers in the USA are DCIS. Autopsy series show a 9% incidence of asymptomatic DCIS, suggesting that much of the screen detected DCIS would have remained occult (8). This does not mean that it should not be treated; it invites research into ways of identifying high-risk lesions. For EGC, a proportion of screen detected lesions is likely to be similarly clinically unimportant. It is important that patients and clinicians realize this, but it does not invalidate the concept of EGC as a significant lesion.
International differences between pathologists over the use of the terms `well differentiated carcinoma' or `dysplasia' only concern at most 20% of EGC. There is no dispute over the remaining 80%. A paper by Schlemper et al. (9) showing inconsistency in biopsy diagnosis between Western and Japanese pathologists focused on highly selected cases treated by EMR. The lesions in question were diagnosed as well differentiated (intestinal-type) adenocarcinoma in Japan. Thirteen lesions had some elevated component and another six were LMMCs. The results show that many invasive cancers may be diagnosed as severe dysplasia by the Western pathologists (potential under-diagnosis) and that many superficial elevated lesions or LMMCs are diagnosed as adenocarcinoma in Japan (potential over-diagnosis). Western pathologists are clearly concerned about over-diagnosis leading to unnecessary surgery. Japanese pathologists, on the other hand, are more concerned about overlooking cancers, probably because of the much higher incidence of true gastric carcinoma.
Western surgeons often find that patients with a biopsy of severe dysplasia are subsequently diagnosed with carcinoma. Sometimes these cancers are incurable. Eleven of 13 patients diagnosed with high grade dysplasia from Leeds, UK, had gastric carcinoma within 15 months, suggesting such lesions should be treated with respect, irrespective of semantics (10). There is an interesting paradox between the West and Japan; Western surgeons treat lesions diagnosed as severe dysplasia by open surgery while the Japanese diagnose them as cancer but treat them by EMR (11).
Accurate identification of markers that the disease has reached the point of no return is needed. Recently much progress has been made towards identification of the multiple molecular events that occur during gastric carcinogenesis (12). Application of this knowledge by means of immunohistochemical and molecular analysis of biopsy specimen is being pursued in a prospective study in Hiroshima. Until proven markers are available, it is wise to consider well differentiated EGC a real disease, not a `pseudo-disease'. Because EMR is available to treat many such lesions with minimal morbidity in Japan, there is little reason other than cost not to do so.
The little available natural history data suggests that most EGC will progress to advanced disease if untreated (2,3). In addition, EGC symptoms resemble those of benign disease. Clinicians faced with patients with unexplained dyspepsia must therefore maintain an index of suspicion. A recent report from Gloucester, UK, found that 24 out of 25 patients under 55 years old diagnosed with gastric cancer had signs or symptoms other than dyspepsia, mainly loss of body weight. Only one had EGC (T1,N0), and 21 relapsed after a short follow up (13). Another report from Middlesborough, UK, showed that one in six patients dying from gastric cancer had a normal or benign endoscopy for dyspepsia within 3 years of diagnosis (14). Given the doubling time for EGC the cancers must have been present at the earlier investigation, but were overlooked. This all suggests that improved knowledge and awareness of the various forms of EGC is required, rather than a reassessment of its very existence.
Everett and Axon referred to a prospective non-randomized study to detect stump cancer as evidence that increased detection of EGC does not reduce deaths from gastric cancer (4). The outcomes of endoscopically screened and non-screened groups were compared. However this study does not provide such evidence for two reasons. First, the majority of the non-screened population were those who did not accept an offer of endoscopic survey for remnant stomach cancer, while the screened population were those who accepted it. Enormous bias exists with such a selection method. Second, 11 out of 12 patients who died from gastric stump cancer in the screened group may have been saved if the program had been carried out in Japan. Two were treated for EGC by a local resection of the stump without lymph node dissection and eventually developed local recurrence. In spite of regular invitation, another five patients did not undergo endoscopy at least 6 years before they had incurable disease. Another patient whose previous biopsy was diagnosed as dysplasia did not undergo endoscopy for 5 years. For one patient endoscopic biopsy was not performed because regular use of warfarin, resulting in the delayed diagnosis of signet cell carcinoma. Two other patients developed symptoms and eventually were diagnosed with gastric cancer in the 3-year interval between examinations. In summary, the study highlights difficulties in screening, but does not invalidate the concept of EGC.
The absolute number of deaths from gastric cancer in Japan has remained constant for 20 years in spite of a remarkable increase of EGC. Surely EGC is therefore unrelated to advanced cancer? The pit-fall in this argument is that the death rate increases dramatically with age (15). It accounts for less than 10 per 100 000 population at the age of 40 and about 100 at 60, while it increases to more than 500 at 80 years old. Considering the rapid increase of the aged population over 60 in Japan, the number of gastric cancer deaths should have increased dramatically. In fact, the total number of deaths is stable and age adjusted death rate is decreasing. If EGC was overlooked, we probably could not have prevented a major increase in the number of deaths from gastric carcinoma.
Recently one of authors (MS) cared for a patient who had refused treatment of an EGC. He came to our hospital with a depressed type EGC, already followed up for a couple of years in another hospital. Despite our strong recommendation, he preferred to be observed and the lesion progressed at each endoscopy. Ten months after his first visit, the lesion was diagnosed as advanced carcinoma. He then agreed to surgery. Eventually, a cancer with lymph node metastasis to the second tier was removed. After one year, he developed lymph node recurrence which eventually became systemic disease. He died recently with regret for not undergoing treatment when the diagnosis was made. This case demonstrated clearly that we should prevent the idea that EGC is a pseudo-disease from spreading out widely.
It is very difficult to estimate how long any lesion can be safely observed without treatment. What we can tell our patients is the diagnosis at each moment, the possibility of lymph node metastasis and the risk of progress. As we cannot distinguish benign or slow growing lesions from those with potential for constant evolution, we must not regard EGC as a pseudo-disease.
References
|
Mitsuru Sasako, M.D. Department of Surgical Oncology National Cancer Center Hospital, Tokyo and Gregory Bruce Mann, M.D. Visiting Doctor to the National Cancer Center Hospital by a Fellowship from the Foundation for Promotion of Cancer Research Department of Surgery University of Melbourne, Melbourne |
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 Aug 1998
Copyright©Japanese Journal of Clinical Oncology, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||