Search

Background: Use of the conventional Western and Japanese classification systems of gastrointestinal epithelial neoplasia results in large differences among pathologists in the diagnosis of oesophageal, gastric, and colorectal neoplastic lesions., Aim: To develop common worldwide terminology for gastrointestinal epithelial neoplasia., Methods: Thirty one pathologists from 12 countries reviewed 35 gastric, 20 colorectal, and 21 oesophageal biopsy and resection specimens. The extent of diagnostic agreement between those with Western and Japanese viewpoints was assessed by kappa statistics. The pathologists met in Vienna to discuss the results and to develop a new consensus terminology., Results: The large differences between the conventional Western and Japanese diagnoses were confirmed (percentage of specimens for which there was agreement and kappa values: 37% and 0.16 for gastric; 45% and 0.27 for colorectal; and 14% and 0.01 for oesophageal lesions). There was much better agreement among pathologists (71% and 0.55 for gastric; 65% and 0.47 for colorectal; and 62% and 0.31 for oesophageal lesions) when the original assessments of the specimens were regrouped into the categories of the proposed Vienna classification of gastrointestinal epithelial neoplasia: (1) negative for neoplasia/dysplasia, (2) indefinite for neoplasia/dysplasia, (3) non-invasive low grade neoplasia (low grade adenoma/dysplasia), (4) non-invasive high grade neoplasia (high grade adenoma/dysplasia, non-invasive carcinoma and suspicion of invasive carcinoma), and (5) invasive neoplasia (intramucosal carcinoma, submucosal carcinoma or beyond)., Conclusion: The differences between Western and Japanese pathologists in the diagnostic classification of gastrointestinal epithelial neoplastic lesions can be resolved largely by adopting the proposed terminology, which is based on cytological and architectural severity and invasion status.

The use of oral non-steroidal anti-inflammatory drugs (NSAIDs) in 31 patients with collagenous colitis and in 31 matched control patients with irritable bowel syndrome or colonic diverticular disease who had also undergone colonoscopy and biopsy was investigated. The long term use (greater than 6 months) of NSAIDs was significantly commoner in the study group (19/31) than in the control group (4/31) (p less than 0.02), even assuming the most adverse drug history in six patients in whom this could not be established. In all patients with collagenous colitis taking NSAIDs, diarrhoea followed the use of these drugs, and by a mean (SD) of 5.5 (4.4) years (range 0.5 to 15 years). In three patients with collagenous colitis, diarrhoea improved after withdrawing NSAIDs; rechallenge in one was followed by a recurrence of diarrhoea, which improved after withdrawing the drug again. It is suggested that NSAIDs may play an aetiological role in the diarrhoea and thickened collagen band in some patients with collagenous colitis.

We investigated the distribution of the collagen band in 33 patients with collagenous colitis to estimate the likelihood of the disease being diagnosed in biopsy specimens from the left side of the colon, such as those obtained using flexible sigmoidoscopy. To be included in this study patients had a subepithelial collagen band greater than or equal to 10 microns, an increase in chronic inflammatory cells in the same specimen, and diarrhoea for which there was no other apparent cause. In 17 patients undergoing full colonoscopy with a thickened collagen band, collagenous colitis was frequently patchy, even though overall the thickened collagen band was almost equally distributed throughout the colon. Rectal biopsy specimens showed a normal collagen band in 73% of patients, while a thickened collagen band was found in 82% of patients in at least one specimen from the left side of the colon. Three patients had a thickened collagen band only in the caecum. In three of eight rectal biopsy specimens with a normal collagen band there was no mucosal inflammation to raise the possibility of proximal disease, although all but one specimen with a normal collagen band from the sigmoid and descending colon were inflamed. Rectal biopsy alone is therefore a relatively poor method of making the diagnosis. Flexible sigmoidoscopy with multiple biopsy specimens from several sites is a reasonable initial investigation but not sufficient to exclude collagenous colitis when based on the presence of a thickened collagen band alone. Should left sided biopsy specimens show a normal collagen band but an inflamed mucosa, total colonoscopy with multiple specimens including the caecum may be required to establish the diagnosis.

A group of 18 patients with stable ulcerative colitis involving the entire colon for at least eight years was subjected to a biopsy of normal appearing rectal mucosa and followed prospectively over four years for the development of either dysplasia or cancer. Goblet cell glycoconjugate structure was examined in the rectal biopsies using fluorescein conjugated lectins. At the beginning of the study, 13 of the 18 patients had abnormalities of goblet cell mucin or cytoplasmic glycoconjugates in the rectal biopsies. Dysplasia subsequently developed in six and carcinoma in one of these patients. Among the five patients with normal lectin binding studies in the initial rectal biopsies, colonic dysplasia has subsequently developed in one. The abnormalities seen in the rectal goblet cells resembled in part those previously seen in immature and neoplastic colonic cells. The dysplastic tissues all contained the form of mucin which has been found in other neoplastic colonic tissues. This preliminary report after four years of prospective study suggests that abnormalities of glycoconjugate structure may be associated with, and may precede, neoplastic events in the setting of chronic ulcerative colitis.

Of 111 carcinomas developing in 73 patients with ulcerative colitis, 46 (41.5%) arose in the rectum where they are directly accessible to sigmoidoscopy. Fifty-eight per cent of single carcinomas developed in the rectum. The extent and frequency of rectal dysplasia was assessed by examining slides of rectal mucosa with an eyepiece micrometer. Slides from 46 patients with carcinoma and 22 patients with dysplasia but no carcinoma in whom proctectomy or proctocolectomy had been carried out were examined by this method. Thirteen of 15 patients with carcinoma of the colon (87%) and 21/22 patients (95%) with large bowel dysplasia showed evidence only of rectal dysplasia. However, there was marked variability in the proportion of dysplastic rectal mucosa even in those patients with rectal carcinoma, while in some patients dysplasia was limited to a small focus. Because of the possibility of false negative biopsies due to sampling error, multiple biopsies should be taken to detect dysplasia. Their state should be recorded and deliberately varied at subsequent visits. Careful sigmoidoscopy and multiple biopsies in this study had potential value as an aid in the detection of 85-90% of all carcinomas. In practice the figure would almost certainly be lower due to intrinsic bias (see discussion) so that, although regular sigmoidoscopy and biopsy would be of great value when colonoscopy is not available, the latter should be included in any long-term programme of carcinoma prevention.