The paper by Ono and colleagues1 in this issue of Gut documents the remarkable results achieved at the Tokyo National Cancer Centre Hospital (NCCH) with endoscopic management of early gastric cancer (EGC) (see page225). EGCs constitute a much higher proportion of the total number of gastric cancers in Japan than is the case in the West. There are a number of reasons for this: firstly, many asymptomatic subjects in Japan are screened for gastric cancer (usually by barium radiology, followed, if necessary, by endoscopy); secondly, Japanese diagnostic gastroscopy is a much more careful procedure than the “smash and grab” style of endoscopy which is typical in this country: the stomach is inflated to a greater degree, indigo carmine dye spraying is used to examine any suspicious area, and simethicone (to eliminate bubbles and froth) and hyoscine (to paralyse the stomach) are used routinely. There is some evidence that early lesions are missed in the UK: a recent audit in one centre showed that 11 of 81 patients presenting with advanced gastric cancer had undergone endoscopy within the previous two years.2

Is EGC in Japan the same disease as EGC in the West? A recent analysis of published data from Japan and the rest of the world3concluded that histological type, macroscopic appearance, degree of invasion, and frequency of lymph node metastases are the same in Europe as Japan. Japanese classification of tumour invasion, as Onoet al have pointed out, differs, with lesions classified as “high grade dysplasia” in the West called “intramucosal carcinoma” in Japan. A new terminology—the Vienna classification of gastrointestinal neoplasia4—has been proposed in an effort to overcome this discrepancy.

Endoscopic mucosal resection (EMR) for EGCs confined to the mucosa seems a very attractive alternative to surgery. Onoet al wish to “promote its use around the world” and their paper is a powerful argument in favour of EMR. Could EMR become a standard therapy for mucosal EGC in this country? EMR demands a high level of endoscopic expertise, as well as a cooperative patient. I have had the opportunity to examine the case from both sides, having visited the NCCH and having played host to endoscopists from the NCCH who worked in our unit over the past few years. I have personally witnessed an EMR at the NCCH which took over three hours to complete (the lesion was large); the patient was lightly sedated with midazolam and remained cooperative throughout. Our patients are less likely to tolerate these procedures as well as the Japanese. Endoscopists from the NCCH carried out a number of EMRs for EGC in our unit in April of this year as part of a “live” demonstration in a course on endoscopic management of early gastrointestinal cancer, and all patients required a general anaesthetic. As well as anaesthetic input, EMR requires special equipment and a dual channel therapeutic gastroscope. Although the technique of “strip biopsy” is relatively straightforward, and suitable for smaller lesions, it is more likely to result in resection in multiple fragments which makes histological evaluation of resection much more difficult. (In this series, completeness of resection could not be evaluated in 20%.)

The NCCH endoscopists now prefer the newer technique employing the insulation tipped diathermic knife (IT knife) which is not commercially available in the UK. This technique is technically much more demanding and time consuming but can be used to resect much larger lesions: I have seen an 8 cm EGC resected with this technique (the patient was unfit for surgery). EMR is not generally suitable however for large lesions: in this series, complete resection was achieved in only 38% of lesions >3 cm.

Alternatives to strip biopsy have been described by other Japanese groups: these include aspiration mucosectomy5 and EMR using a ligating device.6 Proponents of aspiration mucosectomy claim it is technically easier than strip biopsy.

Japanese endoscopists are expert in assessing depth of invasion of gastric cancers purely on the basis of endoscopic features.7 Some criteria, such as size and presence of ulceration, are straightforward, but others, such as the macroscopic configuration (superficial, elevated, depressed, etc) and mucosal fold pattern, are more subjective and may present difficulty for western endoscopists. Ono et al “do not use endoscopic ultrasonography (EUS) routinely, as it is not sensitive enough to evaluate minute invasion to the submucosa.” Nevertheless, 15% of resected lesions showed submucosal invasion. I suspect that Western endoscopists would opt for routine use of EUS before embarking on EMR.

Endoscopists inexperienced with EMR will have concerns about complications of bleeding and haemorrhage. Ono et al had a relatively high incidence of perforation at 5% (the risk of bleeding is not stated). Remarkably, most of these perforations were managed by endoscopic clipping. Japan has an entirely different medical culture to the UK: litigation and complaints are uncommon, and advice given by doctors is usually accepted without question. British endoscopists will be warier of new techniques with potentially serious complications. In this brave new world of clinical governance, who is to determine what an acceptable complication rate should be for a new procedure? The current political climate in the NHS discourages the development of techniques such as EMR. Few centres in this country are likely to diagnose enough EGCs to become expert in EMR, and it would be desirable, therefore, for this technique to become concentrated in a few specialised centres.

Patients with EGC can expect a five year survival rate in excess of 90% with surgery8 ,9; why then should we in the West embrace EMR? Ono et al have pointed out a number of compelling reasons: firstly, EGCs fulfilling the criteria for EMR have a lower risk of lymph node metastasis (0.36%) than the mortality rate from surgery for EGC at their centre (0.5%); secondly, EMR is considerably cheaper compared with conventional surgery; and thirdly, there were no gastric cancer related deaths in this series during a median follow up period of 38 months. Surgery will continue to play a key role in the management of EGC: despite the success of endoscopic therapy, most EGCs at the Tokyo NCCH are still treated surgically.

EMR in this country is likely to develop initially as therapy for EGC in frail patients who would pose a high risk for radical surgery; with increasing experience, we can aspire to the outcomes achieved by the Japanese.