ORIGINAL ARTICLE 

Gastric adenocarcinoma missed at endoscopy

A. AMIN, H. GILMOUR*, L. GRAHAM, S. PATERSON-BROWN, J. TERRACE and T.J. CROFTS

Department of Surgey and *Department of Pathology, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh, UK

Introduction Methods

Results Discussion References

Keywords: gastric adenocarcinoma, diagnostic delay, endoscopy

Background and purpose: The diagnosis of gastric cancer is based on histological confirmation at endoscopy with the emphasis on early detection to improve prognosis. The aims of this study were to identify the proportion of patients with gastric adenocarcinoma in whom the diagnosis was missed at first endoscopy and the subsequent delay which occurred before the histological diagnosis was established. Methods: Retrospective review of 137 consecutive patients with biopsy-proven gastric adenocarcinoma presenting to one surgical unit over a five-year period. Results: Two patients with a biopsy diagnosis at laparotomy and 6 patients in whom case notes could not be traced were excluded from the study. Of the remaining 129 patients, the diagnosis of gastric adenocarcinoma was missed at first endoscopy in 18(14 %). The median delay to histological diagnosis in this subgroup of patients was 13 weeks (range 3-102). Conclusion: Delays in establishing the diagnosis of gastric adenocarcinoma following initial endoscopy occur in a number of patients. Greater suspicion and a more rigorous protocol for repeat endoscopy and biopsy must be implemented in order to reduce the number of missed diagnoses after initial endoscopy

J.R.Coll.Surg.Edinb., 47, October 2002, 681-684 

INTRODUCTION

Gastric cancer is the fourth commonest cause of death from malignant disease in the United Kingdom, and the overall 5-year survival is between 5-10% due to the advanced stage at presentation.^1-3 A policy of investigating dyspeptic symptoms in patients aged over 45 years with prompt endoscopy in order to diagnose gastric cancer at an earlier stage, has been suggested to improve prognosis.^4,5

The role of endoscopy in providing a histological diagnosis of gastric cancer is well-established, however, long delays can still occur.^6-8 In a study of 115 patients from Leeds with carcinoma of the oesophagus and stomach, there was a median delay of only 2.8 weeks from referral to hospital to histological diagnosis, but in a sub-group of 17 patients this delay ranged between 12-87 (median 33) weeks.⁹ Reasons behind this longer delay- whether due to endoscopic error or due to delay in referral from the outpatient clinic -were not reported. There is also a recognised false negative rate for gastric carcinoma diagnosed following endoscopy and biopsy ranging between 14-27%.^{4, 10, 11} However, the identification of missed gastric adenocarcinoma and subsequent diagnostic delay was not the primary objective in any of these studies.

The aims of this study, therefore, were to identify the proportion of patients with gastric adenocarcinoma in whom the diagnosis was missed at first endoscopy and the subsequent delay which occurred before the histological diagnosis was established.

METHODS

All patients with biopsy-proven gastric adenocarcinoma diagnosed between January 1994 and December 1999 were identified from a prospectively collected surgical database which records details of all patients with oesophago-gastric cancer referred to the University Department of Surgery at the Royal Infirmary of Edinburgh for a surgical opinion. This unit provides the regional service for all upper gastrointestinal surgery in Edinburgh and the Borders. Patient case notes were analysed retrospectively and the local pathology database was used to cross-reference the biopsy and pathology reports for every patient to ensure completeness.

Diagnostic delay was measured as the interval between the date of the biopsy report at first endoscopy to the date of the biopsy report confirming the histological diagnosis of gastric adenocarcinoma. For patients in whom no biopsy was obtained at first endoscopy, this delay was the interval between the date of the first endoscopy to the date of the biopsy report confirming histological diagnosis. The delay was then approximated to the nearest week. Patients undergoing endoscopy more than 2 years before a histological diagnosis of gastric adenocarcinoma, were not considered to have had the diagnosis missed at the time of this earlier endoscopy.

RESULTS

One hundred and thirty seven consecutive patients with biopsy-proven gastric adenocarcinoma were identified over the five-year period. Eight patients were excluded from the study; two because the diagnosis was made at laparotomy for a perforated ulcer and 6 because the case notes could not be traced. Of the remaining 129 patients, 79 were male and the median age at diagnosis was 72 years (range 32-91).

A histological diagnosis of gastric adenocarcinoma was established in 101(78%) patients at first endoscopy. (Table 1) In 10(8%) patients the histology was inconclusive at first endoscopy, but the biopsy specimens were recognised to be either too superficial for adequate microscopic assessment or showed cytological atypia without conclusive evidence of malignancy. In 8 of these 10 patients, there was macroscopic evidence of malignancy at endoscopy and a histological diagnosis was obtained within four weeks in all 10 after repeat biopsy. For the purposes of this study, these patients were not considered to have a diagnosis of gastric adenocarcinoma ‘missed’ at first endoscopy.

In the remaining 18(14%) patients with gastric adenocarcinoma, the diagnosis was missed at first endoscopy with symptoms subsequently attributed to benign pathology. One adenocarcinoma was missed in a gastric remnant following previous surgery for peptic ulcer disease and the remaining seventeen were missed in an ‘unoperated stomach’. Histological diagnosis was eventually established in these 18 patients following a median delay of 13 weeks (range 3-102).

Table 1: Outcome after first endoscopy and diagnostic delay

Overall, a histological diagnosis of gastric adenocarcinoma was established after the first, second, third and fourth endoscopy in 78%, 94%, 99% and 100% patients, respectively.

Table 2 outlines findings at initial endoscopy for each of the 18 patients in whom gastric adenocarcinoma was missed. The indication for a repeat endoscopy (at which histological diagnosis was established) and the diagnostic delay is also summarised. In seven patients no biopsy specimens were obtained at first endoscopy. Although 4 of these 7 patients had normal endoscopic findings, three patients with visible endoscopic lesions - one small ulcer, one healed ulcer and one stricture -were not biopsied. Five of these seven patients were diagnosed when admitted to hospital as an emergency at a later date and two diagnosed when referred back for repeat endoscopy by the general practitioner. Longest delays occurred in this sub-group of seven patients. The remaining eleven patients were diagnosed at follow-up endoscopy after a course of medical treatment although more than two endoscopies were necessary in some patients before a histological diagnosis was established. Malignancy was suspected endoscopically in 4 patients at first endoscopy - three ulcers and one polypoid raised lesion - but histology proved to be benign. Thus, biopsy of all endoscopically visible lesions and immediate repeat biopsy of all endoscopically suspicious lesions with benign histology, might have confirmed adenocarcicoma in seven patients preventing subsequent delay.

Table 2: Biopsy number and macroscopic suspicion of malignancy at first endoscopy for the 17 patients in whom gastric adenocarcinoma was missed

Table 3 shows the pathology of the 18 missed adenocarcinomas. The majority of the adenocarcinomas were poorly differentiated. Only 4 of the 18 missed adenocarcinomas were stage IV and irresectable at the time of diagnosis. Two of these 4 patients were unsuitable for any surgical intervention and two patients had palliative bypass procedures. One patient with clinically resectable disease was unfit for surgery due to coexisting medical conditions. Table 3 also shows the site of the 18 missed adenocarcinomas - proximal or lesser curve lesions accounted for two thirds of all the missed cancers.

Endoscopies were performed by several different endoscopists. Of the 111 adenocarcinomas diagnosed at the time of first endoscopy (10 patients with suspicious but inconclusive histology who were diagnosed within four weeks are included), 63(57%) endoscopies were performed by consultants, 41(37%) by middle graders (Specialist Registrar, Staff Grade or Research Fellow) and 7(6%) by other grades (General Practitioner or Specialist Nurse Endoscopist). Of the 18 missed adenocarcinomas at first endoscopy, 7(39%) endoscopies were performed by consultants and 11(61%) by middle graders.

Table 3: Pathology and site of the 17 missed gastric adenocarcinomas

DISCUSSION

Two important points emerge from this study. Firstly, a significant proportion of patients with gastric cancer were misdiagnosed at first endoscopy. Secondly, and not surprisingly, long delays can subsequently occur following initial misdiagnosis. The false negative rate of 22% (only 78% were diagnosed at first endoscopy) is similar to other reported data.^{4, 10, 11} In 10 patients, however, the need for obtaining a better representative biopsy sample was identified at the time of this first endoscopy resulting in minimal diagnostic delay. However, it is the 18(14%) patients who were misdiagnosed at first endoscopy resulting in a median diagnostic delay of 13 weeks (range 3-102) in whom there is most concern. As many of these patients will also have waited several weeks for their initial endoscopy, the overall delay will be much greater. As this study only reports that group of patients referred for a surgical opinion, the overall figure may also be higher.

Diagnostic error at endoscopy is multifactorial. It is well documented that gastric lesions, especially gastric ulcers, subsequently proven to be malignant can appear benign endoscopically and failure on part of the endoscopist to biopsy these seemingly benign abnormalities contributes to diagnostic error.^6-8 In order to reduce false-negative diagnoses, multiple biopsies are recommended and for gastric ulcers these should be obtained from both the rim and base.^6,7 Recommendations on the optimum number of biopsies differ but 10 or more biopsies may have to be taken as routine from all gastric ulcers to ensure high accuracy rates.⁷ Use of the powerful acid suppressing proton pump inhibitors can mask endoscopic signs of early gastric cancer resulting in false-negative diagnoses.¹¹ Inappropriate use of proton pump inhibitors would also contribute to diagnostic delay as rapid control of symptoms would lead the patient or general practitioner to defer referral for prompt endoscopy.¹³

In the present study, experience of the endoscopist may have been a factor in initial misdiagnosis - nearly two-thirds of the endoscopies for the missed cancers were performed by middle-graders. However, the most likely cause of missed diagnosis was a high threshold for biopsy. Abnormal findings were observed at endoscopy in seven patients in whom no biopsies or benign biopsies were obtained and a more aggressive biopsy policy may have prevented furthur delay. This was of particular concern in the three patients with endoscopically benign lesions which were not biopsied. These three patients suffered long diagnostic delays (18, 58 and 80 weeks) and were found to have advanced disease (Stage II, II and IV respectively) at diagnosis. Four patients with normal endoscopic findings were diagnosed with gastric adenocarcinoma within 2 years at repeat endoscopy (Table 2). At the time of diagnosis, three of these patients had advanced disease and one had stage I disease. Small and early gastric cancers can be hard to detect and it is possible that lesions were missed in these four patients during the initial endoscopy as the endoscopist was unaware of their appearance. It was difficult to reliably comment on the influence of anti-secretory drugs on missed diagnoses in this study as it was unclear from patient case records whether these drugs had been prescribed prior to initial endoscopy.

Present evidence suggests that the natural history of gastric cancer is that of a slow growing tumour which takes 3-4 years to progress from early to advanced disease.^14-16 It is essential that all lesions should be considered as potentially invasive if the number of missed diagnoses is to be reduced.¹⁷ Accurate data on doubling time for gastric cancer is scarce, but has been estimated at between 1.5-10 years for early, and 2-12 months for more advanced tumours.18,19 It can be argued, however, that these findings are a consequence of lead-time bias.^20,21 Stage I disease accounted for 7 of the 18 patients misdiagnosed at first endoscopy and since the diagnostic delay was less than 16 weeks in all six patients, this is unlikely to have influenced the chance of potential cure. However, seven patients were stage III or IV and diagnostic delay may have influenced survival in this group.

In conclusion, this study, although reporting small numbers, has confirmed that delays in establishing the diagnosis of gastric adenocarcinoma following initial endoscopy occur in a surprisingly large number of patients. Furthermore, greater suspicion and a more rigorous protocol for repeat endoscopy and biopsy must be implemented in order to reduce the number of missed diagnoses after initial endoscopy. These findings have significant implications for the delivery of endoscopic services, while at the same time highlighting potential problems in endoscopic training and supervision.

REFERENCES

1. Office of Population Censuses and Surveys. Cancer statistics. London HMSO,1986
2. Sue-Ling H, Martin I, Griffith J, Ward DC, Quirke P, Dixon MF, Axon AT, McMahon MJ, Johnston D. Early gastric cancer: 46 cases treated in one surgical department. Gut 1992; 33:1318-1322
3. Sue-Ling HM, Johnston D, Martin IG, Dixon MF, Landsdown MR, McMahon MJ, Axon AT. Gastric cancer: a curable disease in Britain. BMJ 1993; 307: 591-596
4. Hallissey M, Allum W, Jewkes A, Ellis DJ, Fielding JW. Early detection of gastric cancer. BMJ 1990; 301: 513-515
5. Axon AT, Bell GD, Jones RH, Quine MA, Mcloy RF. Guidelines for appropriate indications for upper gastrointestinal endoscopy. BMJ 1995; 310:853-856
6. Graham DY, Schwartz JT, Cain GD, Gyorkey.Prospective evaluation of biopsy number in the diagnosis of esophageal and gastric carcinoma. Gastroenterology 1982; 82:228-231
7. Dekker W, Tytgat G. Diagnostic accuracy of fiberendoscopy in the detection of upper intestinal malignancy. A follow up analysis. Gastroenterology 1977; 73:710-714
8. Witzel L, Halter F, Gretillat PA, Scheurer U, Keller M. Evaluation of specific value of endoscopic biopsies and brush cytology for malignancies of the oesophagus and stomach. Gut 1976; 17:375-377
9. Martin IG, Young S, Sue-Ling H, Johnston D. Delays in the diagnosis of oesophagogastric cancer: a consecutive case series. BMJ 1997; 314:467-470
10. Jorde R, Ostensen H, Bostad LH, Burhol PG, Langmark FT. Cancer detection in biopsy specimens taken from different types of gastric lesions. Cancer1986; 58:376-382
11. Suvakovic Z, Bramble MG, Jones R, Wilson C, Idle N, Ryott J. Improving the detection rate of early gastric cancer requires more than open access gastroscopy: a five year study. Gut 1997; 41:308-313
12. Solbin L, Wittekind C. Classification of malignant tumours.TNM 5th ed. Union Internationale Contre le cancer,1997
13. Griffin SM, Raimes SA. Proton pump inhibitors may mask early gastric cancer. BMJ 1998; 317:1606-1607 [editorial]
14. Tsukuma H, Mishima T, Oshima A. Prospective study of early gastric cancer. Int J Cancer 1983; 31:421-426
15. Tsukuma H, Oshima A, Narahara H, Morii T. Natural history of early gastric cancer: a non-concurrent, long-term, follow up study. Gut 2000;47:618-621
16. Eckardt VF, Willems D, Kanzler G, Remmele W, Bettendorf U, Paulus W. Eighty months persistence of poorly differentiated early gastric cancer. Gastroenterology 1984; 87:719-724
17. Everett S, Axon A. Early gastric cancer: disease or pseudodisease. Lancet 1998; 351:1350-1352
18. Haruma K, Tokutomi T, Nakanishi T, Sato T, Katsuta S, Tokumo K, Sumii K, Kajiyama G, Tahara E. Observations on the growth rate of polypoidal type gastric cancers with doubling times based on a retrospective roentgenological study. Gan No Rinsho 1988; 34:183-190
19. Kohli Y, Kawai K, Fujita S. Analytical studies on the growth of human gastric cancer. J Clin Gastroenterol 1981; 3:129-133
20. Schmidt J. Attempts to reduce delays in diagnosis should be done in a controlled trial. BMJ 1997; 315: 426-427 [letter]
21. Renehan A, Tweedle D. Misconceptions exist over whether delay in diagnosis influences survival. BMJ 1997; 315:427-428 [letter]

Copyright: 6 August 2002

Correspondence: T.J. Crofts, University Department of Surgery, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh, EH3 9YW