- Email: [email protected]
Most dysplasia in ulcerative colitis is visible at colonoscopy
Most dysplasia in ulcerative colitis is visible at colonoscopy Matthew D. Rutter, MBBS, MRCP, Brian P. Saunders, MBBS, MD, MRCP, Kay H. Wilkinson, BA, BM, BCh, Michael A. Kamm, MBBS, MD, FRCP, FRACP, Christopher B. Williams, BM, FRCP, FRCS, Alastair Forbes, BSc MD, FRCP, ILTM London, United Kingdom
Background: Patients with long-standing extensive ulcerative colitis are at increased risk for colorectal carcinoma. Because most dysplasia is believed to be macroscopically invisible, recommended surveillance protocols include multiple non-targeted colonic biopsies. It was hypothesized by us that most dysplasia is actually colonoscopically visible. This study assessed the proportion of dysplasia that has been detected macroscopically in patients who underwent colonoscopy surveillance at our center. Methods: A retrospective review was conducted of colonoscopically detected neoplasia (dysplasia or cancer) in patients with ulcerative colitis who underwent surveillance from 1988 through 2002. An established surveillance protocol was used in all cases that included random segmental biopsies every 10 cm throughout the length of the colon, in addition to targeted biopsies of macroscopic lesions. Neoplasia detection was categorized as resulting from either targeted or nontargeted (‘‘random’’) biopsies. Follow-up information was obtained to the study end. Results: A total of 525 patients underwent 2204 surveillance colonoscopies. A total of 110 neoplastic areas were detected in 56 patients: 85 (77.3%) were macroscopically visible at colonoscopy, and 25 (22.7%) were macroscopically invisible. Fifty patients (89.3%) had macroscopically detectable neoplasia, and 6 (10.7%) had macroscopically invisible lesions. The frequency of cancer in patients who had endoscopic resection of neoplasia did not differ from that for the surveillance population as a whole (1/40 vs. 18/525; p = 1.0, Fisher exact test), irrespective of whether the lesion was thought to be an adenoma or a dysplasia-associated lesion/mass. Conversely, a high proportion of unresectable lesions harbored cancer. Conclusions: Most dysplastic lesions in ulcerative colitis are visible at colonoscopy. From a clinical perspective, the endoscopic resectability of a lesion is more important than whether it is thought to be a sporadic adenoma or a dysplasia-associated lesion/mass. (Gastrointest Endosc 2004;60: 334-9.)
Patients with long-standing extensive ulcerative colitis (UC) are at increased risk for the development of colorectal carcinoma. In many centers, colonoscopic surveillance of such patients is performed in an attempt to reduce the associated mortality. Colonoscopic surveillance relies on the detection of premalignant dysplastic tissue, either before the development of carcinoma or when the cancer is early stage with a more favorable prognosis. RecomReceived Janury 15, 2004. For revision April 6, 2004. Accepted May 7, 2004. Current affiliation: St Mark’s Hospital, Harrow, London, United Kingdom. Presented at the British Society of Gastroenterology Annual Conference, March 23-26, 2003, and Digestive Diseases Week, May 18-21, 2003, Orlando, Florida (Gastrointest Endosc 2003;57:AB215). Reprint requests: M.D. Rutter, Department of Gastroenterology, University Hospital of North Tees, Stockton-on-Tees, Teesside, TS19 8PE, UK. Copyright Ó 2004 by the American Society for Gastrointestinal Endoscopy 0016-5107/$30.00 PII: S0016-5107(04)01710-9 334
GASTROINTESTINAL ENDOSCOPY
mended surveillance protocols include multiple non-targeted (‘‘random’’) biopsies throughout the colon to detect dysplasia, based on historical evidence that dysplasia occurs in flat mucosa and as a widespread ‘‘field-effect’’ distant to cancer sites in 96% to 100% of cases.1,2 Guidelines promulgated in Britain and in the United States advise that two to 4 nontargeted biopsy specimens be taken for every 10 cm of colon and rectum.3,4 Although it is recognized that dysplasia can occur in mass lesions within colitic mucosa, the so-called dysplasia-associated lesion/ mass (DALM),5 it is still widely believed that most dysplasia is invisible at colonoscopy and can only be detected microscopically.6,7 Over recent years, our experience has been that much of the dysplasia detected during colonoscopic surveillance has been found within macroscopically visible lesions. It, therefore, was hypothesized that with the improvements in colonoscopic techniques and equipment, the majority of neoplastic lesions (either dysplasia or cancer) are macroscopically visible at colonoscopy. The aim of this study was to VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
ascertain the proportion of dysplasia that is macroscopically evident at surveillance colonoscopy. PATIENTS AND METHODS Surveillance program There is a well-established surveillance program at our hospital for patients with long-standing extensive UC.8,9 Patients with histopathologically proven UC with macroscopic inflammation proximal to the splenic flexure are counseled regarding their increased risk for colorectal cancer and are offered colonoscopic surveillance at 2-year intervals beginning 8 years after the onset of colitic symptoms. At colonoscopy, abnormal-appearing areas are described, biopsy specimens are taken, and the abnormality is photographed where possible. In addition, the segmental macroscopic appearance of the colon is documented, and segmental biopsy specimens are taken in all patients (8-12 biopsy specimens on average), with additional biopsy specimens from any mucosal irregularity. For the diagnosis of dysplasia in any biopsy specimen, two experienced pathologists must independently concur on the diagnosis, based on the standardized classification (1983) proposed by Riddell et al.10 Patients who have confirmed dysplasia (either low or high grade) are advised to have panproctocolectomy. Colonoscopic data are recorded in a computer database (Metabase; Metasa Ltd., London, U.K.), and pertinent histopathologic, colonoscopic, and clinical data related to each procedure and hospital consultation are recorded in an UC surveillance computer database (Oracle Corp., Redwood City, Calif.).
Study All instances of colonoscopically detected neoplasia (either dysplasia or cancer) in patients participating in the UC surveillance program between January 1, 1988 and January 1, 2002, were retrospectively reviewed. Cases of neoplasia were identified from the surveillance database. Details for each instance of neoplasia were obtained from the surveillance database, and all medical records, colonoscopy reports, and pathology reports were reviewed. When available, colonoscopic photographs also were studied. Cases in which biopsy specimens were indefinite for dysplasia, and those in which neoplasia was present in segments of the colon not currently or previously affected by UC were excluded from the study. By using all information available, one researcher (M.R.) categorized each neoplastic biopsy specimen as taken from either a colonoscopically abnormal mucosal area (‘‘targeted’’) or a mucosal area (‘‘non-targeted’’) in which the colonoscopic appearance did not raise any suspicion of malignancy. Neoplasia was assumed to be non-targeted unless there was clear evidence to the contrary. For visible dysplastic lesions that were documented as being present at more than one procedure, only the initial finding was included in the analysis to preclude double counting. Subanalyses were performed after excluding small (<10 mm) discretely demarcated mildly dysplastic tubular VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
Table 1. Patient demographics Median (range) Age at onset of colitic symptoms Age at first detection of dysplasia Duration of UC at first detection of dysplasia Number of surveillance colonoscopies
33 y (7-62) 56 y (27-81) 21 y (9-57) 5 (1-17)
UC, Ulcerative colitis.
adenomas, because these lesions are considered by some to be of lesser importance than other dysplastic lesions in UC5 and after excluding cancers. Follow-up data were obtained to colectomy, death, or study end (January 1, 2002). These data were obtained from medical records and, where necessary, from primary care physicians, other hospitals, and from the Office of National Statistics NHS Central Register, including cancer diagnosis and death certificate data.
RESULTS Patients Over the 14-year study period, 525 patients with UC underwent colonoscopic surveillance: 2204 colonoscopies were performed. The median surveillance interval was 2.19 years (interquartile range 1.832.45 years). Dysplasia (low or high grade) or cancer was detected colonoscopically in 56 patients (10.7%). These 56 patients (33 men, 23 women) comprised the study population (Table 1). The median age at onset of colitic symptoms for the 56 patients who developed neoplasia (either dysplasia or cancer) was significantly greater than that of the 469 patients in whom no neoplasia was detected (33 years vs. 28 years, respectively; p = 0.0002, 2-tailed unpaired t test). There was no difference in the median duration of colitis between the two groups (27.5 years for patients who developed neoplasia vs. 26 years for patients who did not; p = 0.9, 2-tailed unpaired t test). Neoplasia A total of 110 neoplastic areas (at 88 colonoscopies) were found in the study cohort. The distribution is depicted in Figure 1. Eighty-five neoplastic areas (77.3%) were macroscopically visible at colonoscopy, whereas 25 sites of neoplasia (22.7%) were macroscopically invisible at colonoscopy (in 13 patients). The diameter of the 85 visible lesions ranged from 1 to 40 mm (median 10 mm). At colonoscopy, 74 (87%) of the visible lesions were polypoid, 4 had an irregular outline, one was described as a plaque, and 6 were described as macroscopic cancers (one of which was confirmed histopathologically only in the colectomy specimen). Details of the colonoscopic and histopathologic findings are summarized by Figure 2. GASTROINTESTINAL ENDOSCOPY
335
M Rutter, B Saunders, K Wilkinson, et al.
Ulcerative colitis: visibility of dysplasia at colonoscopy
Table 2. Follow-up of dysplasia
Initial diagnosis of dysplasia Macroscopically invisible dysplasia Macroscopically visible dysplasia* Mildly dysplastic TA <10 mm
Figure 1. Distribution of neoplasia in colon.
No. patients
No. lesions
Follow-up (y)
No. cancers
50
104
183
6
12
24
37
31
55
115
4
17
25
56
1
2y
TA, Tubular adenoma. *Excluding mildly dysplastic tubular adenomas <10 mm in size. yIncludes one patient who also developed macroscopically visible dysplasia.
Figure 2. Details of neoplastic lesions. CRC, Colorectal cancer; HGD, high grade dysplasia; LGD, low grade dysplasia; DALM, dysplasia-associated lesion/mass; TA, tubular adenoma; TVA, tubulovillous adenoma; VA, villous adenoma.
Twenty-five lesions were considered endoscopically and histopathologically to be mildly dysplastic tubular adenomas less than 10 mm in diameter (17 patients). Excluding these and the 6 biopsyconfirmed cancers, there were 79 areas of dysplasia (37 patients, 63 colonoscopies), of which 55 (69.6%) were colonoscopically visible and 24 (30.4%) were colonoscopically invisible. Overall, 50 patients (89.3%) had macroscopically detectable neoplasia, whereas only 6 patients with neoplasia (10.7%) had macroscopically invisible lesions. A single neoplastic focus was detected in 72 colonoscopies (82%), two foci were detected in 12 colonoscopies (14%), and, in 4 colonoscopies (5%), more than two foci of neoplasia were found. Neoplasia was detected at more than one colonoscopy in 19 patients (34%). The background mucosa in the neoplastic segment of colon contained active inflammation in 34 cases (31%), chronic inflammation in 62 (56%), and no inflammation in 14 cases (13%).
patients, colorectal cancer was discovered in a further 6 (12.0%), after a total of 183 patient-years’ follow-up (mean 3.7 years; range 1 month to 12 years 3 months). Outcomes are shown in Table 2. Endoscopic resection of lesions Forty patients underwent endoscopic resection of the dysplastic lesion, which included 17 mildly dysplastic tubular adenomas less than 10 mm in diameter and 23 other macroscopically visible lesions. Over a total of 169 patient-years’ follow-up (mean 4.2 years), only one patient developed colorectal cancer (2.5 years after resection of a mildly dysplastic tubular adenoma). The frequency of cancer in patients who had endoscopic resection of the dysplastic lesion did not differ significantly from the overall cancer frequency within the surveillance population (1 of 40 patients [2.5%] vs. with 18 of 525 patients [3.4%]; p = 1.0, Fisher exact test). Colectomy for visible dysplasia
Follow-up The initial neoplastic lesion was a histopathologically confirmed cancer in 6 patients. Of the other 50 336
GASTROINTESTINAL ENDOSCOPY
In total, 8 patients with macroscopically visible dysplasia were referred for immediate colectomy. In 4, cancer was detected in the colectomy specimen. In VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
3 of the 4 cases, the cancer site was within the colonoscopically detected mass lesion, and, in the other case, the cancer was macroscopically visible to the surgeon and had presumably been missed at the previous colonoscopy. Background mucosal inflammation None of the 7 patients who developed dysplastic lesions within an area of noninflamed mucosa subsequently developed cancer over 23 patient-years’ follow-up (median 25 months, range 4-85 months). DISCUSSION Most dysplasia is visible Before the advent of colonoscopic surveillance, Morson and Pang1 remarked in 1967 that neoplastic changes occurred more commonly in flat instead of polypoid mucosa and that such changes could not be detected by macroscopic examination alone.1 Of note, the same study also found that dysplastic areas varied in macroscopic appearance from almost normal to obviously inflamed and that a thick mucosa with a finely nodular/velvety surface was the most common appearance. Beginning in the early 1970s, colonoscopy (which, unlike studies of colectomy specimens, allows an in vivo magnified inspection of the colonic mucosa) became the mainstay of surveillance for patients with UC, and, over the last 3 decades, colonoscopic imaging technology has improved greatly. However, it is still a widely held belief that dysplasia in UC is invisible at colonoscopy and only detectable in randomly obtained biopsy specimens,6,7,11 with less than 5% of dysplasia being mass related.6 Indeed, UC surveillance has been described as ‘‘an essentially random pursuit of endoscopically invisible neoplastic mucosal changes over a large surface area.’’12 This retrospective study of the detection of dysplasia in patients with UC over a period of 14 years shows that, even after excluding small tubular adenomatous lesions, most surveillance-detected dysplasia is colonoscopically visible. Almost 90% of the patients with dysplastic lesions had developed a colonoscopically visible abnormality that was detected during surveillance. Although the term ‘‘flat dysplasia’’ is not synonymous with colonoscopic invisibility, the results of the present study suggest that most visible lesions are polypoid instead of flat. Careful analysis of prior studies appears to support the finding that the majority of dysplastic lesions are mass related. In the only other study that specifically addressed the colonoscopic morphology of dysplastic lesions, Blackstone et al.5 observed that 12 of 39 dysplastic lesions were mass related (30.8%). VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
Of other studies that mention the presence of mass lesions, Rosenstock et al.13 reported that 13 of 22 dysplastic lesions detected by colonoscopic surveillance were mass related (59.1%),13 and LennardJones et al.,14 in a study of operative specimens, found that 11 of the 18 colectomy specimens containing either high-grade dysplasia or cancer had massrelated dysplastic lesions (61.1% of specimens). Another study from our hospital of colectomy specimens found that 20 of 28 colectomy specimens from patients with dysplasia had a total of 40 visible dysplastic lesions (71.4% of colectomy specimens).15 Even the original study of Morson and Pang1 of rectal biopsy specimens found that two of the 9 patients with dysplasia detected at rigid sigmoidoscopy had an associated polypoidal lesion (28.6%, with a further patient having a polypoidal lesion in the colectomy specimen). One possible explanation for the higher proportion of flat dysplastic lesions in other studies is that it can be extremely difficult to discriminate active inflammation from regenerating epithelium and dysplasia.10 In our center, no biopsy specimen is categorized as dysplastic unless two experienced GI histopathologists concur, and, consequently, the frequency of dysplasia in our patients has been lower than that reported from most centers. Another explanation could be that many supposedly ‘‘random’’ biopsies have actually been targeted to an abnormal area of mucosa by the colonoscopist without documentation of such in the colonoscopy report. Biases For this retrospective study, dysplastic lesions were categorized as being invisible (i.e., detected from non-targeted biopsy specimens) unless there was clear evidence to the contrary. It is possible that some of these biopsies were actually targeted but were undocumented as such at the time. Thus, it is likely that the number of dysplastic foci detected by targeted biopsy specimens is an underestimate, and the number of dysplastic foci detected by non-targeted biopsy specimens has been overestimated. In this respect, a prospective study would be needed to obtain more accurate data. Conversely, it is possible that the reason for the low frequency of dysplasia in non-targeted biopsy specimens in the current study is that too few nontargeted biopsy specimens were taken. Rubin et al.16 calculated that 33 biopsy specimens have to be examined to detect dysplasia with 90% confidence; but, historically, our policy has been to take approximately 8 to 12 non-targeted biopsy specimens. However, in a prospective trial from our center that assessed the value of quadrantic, non-targeted GASTROINTESTINAL ENDOSCOPY
337
M Rutter, B Saunders, K Wilkinson, et al.
biopsies every 10 cm along the colorectum, no dysplasia was detected in almost 3000 random biopsy specimens, thereby raising questions about the value of such biopsies.17 Again, a prospective study of the cost-effectiveness of any potential change in biopsy protocols would be useful, because this is a key issue for any surveillance program. With recent rapid advances in videoendoscopic instrument systems, improved endoscopic skills, and improved detection techniques, such as pan-colonic dye spraying, the proportion of lesions that are discovered macroscopically is likely to increase. It was beyond the limits of this retrospective study to assess the proportion of macroscopically abnormal areas from which biopsy specimens were obtained that were dysplastic. Similarly, the study did not assess the diagnosis of dysplasia during the colonoscopic procedure (as opposed to detection of dysplasia in colonoscopic biopsy specimens). Improved instrumentation, such as magnification endoscopy, might allow better discrimination between mucosal abnormalities during colonoscopy.18-20 With regard to the follow-up data presented, it must be noted that these data are censored, because many patients underwent surgery shortly after the diagnosis of dysplasia. This curtails the natural progression of dysplasia. It also introduces a bias, because the patients considered most likely to develop cancer were referred for immediate surgery and, therefore, cancer cannot develop. Also, the follow-up period varied, depending on the proximity of dysplasia detection to the study end point. DALM or sporadic adenoma? There is a continuing debate regarding the diagnosis and the significance of raised dysplastic lesions in UC. Perhaps the only non-contentious aspect of this question is the lesion proximal to the extent of colitic inflammation; these can safely be considered sporadic adenomas, dealt with endoscopically, and carry no increased risk of cancer long term.5,9 Thus, such lesions were excluded from the current analysis. Raised lesions within the distribution of colitic inflammation have been termed DALMs,5 and several studies have shown that there is a significant risk that these harbor cancer.9,21,22 However, it is now recognized that many such lesions can safely be removed endoscopically,23,24 and various attempts have been made to endoscopically or histopathologically categorize such lesions as ‘‘adenoma-like masses’’ or DALMs.25-27 It is felt by us that, at least from a clinical perspective, all such lesions can be considered part of the same spectrum, analogous to the spectrum 338
GASTROINTESTINAL ENDOSCOPY
Ulcerative colitis: visibility of dysplasia at colonoscopy
seen in sporadic adenomas, with small, solitary endoscopically resectable, low-grade dysplastic lesions having a low risk of malignancy and multiple, large, endoscopically unresectable high-grade dysplastic lesions carrying a high risk of malignant transformation and, more importantly, of actually harboring cancer.5,9 For these reasons, no attempt was made in the present study to retrospectively recategorize lesions as DALMs or adenomas. Although the clinical categorization at the time of diagnosis is given, the distinction between, e.g., a sessile adenoma within colitic mucosa and a sessile DALM often is arbitrary, and indeed all would be considered DALMs by the original definition. The data from the current study support this hypothesis: the cancer risk in patients who undergo complete endoscopic resection of a macroscopically visible dysplastic lesion is low and no different from that for a surveillance cohort overall. Conversely, of the 8 patients with macroscopically visible lesions for which endoscopic therapy was considered inappropriate and who, therefore, underwent immediate colectomy, 4 (50%) had colorectal cancer in the specimen and all but one of the cancers were detected within the lesion itself.
Clinical implications Contrary to the results of prior studies and popular opinion, the dysplasia detected during our surveillance program was visible at colonoscopy in the majority of cases. Because dysplasia was macroscopically invisible in a minority of cases, the value of non-targeted colorectal biopsies is lower than has been assumed. Nevertheless, in this retrospective review, there was no documentation of a colonoscopically visible lesion for 23% of dysplastic foci. Hence, a prospective assessment of the value of non-targeted colonic biopsies is required before it is possible to consider discontinuation of the non-targeted biopsy regimen. The results of the present study should encourage meticulous examination of the colon for mucosal abnormalities and indicate that, after consideration of other risk factors, endoscopic therapy may be appropriate for a larger proportion of patients with dysplastic lesions. In our view, the most important criterion in managing visible dysplasia is whether the lesion is endoscopically resectable. Lesions that are not resectable require urgent surgery because of the high frequency of cancer in such lesions. However, lesions that can be entirely removed by endoscopic resection confer a good prognosis, regardless of whether the lesion has been termed a DALM or a sporadic adenoma. VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
REFERENCES 1. Morson BC, Pang LS. Rectal biopsy as an aid to cancer control in ulcerative colitis. Gut 1967;8:423-34. 2. Hulten L, Kewenter J, Ahren C. Precancer and carcinoma in chronic ulcerative colitis. A histopathological and clinical investigation. Scand J Gastroenterol 1972;7:663-9. 3. The role of colonoscopy in the management of patients with inflammatory bowel disease. American Society for Gastrointestinal Endoscopy. Gastrointest Endosc 1998;48: 689-90. 4. Eaden JA, Mayberry JF. Guidelines for screening and surveillance of asymptomatic colorectal cancer in patients with inflammatory bowel disease. Gut 2002;51(Suppl 5): V10-2. 5. Blackstone MO, Riddell RH, Rogers BH, Levin B. Dysplasiaassociated lesion or mass (DALM) detected by colonoscopy in long-standing ulcerative colitis: an indication for colectomy. Gastroenterology 1981;80:366-74. 6. Tytgat GN, Dhir V, Gopinath N. Endoscopic appearance of dysplasia and cancer in inflammatory bowel disease [review]. Eur J Cancer 1995;31A:1174-7. 7. Sharan R, Schoen RE. Cancer in inflammatory bowel disease. An evidence-based analysis and guide for physicians and patients. Gastroenterol Clin North Am 2002;31:237-54. 8. Lennard-Jones JE, Morson BC, Ritchie JK, Williams CB. Cancer surveillance in ulcerative colitis. Experience over 15 years. Lancet 1983;II:149-52. 9. Connell WR, Lennard-Jones JE, Williams CB, Talbot IC, Price AB, Wilkinson KH. Factors affecting the outcome of endoscopic surveillance for cancer in ulcerative colitis [see comments]. Gastroenterology 1994;107:934-44. 10. Riddell RH, Goldman H, Ransohoff DF, Appelman HD, Fenoglio CM, Haggitt RC, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14:931-68. 11. Lewis JD, Deren JJ, Lichtenstein GR. Cancer risk in patients with inflammatory bowel disease [review]. Gastroenterol Clin North Am 1999;28:459-77. 12. Bernstein CN. The case against dysplasia surveillance in ulcerative colitis. In: McLeod RS, Martin F, Sutherland LR, Wallace JL, Williams CN, editors. Trends in inflammatory bowel disease therapy. Dordrecht (Netherlands): Kluwer; 1996. p. 268-76. 13. Rosenstock E, Farmer RG, Petras R, Sivak MVJ, Rankin GB, Sullivan BH. Surveillance for colonic carcinoma in ulcerative colitis. Gastroenterology 1985;89:1342-6.
VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
14. Lennard-Jones JE, Melville DM, Morson BC, Ritchie JK, Williams CB. Precancer and cancer in extensive ulcerative colitis: findings among 401 patients over 22 years. Gut 1990;31:800-6. 15. Butt JH, Konishi F, Morson BC, Lennard-Jones JE, Ritchie JK. Macroscopic lesions in dysplasia and carcinoma complicating ulcerative colitis. Dig Dis Sci 1983;28:18-26. 16. Rubin CE, Haggitt RC, Burmer GC, Brentnall TA, Stevens AC, Levine DS, et al. DNA aneuploidy in colonic biopsies predicts future development of dysplasia in ulcerative colitis. Gastroenterology 1992;103:1611-20. 17. Rutter MD, Saunders BP, Schofield G, Forbes A, Price AB, Talbot IC. Pancolonic indigo carmine dye spraying for the detection of dysplasia in ulcerative colitis. Gut 2004;53: 256-60. 18. Axelrad AM, Fleischer DE, Geller AJ, Nguyen CC, Lewis JH, Al Kawas FH, et al. High-resolution chromoendoscopy for the diagnosis of diminutive colon polyps: implications for colon cancer screening. Gastroenterology 1996;110:1253-8. 19. Kim CY, Fleischer DE. Colonic chromoscopy. A new perspective on polyps and flat adenomas [review]. Gastrointest Endosc Clin N Am 1997;7:423-37. 20. Tada M, Katoh S, Kohli Y, Kawai K. On the dye spraying method in colonofiberscopy. Endoscopy 1977;8:70-4. 21. Blackstone MO. DALM in ulcerative colitis cancer surveillance [letter]. Gastroenterology 1986;91:266-7. 22. Bernstein CN, Shanahan F, Weinstein WM. Are we telling patients the truth about surveillance colonoscopy in ulcerative colitis? [review; see comments]. Lancet 1994;343:71-4. 23. Rubin PH, Friedman S, Harpaz N, Goldstein E, Weiser J, Schiller J, et al. Colonoscopic polypectomy in chronic colitis: conservative management after endoscopic resection of dysplastic polyps. Gastroenterology 1999;117:1295-300. 24. Engelsgjerd M, Farraye FA, Odze RD. Polypectomy may be adequate treatment for adenoma-like dysplastic lesions in chronic ulcerative colitis [see comments]. Gastroenterology 1999;117:1288-94. Discussion, 1488-91. 25. Odze RD. Adenomas and adenoma-like DALMs in chronic ulcerative colitis: a clinical, pathological, and molecular review. Am J Gastroenterol 1999;94:1746-50. 26. Bernstein CN. ALMs versus DALMs in ulcerative colitis: polypectomy or colectomy? Gastroenterology 1999;117:1488-92. 27. Torres C, Antonioli D, Odze RD. Polypoid dysplasia and adenomas in inflammatory bowel disease: a clinical, pathologic, and follow-up study of 89 polyps from 59 patients. Am J Surg Pathol 1998;22:275-84.
GASTROINTESTINAL ENDOSCOPY
339
Background: Patients with long-standing extensive ulcerative colitis are at increased risk for colorectal carcinoma. Because most dysplasia is believed to be macroscopically invisible, recommended surveillance protocols include multiple non-targeted colonic biopsies. It was hypothesized by us that most dysplasia is actually colonoscopically visible. This study assessed the proportion of dysplasia that has been detected macroscopically in patients who underwent colonoscopy surveillance at our center. Methods: A retrospective review was conducted of colonoscopically detected neoplasia (dysplasia or cancer) in patients with ulcerative colitis who underwent surveillance from 1988 through 2002. An established surveillance protocol was used in all cases that included random segmental biopsies every 10 cm throughout the length of the colon, in addition to targeted biopsies of macroscopic lesions. Neoplasia detection was categorized as resulting from either targeted or nontargeted (‘‘random’’) biopsies. Follow-up information was obtained to the study end. Results: A total of 525 patients underwent 2204 surveillance colonoscopies. A total of 110 neoplastic areas were detected in 56 patients: 85 (77.3%) were macroscopically visible at colonoscopy, and 25 (22.7%) were macroscopically invisible. Fifty patients (89.3%) had macroscopically detectable neoplasia, and 6 (10.7%) had macroscopically invisible lesions. The frequency of cancer in patients who had endoscopic resection of neoplasia did not differ from that for the surveillance population as a whole (1/40 vs. 18/525; p = 1.0, Fisher exact test), irrespective of whether the lesion was thought to be an adenoma or a dysplasia-associated lesion/mass. Conversely, a high proportion of unresectable lesions harbored cancer. Conclusions: Most dysplastic lesions in ulcerative colitis are visible at colonoscopy. From a clinical perspective, the endoscopic resectability of a lesion is more important than whether it is thought to be a sporadic adenoma or a dysplasia-associated lesion/mass. (Gastrointest Endosc 2004;60: 334-9.)
Patients with long-standing extensive ulcerative colitis (UC) are at increased risk for the development of colorectal carcinoma. In many centers, colonoscopic surveillance of such patients is performed in an attempt to reduce the associated mortality. Colonoscopic surveillance relies on the detection of premalignant dysplastic tissue, either before the development of carcinoma or when the cancer is early stage with a more favorable prognosis. RecomReceived Janury 15, 2004. For revision April 6, 2004. Accepted May 7, 2004. Current affiliation: St Mark’s Hospital, Harrow, London, United Kingdom. Presented at the British Society of Gastroenterology Annual Conference, March 23-26, 2003, and Digestive Diseases Week, May 18-21, 2003, Orlando, Florida (Gastrointest Endosc 2003;57:AB215). Reprint requests: M.D. Rutter, Department of Gastroenterology, University Hospital of North Tees, Stockton-on-Tees, Teesside, TS19 8PE, UK. Copyright Ó 2004 by the American Society for Gastrointestinal Endoscopy 0016-5107/$30.00 PII: S0016-5107(04)01710-9 334
GASTROINTESTINAL ENDOSCOPY
mended surveillance protocols include multiple non-targeted (‘‘random’’) biopsies throughout the colon to detect dysplasia, based on historical evidence that dysplasia occurs in flat mucosa and as a widespread ‘‘field-effect’’ distant to cancer sites in 96% to 100% of cases.1,2 Guidelines promulgated in Britain and in the United States advise that two to 4 nontargeted biopsy specimens be taken for every 10 cm of colon and rectum.3,4 Although it is recognized that dysplasia can occur in mass lesions within colitic mucosa, the so-called dysplasia-associated lesion/ mass (DALM),5 it is still widely believed that most dysplasia is invisible at colonoscopy and can only be detected microscopically.6,7 Over recent years, our experience has been that much of the dysplasia detected during colonoscopic surveillance has been found within macroscopically visible lesions. It, therefore, was hypothesized that with the improvements in colonoscopic techniques and equipment, the majority of neoplastic lesions (either dysplasia or cancer) are macroscopically visible at colonoscopy. The aim of this study was to VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
ascertain the proportion of dysplasia that is macroscopically evident at surveillance colonoscopy. PATIENTS AND METHODS Surveillance program There is a well-established surveillance program at our hospital for patients with long-standing extensive UC.8,9 Patients with histopathologically proven UC with macroscopic inflammation proximal to the splenic flexure are counseled regarding their increased risk for colorectal cancer and are offered colonoscopic surveillance at 2-year intervals beginning 8 years after the onset of colitic symptoms. At colonoscopy, abnormal-appearing areas are described, biopsy specimens are taken, and the abnormality is photographed where possible. In addition, the segmental macroscopic appearance of the colon is documented, and segmental biopsy specimens are taken in all patients (8-12 biopsy specimens on average), with additional biopsy specimens from any mucosal irregularity. For the diagnosis of dysplasia in any biopsy specimen, two experienced pathologists must independently concur on the diagnosis, based on the standardized classification (1983) proposed by Riddell et al.10 Patients who have confirmed dysplasia (either low or high grade) are advised to have panproctocolectomy. Colonoscopic data are recorded in a computer database (Metabase; Metasa Ltd., London, U.K.), and pertinent histopathologic, colonoscopic, and clinical data related to each procedure and hospital consultation are recorded in an UC surveillance computer database (Oracle Corp., Redwood City, Calif.).
Study All instances of colonoscopically detected neoplasia (either dysplasia or cancer) in patients participating in the UC surveillance program between January 1, 1988 and January 1, 2002, were retrospectively reviewed. Cases of neoplasia were identified from the surveillance database. Details for each instance of neoplasia were obtained from the surveillance database, and all medical records, colonoscopy reports, and pathology reports were reviewed. When available, colonoscopic photographs also were studied. Cases in which biopsy specimens were indefinite for dysplasia, and those in which neoplasia was present in segments of the colon not currently or previously affected by UC were excluded from the study. By using all information available, one researcher (M.R.) categorized each neoplastic biopsy specimen as taken from either a colonoscopically abnormal mucosal area (‘‘targeted’’) or a mucosal area (‘‘non-targeted’’) in which the colonoscopic appearance did not raise any suspicion of malignancy. Neoplasia was assumed to be non-targeted unless there was clear evidence to the contrary. For visible dysplastic lesions that were documented as being present at more than one procedure, only the initial finding was included in the analysis to preclude double counting. Subanalyses were performed after excluding small (<10 mm) discretely demarcated mildly dysplastic tubular VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
Table 1. Patient demographics Median (range) Age at onset of colitic symptoms Age at first detection of dysplasia Duration of UC at first detection of dysplasia Number of surveillance colonoscopies
33 y (7-62) 56 y (27-81) 21 y (9-57) 5 (1-17)
UC, Ulcerative colitis.
adenomas, because these lesions are considered by some to be of lesser importance than other dysplastic lesions in UC5 and after excluding cancers. Follow-up data were obtained to colectomy, death, or study end (January 1, 2002). These data were obtained from medical records and, where necessary, from primary care physicians, other hospitals, and from the Office of National Statistics NHS Central Register, including cancer diagnosis and death certificate data.
RESULTS Patients Over the 14-year study period, 525 patients with UC underwent colonoscopic surveillance: 2204 colonoscopies were performed. The median surveillance interval was 2.19 years (interquartile range 1.832.45 years). Dysplasia (low or high grade) or cancer was detected colonoscopically in 56 patients (10.7%). These 56 patients (33 men, 23 women) comprised the study population (Table 1). The median age at onset of colitic symptoms for the 56 patients who developed neoplasia (either dysplasia or cancer) was significantly greater than that of the 469 patients in whom no neoplasia was detected (33 years vs. 28 years, respectively; p = 0.0002, 2-tailed unpaired t test). There was no difference in the median duration of colitis between the two groups (27.5 years for patients who developed neoplasia vs. 26 years for patients who did not; p = 0.9, 2-tailed unpaired t test). Neoplasia A total of 110 neoplastic areas (at 88 colonoscopies) were found in the study cohort. The distribution is depicted in Figure 1. Eighty-five neoplastic areas (77.3%) were macroscopically visible at colonoscopy, whereas 25 sites of neoplasia (22.7%) were macroscopically invisible at colonoscopy (in 13 patients). The diameter of the 85 visible lesions ranged from 1 to 40 mm (median 10 mm). At colonoscopy, 74 (87%) of the visible lesions were polypoid, 4 had an irregular outline, one was described as a plaque, and 6 were described as macroscopic cancers (one of which was confirmed histopathologically only in the colectomy specimen). Details of the colonoscopic and histopathologic findings are summarized by Figure 2. GASTROINTESTINAL ENDOSCOPY
335
M Rutter, B Saunders, K Wilkinson, et al.
Ulcerative colitis: visibility of dysplasia at colonoscopy
Table 2. Follow-up of dysplasia
Initial diagnosis of dysplasia Macroscopically invisible dysplasia Macroscopically visible dysplasia* Mildly dysplastic TA <10 mm
Figure 1. Distribution of neoplasia in colon.
No. patients
No. lesions
Follow-up (y)
No. cancers
50
104
183
6
12
24
37
31
55
115
4
17
25
56
1
2y
TA, Tubular adenoma. *Excluding mildly dysplastic tubular adenomas <10 mm in size. yIncludes one patient who also developed macroscopically visible dysplasia.
Figure 2. Details of neoplastic lesions. CRC, Colorectal cancer; HGD, high grade dysplasia; LGD, low grade dysplasia; DALM, dysplasia-associated lesion/mass; TA, tubular adenoma; TVA, tubulovillous adenoma; VA, villous adenoma.
Twenty-five lesions were considered endoscopically and histopathologically to be mildly dysplastic tubular adenomas less than 10 mm in diameter (17 patients). Excluding these and the 6 biopsyconfirmed cancers, there were 79 areas of dysplasia (37 patients, 63 colonoscopies), of which 55 (69.6%) were colonoscopically visible and 24 (30.4%) were colonoscopically invisible. Overall, 50 patients (89.3%) had macroscopically detectable neoplasia, whereas only 6 patients with neoplasia (10.7%) had macroscopically invisible lesions. A single neoplastic focus was detected in 72 colonoscopies (82%), two foci were detected in 12 colonoscopies (14%), and, in 4 colonoscopies (5%), more than two foci of neoplasia were found. Neoplasia was detected at more than one colonoscopy in 19 patients (34%). The background mucosa in the neoplastic segment of colon contained active inflammation in 34 cases (31%), chronic inflammation in 62 (56%), and no inflammation in 14 cases (13%).
patients, colorectal cancer was discovered in a further 6 (12.0%), after a total of 183 patient-years’ follow-up (mean 3.7 years; range 1 month to 12 years 3 months). Outcomes are shown in Table 2. Endoscopic resection of lesions Forty patients underwent endoscopic resection of the dysplastic lesion, which included 17 mildly dysplastic tubular adenomas less than 10 mm in diameter and 23 other macroscopically visible lesions. Over a total of 169 patient-years’ follow-up (mean 4.2 years), only one patient developed colorectal cancer (2.5 years after resection of a mildly dysplastic tubular adenoma). The frequency of cancer in patients who had endoscopic resection of the dysplastic lesion did not differ significantly from the overall cancer frequency within the surveillance population (1 of 40 patients [2.5%] vs. with 18 of 525 patients [3.4%]; p = 1.0, Fisher exact test). Colectomy for visible dysplasia
Follow-up The initial neoplastic lesion was a histopathologically confirmed cancer in 6 patients. Of the other 50 336
GASTROINTESTINAL ENDOSCOPY
In total, 8 patients with macroscopically visible dysplasia were referred for immediate colectomy. In 4, cancer was detected in the colectomy specimen. In VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
3 of the 4 cases, the cancer site was within the colonoscopically detected mass lesion, and, in the other case, the cancer was macroscopically visible to the surgeon and had presumably been missed at the previous colonoscopy. Background mucosal inflammation None of the 7 patients who developed dysplastic lesions within an area of noninflamed mucosa subsequently developed cancer over 23 patient-years’ follow-up (median 25 months, range 4-85 months). DISCUSSION Most dysplasia is visible Before the advent of colonoscopic surveillance, Morson and Pang1 remarked in 1967 that neoplastic changes occurred more commonly in flat instead of polypoid mucosa and that such changes could not be detected by macroscopic examination alone.1 Of note, the same study also found that dysplastic areas varied in macroscopic appearance from almost normal to obviously inflamed and that a thick mucosa with a finely nodular/velvety surface was the most common appearance. Beginning in the early 1970s, colonoscopy (which, unlike studies of colectomy specimens, allows an in vivo magnified inspection of the colonic mucosa) became the mainstay of surveillance for patients with UC, and, over the last 3 decades, colonoscopic imaging technology has improved greatly. However, it is still a widely held belief that dysplasia in UC is invisible at colonoscopy and only detectable in randomly obtained biopsy specimens,6,7,11 with less than 5% of dysplasia being mass related.6 Indeed, UC surveillance has been described as ‘‘an essentially random pursuit of endoscopically invisible neoplastic mucosal changes over a large surface area.’’12 This retrospective study of the detection of dysplasia in patients with UC over a period of 14 years shows that, even after excluding small tubular adenomatous lesions, most surveillance-detected dysplasia is colonoscopically visible. Almost 90% of the patients with dysplastic lesions had developed a colonoscopically visible abnormality that was detected during surveillance. Although the term ‘‘flat dysplasia’’ is not synonymous with colonoscopic invisibility, the results of the present study suggest that most visible lesions are polypoid instead of flat. Careful analysis of prior studies appears to support the finding that the majority of dysplastic lesions are mass related. In the only other study that specifically addressed the colonoscopic morphology of dysplastic lesions, Blackstone et al.5 observed that 12 of 39 dysplastic lesions were mass related (30.8%). VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
Of other studies that mention the presence of mass lesions, Rosenstock et al.13 reported that 13 of 22 dysplastic lesions detected by colonoscopic surveillance were mass related (59.1%),13 and LennardJones et al.,14 in a study of operative specimens, found that 11 of the 18 colectomy specimens containing either high-grade dysplasia or cancer had massrelated dysplastic lesions (61.1% of specimens). Another study from our hospital of colectomy specimens found that 20 of 28 colectomy specimens from patients with dysplasia had a total of 40 visible dysplastic lesions (71.4% of colectomy specimens).15 Even the original study of Morson and Pang1 of rectal biopsy specimens found that two of the 9 patients with dysplasia detected at rigid sigmoidoscopy had an associated polypoidal lesion (28.6%, with a further patient having a polypoidal lesion in the colectomy specimen). One possible explanation for the higher proportion of flat dysplastic lesions in other studies is that it can be extremely difficult to discriminate active inflammation from regenerating epithelium and dysplasia.10 In our center, no biopsy specimen is categorized as dysplastic unless two experienced GI histopathologists concur, and, consequently, the frequency of dysplasia in our patients has been lower than that reported from most centers. Another explanation could be that many supposedly ‘‘random’’ biopsies have actually been targeted to an abnormal area of mucosa by the colonoscopist without documentation of such in the colonoscopy report. Biases For this retrospective study, dysplastic lesions were categorized as being invisible (i.e., detected from non-targeted biopsy specimens) unless there was clear evidence to the contrary. It is possible that some of these biopsies were actually targeted but were undocumented as such at the time. Thus, it is likely that the number of dysplastic foci detected by targeted biopsy specimens is an underestimate, and the number of dysplastic foci detected by non-targeted biopsy specimens has been overestimated. In this respect, a prospective study would be needed to obtain more accurate data. Conversely, it is possible that the reason for the low frequency of dysplasia in non-targeted biopsy specimens in the current study is that too few nontargeted biopsy specimens were taken. Rubin et al.16 calculated that 33 biopsy specimens have to be examined to detect dysplasia with 90% confidence; but, historically, our policy has been to take approximately 8 to 12 non-targeted biopsy specimens. However, in a prospective trial from our center that assessed the value of quadrantic, non-targeted GASTROINTESTINAL ENDOSCOPY
337
M Rutter, B Saunders, K Wilkinson, et al.
biopsies every 10 cm along the colorectum, no dysplasia was detected in almost 3000 random biopsy specimens, thereby raising questions about the value of such biopsies.17 Again, a prospective study of the cost-effectiveness of any potential change in biopsy protocols would be useful, because this is a key issue for any surveillance program. With recent rapid advances in videoendoscopic instrument systems, improved endoscopic skills, and improved detection techniques, such as pan-colonic dye spraying, the proportion of lesions that are discovered macroscopically is likely to increase. It was beyond the limits of this retrospective study to assess the proportion of macroscopically abnormal areas from which biopsy specimens were obtained that were dysplastic. Similarly, the study did not assess the diagnosis of dysplasia during the colonoscopic procedure (as opposed to detection of dysplasia in colonoscopic biopsy specimens). Improved instrumentation, such as magnification endoscopy, might allow better discrimination between mucosal abnormalities during colonoscopy.18-20 With regard to the follow-up data presented, it must be noted that these data are censored, because many patients underwent surgery shortly after the diagnosis of dysplasia. This curtails the natural progression of dysplasia. It also introduces a bias, because the patients considered most likely to develop cancer were referred for immediate surgery and, therefore, cancer cannot develop. Also, the follow-up period varied, depending on the proximity of dysplasia detection to the study end point. DALM or sporadic adenoma? There is a continuing debate regarding the diagnosis and the significance of raised dysplastic lesions in UC. Perhaps the only non-contentious aspect of this question is the lesion proximal to the extent of colitic inflammation; these can safely be considered sporadic adenomas, dealt with endoscopically, and carry no increased risk of cancer long term.5,9 Thus, such lesions were excluded from the current analysis. Raised lesions within the distribution of colitic inflammation have been termed DALMs,5 and several studies have shown that there is a significant risk that these harbor cancer.9,21,22 However, it is now recognized that many such lesions can safely be removed endoscopically,23,24 and various attempts have been made to endoscopically or histopathologically categorize such lesions as ‘‘adenoma-like masses’’ or DALMs.25-27 It is felt by us that, at least from a clinical perspective, all such lesions can be considered part of the same spectrum, analogous to the spectrum 338
GASTROINTESTINAL ENDOSCOPY
Ulcerative colitis: visibility of dysplasia at colonoscopy
seen in sporadic adenomas, with small, solitary endoscopically resectable, low-grade dysplastic lesions having a low risk of malignancy and multiple, large, endoscopically unresectable high-grade dysplastic lesions carrying a high risk of malignant transformation and, more importantly, of actually harboring cancer.5,9 For these reasons, no attempt was made in the present study to retrospectively recategorize lesions as DALMs or adenomas. Although the clinical categorization at the time of diagnosis is given, the distinction between, e.g., a sessile adenoma within colitic mucosa and a sessile DALM often is arbitrary, and indeed all would be considered DALMs by the original definition. The data from the current study support this hypothesis: the cancer risk in patients who undergo complete endoscopic resection of a macroscopically visible dysplastic lesion is low and no different from that for a surveillance cohort overall. Conversely, of the 8 patients with macroscopically visible lesions for which endoscopic therapy was considered inappropriate and who, therefore, underwent immediate colectomy, 4 (50%) had colorectal cancer in the specimen and all but one of the cancers were detected within the lesion itself.
Clinical implications Contrary to the results of prior studies and popular opinion, the dysplasia detected during our surveillance program was visible at colonoscopy in the majority of cases. Because dysplasia was macroscopically invisible in a minority of cases, the value of non-targeted colorectal biopsies is lower than has been assumed. Nevertheless, in this retrospective review, there was no documentation of a colonoscopically visible lesion for 23% of dysplastic foci. Hence, a prospective assessment of the value of non-targeted colonic biopsies is required before it is possible to consider discontinuation of the non-targeted biopsy regimen. The results of the present study should encourage meticulous examination of the colon for mucosal abnormalities and indicate that, after consideration of other risk factors, endoscopic therapy may be appropriate for a larger proportion of patients with dysplastic lesions. In our view, the most important criterion in managing visible dysplasia is whether the lesion is endoscopically resectable. Lesions that are not resectable require urgent surgery because of the high frequency of cancer in such lesions. However, lesions that can be entirely removed by endoscopic resection confer a good prognosis, regardless of whether the lesion has been termed a DALM or a sporadic adenoma. VOLUME 60, NO. 3, 2004
Ulcerative colitis: visibility of dysplasia at colonoscopy
REFERENCES 1. Morson BC, Pang LS. Rectal biopsy as an aid to cancer control in ulcerative colitis. Gut 1967;8:423-34. 2. Hulten L, Kewenter J, Ahren C. Precancer and carcinoma in chronic ulcerative colitis. A histopathological and clinical investigation. Scand J Gastroenterol 1972;7:663-9. 3. The role of colonoscopy in the management of patients with inflammatory bowel disease. American Society for Gastrointestinal Endoscopy. Gastrointest Endosc 1998;48: 689-90. 4. Eaden JA, Mayberry JF. Guidelines for screening and surveillance of asymptomatic colorectal cancer in patients with inflammatory bowel disease. Gut 2002;51(Suppl 5): V10-2. 5. Blackstone MO, Riddell RH, Rogers BH, Levin B. Dysplasiaassociated lesion or mass (DALM) detected by colonoscopy in long-standing ulcerative colitis: an indication for colectomy. Gastroenterology 1981;80:366-74. 6. Tytgat GN, Dhir V, Gopinath N. Endoscopic appearance of dysplasia and cancer in inflammatory bowel disease [review]. Eur J Cancer 1995;31A:1174-7. 7. Sharan R, Schoen RE. Cancer in inflammatory bowel disease. An evidence-based analysis and guide for physicians and patients. Gastroenterol Clin North Am 2002;31:237-54. 8. Lennard-Jones JE, Morson BC, Ritchie JK, Williams CB. Cancer surveillance in ulcerative colitis. Experience over 15 years. Lancet 1983;II:149-52. 9. Connell WR, Lennard-Jones JE, Williams CB, Talbot IC, Price AB, Wilkinson KH. Factors affecting the outcome of endoscopic surveillance for cancer in ulcerative colitis [see comments]. Gastroenterology 1994;107:934-44. 10. Riddell RH, Goldman H, Ransohoff DF, Appelman HD, Fenoglio CM, Haggitt RC, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14:931-68. 11. Lewis JD, Deren JJ, Lichtenstein GR. Cancer risk in patients with inflammatory bowel disease [review]. Gastroenterol Clin North Am 1999;28:459-77. 12. Bernstein CN. The case against dysplasia surveillance in ulcerative colitis. In: McLeod RS, Martin F, Sutherland LR, Wallace JL, Williams CN, editors. Trends in inflammatory bowel disease therapy. Dordrecht (Netherlands): Kluwer; 1996. p. 268-76. 13. Rosenstock E, Farmer RG, Petras R, Sivak MVJ, Rankin GB, Sullivan BH. Surveillance for colonic carcinoma in ulcerative colitis. Gastroenterology 1985;89:1342-6.
VOLUME 60, NO. 3, 2004
M Rutter, B Saunders, K Wilkinson, et al.
14. Lennard-Jones JE, Melville DM, Morson BC, Ritchie JK, Williams CB. Precancer and cancer in extensive ulcerative colitis: findings among 401 patients over 22 years. Gut 1990;31:800-6. 15. Butt JH, Konishi F, Morson BC, Lennard-Jones JE, Ritchie JK. Macroscopic lesions in dysplasia and carcinoma complicating ulcerative colitis. Dig Dis Sci 1983;28:18-26. 16. Rubin CE, Haggitt RC, Burmer GC, Brentnall TA, Stevens AC, Levine DS, et al. DNA aneuploidy in colonic biopsies predicts future development of dysplasia in ulcerative colitis. Gastroenterology 1992;103:1611-20. 17. Rutter MD, Saunders BP, Schofield G, Forbes A, Price AB, Talbot IC. Pancolonic indigo carmine dye spraying for the detection of dysplasia in ulcerative colitis. Gut 2004;53: 256-60. 18. Axelrad AM, Fleischer DE, Geller AJ, Nguyen CC, Lewis JH, Al Kawas FH, et al. High-resolution chromoendoscopy for the diagnosis of diminutive colon polyps: implications for colon cancer screening. Gastroenterology 1996;110:1253-8. 19. Kim CY, Fleischer DE. Colonic chromoscopy. A new perspective on polyps and flat adenomas [review]. Gastrointest Endosc Clin N Am 1997;7:423-37. 20. Tada M, Katoh S, Kohli Y, Kawai K. On the dye spraying method in colonofiberscopy. Endoscopy 1977;8:70-4. 21. Blackstone MO. DALM in ulcerative colitis cancer surveillance [letter]. Gastroenterology 1986;91:266-7. 22. Bernstein CN, Shanahan F, Weinstein WM. Are we telling patients the truth about surveillance colonoscopy in ulcerative colitis? [review; see comments]. Lancet 1994;343:71-4. 23. Rubin PH, Friedman S, Harpaz N, Goldstein E, Weiser J, Schiller J, et al. Colonoscopic polypectomy in chronic colitis: conservative management after endoscopic resection of dysplastic polyps. Gastroenterology 1999;117:1295-300. 24. Engelsgjerd M, Farraye FA, Odze RD. Polypectomy may be adequate treatment for adenoma-like dysplastic lesions in chronic ulcerative colitis [see comments]. Gastroenterology 1999;117:1288-94. Discussion, 1488-91. 25. Odze RD. Adenomas and adenoma-like DALMs in chronic ulcerative colitis: a clinical, pathological, and molecular review. Am J Gastroenterol 1999;94:1746-50. 26. Bernstein CN. ALMs versus DALMs in ulcerative colitis: polypectomy or colectomy? Gastroenterology 1999;117:1488-92. 27. Torres C, Antonioli D, Odze RD. Polypoid dysplasia and adenomas in inflammatory bowel disease: a clinical, pathologic, and follow-up study of 89 polyps from 59 patients. Am J Surg Pathol 1998;22:275-84.
GASTROINTESTINAL ENDOSCOPY
339