Applications and pitfalls of endoscopy in inflatnrnatory bowel disease

The role of upper gastrointestinal and hepatobiliary endoscopy and colonoscopy in the diagnosis and management of patients with inflammatory bowel disease (IBD) is reviewed. The differentiation of IBD from other forms of colitis, mass lesions in the colon, strictures and polyps, and the management of the patient with dysplasia including clinical strategies for early detection are discussed. The role and value of endoscopic surveillance programs have yet to be defined. Can J Gastroenterol 1990;4(7):324-330


E NOOSCOPY WITH MULTIPLE
mucosa! biopsies and the potential for therapeutic intervention is becoming increasingly employed in the diagnosis and assessment of patients with inflammatory bowel disease (IBD).Traditionally, endoscopy has been confined co sigmoidoscopy and colonoscopy, but it is also important co consider upper gastrointestinal endoscopy, especially in patients with Crohn's disease, and to perform endoscopic retrograde cholangiopancreatography in patients suspected of having hepacobi I iary complications of either ulcerative col iris or Crohn's disease.This paper wi ll review some of the common issues and pitfalls in the endoscopic diagnosis and management of IBD.ENDOSCOPY IN IBO Upper gastrointestinal endoscopy and mucosa!biopsy may be vaiuablefor the diagnosis of IBO.Crohn's disease may involve multiple sites throughout the gastrointestinal tract.Involvement of the esophagus, stomach or duo, denum tends to occur with well docu, mented small or large bowel disease.Rarely, Crohn's disease may arise solely at a site in the upper gastrointestinal tract.

Division of Gasr:roenterology, Mc Master University Medical
Endoscopically, the mucosa!surfact is granular or nodular with friability, erosions and aphthous ulcers.()c.casion ally, linear ulceration is seen in severe disease, with lack of disten, sibility, inflammation and later duo, denal obstruction.Rarely, a duodeno, colic fistula may be encountered, although these arc usually missed en, doscopically and may only be demonstrated on barium enema.Le~ common still may be an entero, cutaneous fistula from the duodenum.
The incidence and prevalence d.
upper gastrointestinal C rohn's disease are difficult to determine because most cases are poorly documented; series do not represent the population at large, and were reported before diagnostic criteria were well standardized or en, doscopy widely available.Three retro, spective studies suggest a prevalence of between 0.4 and 3% (l -3 ).However, more recent endoscopic reports based on histological criteria suggest that the figures are probably higher.
In an endoscopic study of 45 consecutively hospitalized patients with Crohn's disease and a nondiagnostic barium meal, Korelitz et al ( 4) found 42% of P.atienrs co have abnormal histology.Twenty-four per cenr had biopsies diagnostic and 18% suggestive of Crohn's disease, with granulomas in 7%.The antrum and duodenum were most frequently positive, with the esophagus least often involved; histology was abnormal in 24% of endoscopically normal sites.This study is difficult to evaluate, since hospitalized patients were selected as having no radiological abnormality; furthermore, controls were taken from a different geographic population, with endoscopy and histology interpreted by several different clinicians.
In a prospective study, Schmitz-Moorman et al ( 5) evaluated 225 patients with Crohn's disease of the small or large bowel by upper gastrointestinal endoscopy and histology.
They had 30 control subjects, 11 of whom had ulcerative colitis.Endoscopic lesions were found in the gastric antrum in 49% and in the duodenum in 34%, dnd histology was more sen sitive for detecting abnormalities, although the site of biopsies was nm predetermined and may have been influe nced by the endoscopist.G ranulomas were observed in less than one-third of patients but were more common in younger patients with a shorter duration of disease.
Jouin et al ( 6) performed e ndoscopy on 129 patients from a cohort of 195 patients with Crohn's disease, more commonly finding small superficial aphthous ulcers or patchy erythematous areas commonly in the antrum and proximal duodenum.Lesions were found in 28% of patients, and granulomas in 16%, with granulomas as frequent in the upper as in the lowe r gastrointestinal tract.
Malchow et al (7) prospectively evaluated 1414 unselected biopsies obtained during 509 gastroscopies for concordance between endoscopic and histological diagnosis.Agreement was best for histological changes in the duodenum with a kappa of90 falling co 0.12 for biopsies from the gastric body and to 0.1 0 for biopsies from the antrum.Thus, clinically significant disease is overlooked when endoscopy alone is used, and Crohn's disease may be missed altogether.A recent study has compared upper gastrointestinal endoscopic and histological findings in Crohn's patients and control groups of ulcerative colitis and dyspepsia patients (8).There was complete or partial agreement between endoscopy and histology in 85% of all patients, with some overlap in ulcerative colitis and Crohn's disease possibly due in part to drug treatment.There was evidence of Crohn's disease of the upper gastrointestinal tract in 60% of patients with known Crohn's disease.

HEPATOBILIARY DISORDERS INIBD
Sclerosing cholangitis is a progressive, ulumatcly fatal, chronic hepatobiliary disease most commonly occurring in young men, presenting with a chronic cholestatic syndrome and usually associated with chronic 180 (9)(10)(11)(12)(13).Most patients present with jaundice, abdomi nal pain a nd pruritus, although patients may be asymptomatic with only a disturbance of liver function tests.
Endoscopic retrograde cholan giography may be the most sensitive and specific diagnostic test m establishing a diagnosis of sclerosing c ho langitis.The association of sclerosing cholangitis with IBO varies from 4 7% in one pediatric series (13) to JOO% in an adult series from Scandinavia, which reported 4 5 cases of sclerosing cholangitis, 37 patients having ulcerative colitis, six C rohn 's disease, and two unspecified colitis ( 12).
In one study, 50% of patients had pruritus, 66% right upper quadrant pain, 33% fever, jaundice, hepatomega ly and weight loss, and 16% splenomegaly (13).Alkaline phosphatase and gammaglutamyl transferase were elevated in all patients, with some increase in both aspartate aminotransferase and alanine aminotransferase, but only half the patients had an elevation of bilirubin.All patients had abnormal extra-and/or intrahepatic bile ducts on cholangiography.
Endoscopy In IBD Of 681 patients with ulceraLive colitis in Oxford, 2 I (1%) had persistently ahnormal liver function tests, and 17 of these were found by cholangiography to have sclero~ing cholangitis (9).Liver biopsy showed histological changes in only half the patients, making endoscopic retrograde cholangiography necessary to confirm the diagnosis.
C hanges in the mtrahepatic bile ducts include thin-walled tubular and saccular cho langiectases and annular fibrous crests with or without abscess; and fibrous c holangitis with or without ductal dilation and replacement of bile ducts by fibrous cords ( 14).The 'pruned tree' appearance on cholangiography represents t he transitton from patent to cholangiectatic ducts.Radiographic findings at cholangiography include changes in both the extra-and intrahepalic ducts, although different senes report different proportions involved.The most important features included diminished arbo rization, ectasia and stenosis (15,16).Strictures may be multifocal involving either the extra-or intrahepatic ducts, and in advanced disease, confluent stricturing with long connected segments may be found ( 16).Pancreatic abnormalities: T he association of IBO and sclerosing c holangitis with pancreatic abnormalities is controversial.A Scandinavian study of 151 patients with ulcerative colitis found seven with sclerosing cholangitis and four with an abnormal pancreatogram.The changes were attributed to either a common immune mechanism or a mec hanical effect (17).MacCarty et al ( 16) obse rv ed va rious degrees o f pancreatic duct stri cture in 8% of patients, while m a further Scandinavian study endoscopic retrograde pancreacography was normal in all 38 patients with sclerosing cholangitis ( 18).

Carcinoma of the bile ducts:
Longstanding biliary inflammation is associated with carcinoma of the bile duct (usually affecting the extrahepatic bile duccs) ( 19).Tumours may be glandular, cystic or papillary, and localized or infiltrating.In a series from the Cleveland Clinic, six bile duct cancers were found in 1207 cases of ulcerative colitis for a prevnlence of 0.5% and a relative rbk of 31. 3 (20).Patients had a mean age of 38.5 years at the time of Jiagnosis, and a mean Juration of ulcerative colitis of 23.2 years.Colectomy had been undertaken between five anJ 16 years before, thus suggesting no protection from progression of the sclerosing cholangiris and development of malignancy.
There 1s a vanahle temporal relat1onsh1p in the presentation of IBO anJ hepatobiliary disease, with either occurring after the onset of the index disease.In patients with hepatob1liary disease of obscure origin, it is important to question the presence of antecedent bowel symptoms, to unJercake sigmoidoscopy and obtain biopsies, and to consider endoscopic retrograde cholangiopancreatography (21 ).Furthermore, sclerosing cholangitis may be asymptomattc (22).In the study reported by Aadland et al (12), 24 patients were followed over a 10 year period, and cholangiograph1c changes progresseJ from mild to moderate in only three, while in the remainmg 21 patients there wa no change.However, it is important to appreciate that chol-angiograph1c changes may worsen without any change in symptoms, laboratory parameters or liver biopsy (22).
Endoscopic retrograde cholangiography has maJe the diagnosis of sclerosing cholangitis easier and more specific, and it should be consiJered in any patient with persistently abnormal liver function tests.

COLONOSCOPYINIBD
The diagnosis of IBO of the colon is established in the majority of patients on the basis of clinical findings, sigmoidoscopy, rectal biopsy and double contrast barium enema.By no means do all patients require colonoscopy.In most instances colonoscopy is the last step in a succession of diagnostic examinations performed during evaluation of a prohlem in patients with colitis.
While double contrast barium enema has remained the mainstay of diagnosis for IBO, few comparative studies have been undertaken.In a survey of 149 patients with IBO, 23 ( 15%) were.:-cons1Jcred to have total colitis on ban um enema; 51 ( 34%) haJ visual appearances consistent with rocal colitis at colonoscopy; and 92 (62%) had the same on histology.Furthermore, radiological assessment of skip lesions was unreliable, hut it was not clear whether the differences were clinically relevant (23 ).
Most stuJ1es indicate that between 18 and 20% of patients with Crohn's disease have normal radiographic findings hut cndnscop1cally detectable disease, whic:.h is similar to studies detcrminmg the extent of disease in ulcerative colitis (24).Most false negatives are associated with mild disease at proccosigmo1doscopy.Although radiographic studies appear less sensitive than endoscopy for diagnosis, their abilities to discriminate between Crohn's disease and ulcerative colitis are comparable, with accuracies between 95 and 98% (24).

INDICATIONS FOR COLONOSCOPYINIBD
Colonoscopy is usually reserved to evaluate the extent of disease, smce that information is rarely essential in determining the therapeutic approach to the patient.Moreover, colonoscopy durmg acute disease is contraindicated due to the increased risk of perforation.When difficulty arises in the interpretation of a barium enema, colonoscopy is valuable especially for evaluation of strictures or mass lesions.In patients with a diagnosis of nonspecific colitis, colonoscopy and biopsy can be crucial to distinguish between ulcerative colitis, Crohn's disease and a wide range of other inflammatory condmons.The colonic epithelium responds to inflammation in a limited number of ways, and incraluminal appearances are rarely definitive.The correct diagnosis is usually obtained from the h1stopathological assessment of multiple biopsies obtained during colonoscopic examination, combined with clinical and enJoscop1c observations.Colonoscopy may be especially useful in the pre-and postoperative evaluation of Crohn's disease of the large bowel, since radiology cannot provide accurate information after sur-g1cal interventton.Furthermore, moM surgeons are unwilling to make an anastomosis through howcl affecteJ by active disease and request an accur.nepreoperative Jclineation of the 1.hsease before plannmg surgery.Postoperat11"e• ly an ileostomy stoma can usually l"< intubated with a standard colonoscopr or a pediatric instrument and examma-t1on of a cont ment 1leostomy (Kock pouch) can s11nilarly he inspt'Lte<l m patients with pouch Jysfunct11.m 0< suspected pouch ms.Lastly, colonoscopic surveillance is undertaken ~ most gastroenterolog1sts on ,1 regular annual or biannual basis m patienti with total ulcc.:-rativccolitis of eight or more years' Jurat1on to search I • dysplasia and prevent the development of cancer.

DIFFERENTIAL DIAGNOSIS OF CROHN'S DISEASE VERSUS ULCERATIVE COLITIS
On occasion it may be extremek difficult to distmgu1sh between ulcer.itivecolitis and Crohn':.disease, and colonoscopic examinanon of the ent1ri colon with intubation of the terminal ileum may be necessary, together wi1h multiple biopsies taken every 10 cm throughout the colon.
The nonnal mucosa of the rectum and colon is smooth, pale, gl1stenm~ and transparent, varying from grey through pink in the colnn with a bill( tone at the splenic anJ hepatic flexuro where the spleen or liver impress on the colon wall.The vascular pattern show a branching network of superficial ve.1, sels which arc most prominent in th, rectum and increase in Jiameter distal.ly.No bleeding or pus is seen under normal circumstances, and mucus L' rarely present even after bowel prepara, t1on, when mild erythema or edem: may be seen.Fnabilicy 1s not normalli present even after bowel preparation. Multiple biopsies provide the pathologist with an adequate number of samples to differentiate microscor1 cally discontinuous mucosa!mvolve ment from a pattern of increasin2 severity distally, which is seen in ul cerativc colitis.The finJmg d granulomas on colonoscopic biopsiesu clearly helpful, but they occur relative-ly infrequently, in about 5% of cases (25)(26)(27).
Endoscopic patterns of colonic inflammation are well described.Despite the different endoscopic appearances between Crohn's disease and ulcerative colitis, differentiation can be difficult.Endoscopic biopsy may assist with the correct diagnosis.Crohn's disease involves the mucosa and submucosa while ulcerative colitis primarily affects rhe mucosa!layer.Early in the course of Crohn's disease, the surface vascular pattern tends to remain intact with the earliest lesion being a small aphthous ulcer, usually 3 to 4 mm in diameter and often surrounded by a narrow red border of erythemacous mucosa.Cobblestoning is characteristic of submucosal involvement in Crohn's disease.This is the uniform nodularity caused by submucosal edema.Cobblestoning must be distinguished from multiple pseudopolyps in IBO.The length of pseudopolyps is characteristically greater than the widLh of their base with adjacent mucosa usually being flat.The low nodules of cobblestoning are distinctly contiguous.
The earliest manifestation of ulcerative colitis is an increase in mucosa!bloodflow that appears as a diffuse erythema with micro-aneurysmal changes endoscopically.The vascular architecture is often lost due to edema.Other features include granularity and friability due to mucosa!engorgement.As inflammation progresses, minute surface ulcerations occur.Spontaneous bleeding is characteristic of ulcerative colitis.
Several endoscopic features differentiate ulcerative colitis from Crohn's disease (28).Ulcers never occur in ulcerative colitis in an area of otherwise normal mucosa.Ulcers may occur in diffusely abnormal mucosa in both forms of colitis, but if the surrounding mucosa is normal, the diagnosis is never ulcerative colitis.Aphthous ulcers are pathognomonic of Crohn's colitis.Cobblestoning is pathognomonic of Crohn's colitis.Granularity and friability are common early in ulcerative colitis, but may be late findings in Crohn's colitis.
Pera et al ( 29) evaluated 606 colonoscopies in 357 patients with ulcerative colitis, Crohn's disease or indeterminate colitis.Patients were followed for 22 months to obtain endoscopy-independent diagnoses by histology, surgery or autopsy, and an endoscopic score and likelihood ratios were calculated.Colonoscopy had an accuracy of 89% with 4% errors and 7% indeterminate, which occurred predominantly in the presence of severe inflammation.Discontinuous disease, anal lesions and cobblestoning were predictive of Crohn's disease, while erosions, micro-ulceration and granularity were most predictive for ulcerative colitis.

OTHER FORMS OF COLITIS
Other forms of acute and chronic colitis may mimic chronic idiopathic IBO but can often be identified with a careful histary, colonoscopic examination and biopsy (30).
Studies of patients with mucoid, bloody diarrhea suspected of having idiopathic IBO show that between 22 and 38% have some fonn of infectious colitis or another inflammatory colonic disorder (31)(32)(33)(34)).An increased awareness and selective culture media will improve diagnosis.Examination of the ileum is valuable in one-third of patients (35) and should include microbiological examination of biopsies, which may be the only way to confirm yersinia, campylobacter or chlamydia!infection (33).Multiple biopsies taken in the tenninal ileum and every LO cm throughout the colon are needed to confirm microscopic colitis (36, 3 7) or collage nous colitis (36,38).Perioperative colonoscopy: Colonoscopy either before or after surgery is most useful in patients with Crohn's disease, which may affect any part of the gastrointestinal tract.It is important to determine the extent of the disease prior to undertaking segmental resection.In the presence of an ileocolic fistula, it may be possible to determine whether the colon is primarily involved or has been affected secondary to the close proximity of diseased ileum.Usually adjacent segments of colon are normal, and such information may modify the surgical plan (39).Postoperative indications for colonoscopy, with possible examination of an ileostomy stoma or continent ileostomy pouch, include recurrence of diarrhea.
Direct inspection of the colon following ileocolonic resection and anastomosis is best undertaken by colonoscopy, since high quality barium studies are difficult to obtain in the postoperative bowel due to changes in anatomy, spasm of the affected segment ( which prevents good air contrast) and rapid transit of barium.Small aphthous ulcers typical of Crohn's recurrence may be readily seen and biopsied.
The stoma of an ileostomy may be intubated endoscopically to determine whether ileostomy dysfunction is due to adhesions, stricture with partial or intermittent obstruction, or recurrent disease.A further important indication for endoscopic evaluation through an ileostomy includes the patient with anemia or frank bleeding, when recurrent disease may be identified more proximally.The Kock pouch of a continent ileostomy may become inflamed with pouchitis (40) and can be similarly examined to confirm inflammation or the cause of incontinence or difficul, ty with intubation (32).Incontinence of the pouch may result from fistula formation or loss of the nipple, which can reaJily be seen at endoscopy, as may any recurrence of Crohn's disease.A pediatric gastroscope with its shorter bending section and acute angulation or the oblique-viewing endoscope is especially useful for obtaining full visualization of the pouch.

MASS LESIONS
Inflammatory polyps: Mass lesions seen on barium enema may represent an adenomatous polyp, a non-neoplastic pseudopolyp or carcinoma, and require evaluation by colonoscopy.To the endoscopist, pseudopolyps are most frequently small and multiple, occurring throughout the colon, and are found in both ulcerative colitis and Crohn's disease.They appear to be covered by glistening, essentially normal mucosa similar to that surrounJing them.Inflammatory polyps represent regenerative islands of epithelium not involved in the surrounding destructive ulceration, although they are sometimes composed of granulation tissue ( 41).Occasionally they may be large, almost occluding the lumen; they rarely cause intussusception (42).Although pseudopolyps have no malignant potential, when large and solitary, they may be confused with carcinoma ( 43,44 ).If the lesion is solitary or appears different from the surrounding mucosa or other inflammatory polyps, it should be removed for histological appraisal.At polypectomy, care should be taken to ensure effective coagulation, since healing of the polypectomy site may be delayed in the presence of inflammatory disease.Biopsies should be taken from pseudopolyps that are larger than I cm in diameter, have an irregular surface configuranon or are a different colour from surrounding lesions.Polypectomy should be considered only for lesions which may be confused with carcinoma or if a large inflammatory polyp is considered to be responsible for symptoms of obstruction or hemorrhage (28).;\denomatous polyps: Adenomatous polyps occur uncommonly in patients who have ulcerative colitis, and usually resemble pseudopolyps rather than typical adenomas.In a series of 150 coli tics examined by colonoscopy, four (3%) were found to have adenomatous poiyps, of which three were solitary (45).If a diagnosis of adenoma is made on simple biopsy, the lesion should be removed by snare polypectomy as in a noncolitic colon.Strictures: Strictures in ulcerative colitis are seen with a frequency approximately equal to that seen in patients with Crohn's disease (38).Although a stricture encountered in ulcerative colitis should suggest the possibility of malignancy, most result from fibrosis, muscular hypertrophy or spasm, and colonoscopy provides the best means to evaluate strictures, since a direct inspection of the segment may be undertaken and multiple biopsies and cytological samples obtained (46).At colonoscopy the instrument may be passed up to and often through the stricture to determine its precise nature, and strictures previously seen on barium enema may be easily intubated by the adult colonoscope, since air insufflation will distend the narrow lumen when narrowing is due to muscular hypertrophy or spasm (47).Inflammatory strictures are characterized by mucosa!erythema, friability and ulceration, while fibrotic strictures appear thin, short and weblike.If a stricture is unyielding and rigid to the colonoscope, has an abrupt or shelf-like margin and cannot be intubated by the colonoscope, malignancy should be suspected (48).A visual diagnosis alone is not sufficient; it must be supported by multiple biopsies at the edge of and within the strictured segment.Brush cytology may also be useful if the lesion cannot be clearly intubated.Since carcinoma in colitis can extend submucosally, giving negative superficial biopsies (49), lesions which appear endoscopically malignant should always be referred for resection.If a stricture cannot be intubated with the adult colonoscope, a pediatric colonoscope or slim upper gastrointestinal endoscope may be used.

DYSPLASIA AND CANCER IN COLITIS
The recognition of dysplasia as a marker of high nsk of cancer has made a screening program possible in patients with longstanding ulcerative colitis (50).However, the frequency of examination is not clear, and there is as yet no evidence that such a program has significantly altered the course and outcome of the disease.
Currently the pathological staging of patients with carcinoma is purely a matter of chance; some cancers are detected relatively early and fortunately cured, while a similar patient may have disseminated disease and die early.The hope is implicit that early detection of invasive carcinoma will result in an increased cure rate and not just increased survival time, which could be the result of lead time bias.To date no trial has shown the benefit of detecting dysplasia or presymptomatic carcinoma in a surveillance program to prevent deaths from carcinoma; however, such studies may never be carried out because of the potential difficulties in randomizing patients with any form of 'premalignancy' to the appropriate control arm of a randomized clinical trial.
Clinical strategies for early detection of dysplasia and carcinoma are limited, but it is important to distinguish whether the search is primarily for a premalignant -as opposed to an early invasive -lesion.If the endpoint of clinical surveillance studies is to detect lesions endoscopically, for example, many will prove to be invasive carcinoma and will be associated with a definite mortality from disseminated carcinoma.This unexpected advanced carcinoma is less likely when one ~ searching for premalignanc lesions, since invasive and advanced tumour1 will be found less frequently.Macroscopic appearance of dysplasia: At endoscopy, areas of dysplasia may be seen raised above adjacent mucosa as plaques or irregular areas of nodularitl which are often poorly circumscribed, in contrast to the well circumscribed adenoma.
These differences in appearance have three distinct clinical implications.First, when surveillance is being undertaken to detect dysplasia, the possible gross appearances of the lesion being sought must be borne in mmd and sought deliberately, since their sub, tie nature may be readily overlooked.
Second, dysplasia has the potential to be, or actually is, the superficial pan of an invasive carcinoma.Unless or until the lesion is removed, it remai11.1 uncertain whether this change has actually occurred.
Third, the question of whether adenomas can exist or coexist with dysplasia needs to be considered.In ulcerative colitis, lesions indistinguishable from adenomas by all criteria appear to be relatively common, and endoscopic resection of these lesions II! if they were simple adenomas appea11 safe and unassociated with an excess of carcinoma in the lesion or the remainder of the colon.
The last major difference between the diagnosis of dysplasia and adenoma is in the effect that it has on the clinician.There may be a relative ignorance on the part of both the clinician and the pathologist.Given an endoscopic biopsy showing features of dysplasia, th e patho logist has the option of calling the lesion an aJenom a o r dysplasia .If there is little or no clinical information, o r if the patho logist is unfamiliar with the con cept of the dysplasia-assoc iated lesion o r mass ( DALM) (51 ), the lesion will be reported as an adenoma; if the c linician is unfamiliar with the DALM conc ept , the lesion may be treated as an aJ en o ma.If the pathologist makes a d iagnosis of dysplasia, th e clinic ian must immediately conside r wheth er colectomy is the appropriate man agement, since biopsies of dysplasia and adeno ma sh are a common feat ure -n amely that an associated invasive lesion cannot be excluded unless t he lesio n is removed.Implications for patient management: If the policy is to fo llo w low grade dysplasia, it should be recognized that on e is deliberately fo llo wing a lesion which can give rise directly to invasive carcinoma a nd may therefore a lready have realized that potential.Furthe r, th ese carcinomas may no t be readily recognized clinically in pa tients with IBO.
There is a distinctio n between a first 'diagnostic' anJ subsequent 'surve illance' endoscopy if d ysplasia is found.It is likely to be much more serio us when dysplasia is discovered on a first rather than a subsequen t endoscopy because in th e fo rmer the len gth of t ime chat dysplas ia h as been present is unknown and is pro bably more likely to be accompanied by an underlying invasive compo nent (5 2).
Endoscopy and diagnosis of dysplasia: Proc cos igmo idoscopy prov ides a limited view localized to th e rectum and the opportunity to obtain rectal biops ies.R ectal bio psies have been widely advocated fo r fo llow ing pa ti e nts w it h c h ron ic ulc era tive colitis, bu t mo re recent studies have suggested that as many as 75% of patients with colonic dysp lasia do n o t h ave recta l involvement (53,54 ).Colon oscopy provides th e best opportunity to cake multiple b iopsies and obtain m ore representative sampling fro m the who le colo n to seek for dysplas ia.All suspicio us areas sh ould be bio psied , especia lly areas of mucosa!irregulari ty where th e surface appears 'velvety' in appearan ce or an y mass lesion is present.Oysplasia assoc iated with such a lesion increases the proba bili t y o f ca rc ino m a s ignifi cantly ( 51,55 ).Inflammatory polyps and strictures have the ir own rules.Clinical strategies for early detection: Follow-up strategies or optio ns fall into o n e of fo ur ca tegories: do n o thing; regular fo llow-up and bio psy; fo llow-up a nd b iopsy at sh ore intervals; an d exc ision of the d iseased organ .

Endoscopy in 180
The biopsy classification de termined by an internationa l wo rk ing party can be used to guide pa tient management (56).When the biopsy class ification is n ega tive or indefin ite, it is probably negative, and the patient sh o uld continue with regular 12 m onthly fo llow-up; wh en t he interpretatio n is unclear but probably positive, a short interva l fo llow-up of three to six months sho uld be recommended , a nd when low grade dysplasia is confirmed, three mo n t hly fo llow-up must be undertaken and colectomy con sidered if a DALM is p resent.W hen high grade dysplas ia is reported, colecto my sh o uld be unde rtake n , altho ugh som e would advocate con firmation of h igh grade dysplasia before aJvocating surgery.
Doing n othing can be ad voca ted in patients in who m furthe r surgery wo uld not he contemplated even if anything was found, and regular fo llow-up is carried o ut in patients witho ut evidence of J ysplasia.
T h e role and value of endoscopic surveillan ce programs has yet to be J e fined.Advances in endoscopy and the development of n ew techniques such as high magnification endoscopy and laser spectroscopy may provide further sensit ivity and cost effectiveness in can cer detection in the future.