The Cancer Risk in Longstanding Ulcerative Colitis: Surveillance Colonoscopy and Prophylactic Surgery

(1)

Introduction

Colorectal cancer is far from being the most common complication of ulcerative colitis (UC), but it is the one that has been most extensively studied, and the literature on the subject exceeds that dealing with other complications of the disease. Many groups of workers have investigated the complex relationship between ulcerative proctocolitis and cancer of the large bowel.

Incidence

The colitic patient faces a higher risk of large-bowel cancer than does a member of the general population. Opinions as to how large this risk is differ great- ly between different authors. This is hardly surpris- ing, however, as there have been differences in the method of collection of materials, and the duration of the colitis has not always been taken into account [1, 2]. In facing the problem of cancer development in longstanding ulcerative colitis, many important issues have to be considered.

Are all patients subject to the same risk from large-bowel cancer, or can any one group of patients be singled out as being especially at risk? If such a group can be identified, at what stage of their disease are they particularly liable to develop the cancer? If the patient’s risk group can be defined with reason- able precision, is it possible to predict the sort of cancer risk which an individual patient faces? How should these patients be followed up and how should the surveillance programme be designed? Another crucial issue is whether the risk of cancer is sufficient to warrant the use of prophylactic surgery, and if so, when should it be recommended, and with all options present today, what form should the prophylactic surgery take?

The cumulative probability of cancer for all patients with ulcerative colitis, regardless of disease extent, has been demonstrated to be 2 % at 10 years, 8% at 20 years and 18% at 30 years, while the overall

prevalence of colorectal cancer in any patient has been demonstrated to be 3–4% [3]. However, there is convincing evidence that there are colitis patients who have an especially high risk of developing cancer. Investigations into the relationship between cancer and colitis have focussed upon the identification of a “high-risk” group of patients and the one single factor which does exert the most influence regarding cancer risk in colitic patients concerns the extent of the patient’s colitis. It seem to be quite clear that patients who are especially at risk for development of carcinoma are those with total or near-total involvement of the colon (i.e. involvement of the whole left half of the colon and the transverse colon as well).

There is indeed a remarkable unanimity of opinion on this point between the various workers who have studied the problem [4, 5, 6]. There is, additionally, convincing evidence to show that the duration of colitic symptoms has a strong impact on the incidence of cancer in those high-risk patients. In other words, the longer the history of colitis, the greater the risk of carcinoma.

An idea of the risk of cancer facing the individual patient with total involvement can be gained from the many studies that have been undertaken over the years. Remarkably little attention has been paid to the variation in the duration of the disease between different patients when the incidence of cancer in UC has been calculated. In published reports, the crude incidence is often used. However, the use of crude incidence to evaluate the risk of carcinoma in UC is misleading, as it fails to correct for differences in duration between individual patients. The figures may thus underestimate the true incidence of carcinoma in UC. The time factor is essential and by employing the life-table concept much more sophis- ticated and accurate results can be obtained. While there is evidence to show that the cancer risk for a patient with total involvement is low during the first 10 years of colitic symptoms, being around 0.4% per annum, the annual risk of cancer starts to rise sharply after this initial decade, and after 20 years of colitis, the individual with total involvement faces an

The Cancer Risk in Longstanding Ulcerative Colitis:

Surveillance Colonoscopy and Prophylactic Surgery

Leif Hultén

(2)

annual risk of cancer of over 5% for the remainder of his life.

However, the yearly risks accumulate, so that the cumulative risk during a period of 5, 10, 20 or even 25 years is far higher. The cumulative risk of cancer in such patients rises slowly at first to around 4%

after 10 years of colitic symptoms. However, after 10 years of colitis, the cumulative risk increases more and more sharply to reach a startling level of 41–48%

after 25 years of colitic symptoms [5, 6]. Recently, similar results have been presented, while others demonstrate that those with total or extensive colitis (extending proximal to the splenic flexure), colitis of 8 years or more, a family history of colorectal cancer, primary sclerosing cholangitis and an early age of onset of colitis have the greatest risk of developing cancer [3, 7]. Primary sclerosing cholangitis (PSC) occurs in about 2.5–6% of patients with UC, adding considerable risk of cancer compared with UC in general [7]. The mean interval from diagnosis of PSC to dysplasia or cancer is only 2.9 years. Colorectal cancer associated with PSC is more likely to be proximal, to be diagnosed at a more advanced stage, and to be fatal.

While patients with extensive colitis may have an almost 20 times higher risk of cancer compared with the general population of patients, those with left- sided colitis may only have a four times higher risk, and for an individual patient with a distal proctocolitis, the risk is very small, probably about the same as that in the general population of the same age and sex [4, 8].

The age of the patient at the onset of colitis is another factor which has inspired debate. It has often been claimed that colitic patients whose disease began in childhood may have a particular tendency to develop cancer [9]. One explanation that has been put forward is that total involvement – a strong risk factor – is far more common in patients under 20 years of age than in elderly patients. Another important reason why patients who develop ulcerative colitis as children are particularly at risk from carcinoma, is that they are the only patients who will live long enough to develop cancer! Patients who develop total involvement at the age of 50 or later already have a curtailed life expectation and will probably not live long enough to develop cancer.

Therefore, the main factor which undoubtedly exerts an effect upon the cancer risk for colitic patients is the extent of the patient’s colitis.

Tumour Characteristics and Prognosis

Colitis carcinoma occurs as a rule at an earlier age than does ordinary large-bowel cancer, the average

age incidence being 41 years, compared with about 60 years in cases of ordinary carcinoma of the colon and rectum [9]. While most writers claim that the sit- ing of colitis carcinomas does not differ materially from that of ordinary large bowel carcinomas, others have demonstrated that while approximately three- quarters of ordinary carcinomas are concentrated in the rectum and sigmoid, only one-quarter of the colitic carcinoma occurs in the rectosigmoid segment [10]. Another notable feature of colitis carcinomas is the high incidence of multiple primary growths which is about four times greater than that found with carcinoma of the colon in general. The individual growth may exhibit the gross features of an ordinary colorectal cancer but may sometimes be quite atypical, being flat and infiltrating with an ill-defined edge or sometimes appearing rather as a relatively inconspicuous thickening of the colon wall-like fibrous strictures.

Colitic cancers differ even histologically from sporadic cancer as they are more often high grade and mucinous or of a signet-ring cell type, and usually they have spread extensively by the time the patient comes to surgery [11]. Dysplasia in non-colitic cancers is almost always polypoid. In patients with colitis, however, there are flat dysplasia (mostly invisible to an endoscopist), raised dysplasia called DALM (dysplasia-associated lesion or mass) and incidental adenomas.

Because of the active nature of these colitic growths and the delay that often attends their diagnosis, many of them may have reached an advanced inoperable or incurable stage by the time they come to surgery. In the treatment of established colitis carcinomas, if the growth is operable, the ideal is panproctocolectomy since more limited resections are very liable to be followed by the development of further carcinomas in the remaining colon or rectum [11].While reasonably good results may sometimes be achieved after surgery, the average 5-year survival rate is less than after idiopathic cancer [12].

The Justification for Prophylactic Surgery

The high cancer risk and dubious prognosis is dis- quieting and it automatically raises the question as to the advisability of prophylactic excisional surgery.

Over the years, the opinion has been sharply divided as to the justification for this step. While some authors consider that the cancer risk does not neces- sarily justify an increased use of colectomy [13, 14]

others advocate preventive surgery for all patients with a 10-year history of ulcerative colitis and total involvement of the colon and rectum [4, 5, 6]. Still others go even further and recommend surgery for

(3)

all patients with a radiologically abnormal colon four or more years after the onset of disease [15]. Howev- er, even in patients with longstanding colitis, elective proctocolectomy, with its inherent mortality and morbidity, cannot be justified on the basis of cancer prevention in patients whose colitis is restricted to the distal half of the colon. However, in the group of patients with total or subtotal involvement of the colon with a colitic history of 10 or more years, a very strong case can clearly be argued for elective surgery on the basis of cancer prevention alone. In such patients, the risk of cancer with continued conserva- tive treatment is so high as to completely outweigh the hazards of prophylactic surgery.

Cancer Screening and Colonoscopy Surveillance

In the past, the indication for prophylactic colectomy for patients with longstanding ulcerative colitis was mainly based on the disease history and the clinical risk factors mentioned above. Today, with the avail- ability of colonoscopy and the recognition of the dysplasia–precancer–cancer sequence, surveillance colonoscopy with serial colonoscopic examinations and mucosal biopsies is thought to allow for a more adequate individual assessment of the cancer risk and firm recommendation for surgery before the development of cancer. Thus, it is stated that prophylactic surgery should be reserved for patients whose biopsy findings are indicative of heightened cancer risk based on the joint interpretation by the clinician and the pathologist.

Surveillance Guidelines

Guidelines have recently been published by authors from different expert centres [16–19]. Since duration and anatomic extent of the ulcerative colitis have been shown to be the two strong and independent risk factors regarding the development of cancer, special attention to these variables is crucial in the practice of surveillance. The UK guidelines for surveillance in ulcerative colitis [19] state that surveillance colonoscopy should preferably be performed in disease remission, it should be performed in all patients after 8–10 years of disease to clarify disease extent, and regular surveillance should begin after 8–10 years in those with pancolitis and after 10–15 years in those with left-sided colitis. Given that cancer risk increases with the duration of disease, the interval between colonoscopies should decrease in line with increased disease duration. In the second decade of disease, colonoscopy should be carried out every 3 years, in the third decade every 2 years and

every year in the fourth decade. Based on the fact that interval cancers can develop within 2 years after an examination, patients with extensive (or left-sided) colitis who have a negative result should begin surveillance within 1–2 years. With a negative surveillance colonoscopy, subsequent examinations should be performed every 1–2 years. With two negative examinations, the next surveillance examination may be performed in 1–3 years until colitis has been present for 20 years.

To obtain 90% sensitivity for the detection of dysplasia in patients with extensive disease, a minimum of 33 biopsies is suggested, taking 4 biopsies for every 10 cm around the colon. In those with less extensive microscopic disease, four quadrant biopsies taken from the proximal extent of disease and every 10 cm distally would be satisfactory. The commonly followed guideline is that any diagnosis of high-grade dysplasia (and low-grade dysplasia when associated with a DALM) should lead to consideration of colectomy, while low-grade dysplasia in flat mucosa should not be an indication for colectomy but for continuous surveillance with colonoscopy. The issue is controversial. While DALM or cancer in such cases has been proved to occur in more than half the patients at 5 years [20] other authors [21] considered low-grade dysplasia (unrelated to a DALM) less alarming and not sufficiently reliable as a marker to justify prophylactic colectomy.

Careful comparative studies by experienced pathologists show evidence that surveillance and screening programmes may carry a significant rate of histological error due to interobserver variation between the gastrointestinal pathologists when grading dysplasia in ulcerative colitis [22].

The Efficacy of Surveillance

There is evidence showing that colonoscopic surveillance may detect precancerous dysplasia and early treatable cancer. Thus, some studies indicate that patients with cancers detected by surveillance tend to be at a curable stage while patients not adhering to surveillance are most likely to die from cancer [23].

Other studies, however, are less convincing [24], showing evidence that only very few of the malignan- cies found were treatable cancers detected by true surveillance colonoscopy and with only a marginally better success rate. There seems in fact to be no direct evidence that endoscopic surveillance reduces cancer mortality in inflammatory bowel disease [25], and a review reported on in the Cochrane Central Register of Controlled Trials presents a similar message [18].

Although cancers tend to be detected at an earlier stage and has a correspondingly better prognosis in

(4)

patients who are undergoing surveillance, lead-time bias seems to contribute substantially to this appar- ent benefit; and there is no clear evidence that surveillance colonoscopy prolongs survival in patients with extensive colitis. Nevertheless, there seems to be consensus that a surveillance programme should offer colonoscopy to patients with extensive ulcerative colitis of 8 years duration and to those with less extensive disease of 15 years duration. Surveillance colonoscopy should be performed every 3 years for 10 years, every 2 years for 10 and then annually including at least four biopsies taken every 10 cm around the colon with careful biopsy of any macro- scopic lesion.

Surveillance and its Limitations

Mucosal dysplasia documented by surveillance colonoscopy used clinically to identify the “high risk” patients has proven to have significant limitations. Sampling error, presence of acute mucosal inflammation, differences in pathological interpretation (observer variations) and variable patient compliance are factors that might interfere with the diagnostic accuracy.

The reliability of dysplasia as a precursor of cancer is poor [22]. The histological interpretation is often subjective and of doubtful diagnostic value and the decision making resulting from its discovery will, therefore, often be doubtful. The grading of dysplasia, differentiation between true dysplastic changes and active inflammatory changes and differentiation between DALM and “sporadic” adenomas constitute the dilemma for the pathologist. Moreover, an important obstacle to an accurate assessment of dysplasia is an active mucosal inflammation that may result in cytological changes that are hard to distin- guish from dysplasia.

Because dysplasia is focal and patchy, it may not even be detected by extensive colonoscopic biopsies, and established cancers may be missed entirely. Thus, there is evidence showing that a significant fraction of patients may have colorectal cancer in the absence of dysplasia at the colonoscopic surveillance, although this may at least in part be attributable to inadequate surveillance [26]. Based on such observations, the authors claimed that if proctocolectomy is recommended only upon detection of either high-grade dysplasia or dysplasia of any grade, a large proportion of patients may already have an established cancer at the time of surgery. It is obvious that there are practical limitations to the theoretical concept that dysplasia surveillance can reliably serve as a marker for malignant degeneration. The authors also emphasised that failure to operate on high-risk patients – i.e. those

with pancolitis, those with duration of disease more than 10 years and those with disease presenting during childhood – while waiting for dysplasia to be dis- covered by surveillance colonoscopy will invite a high risk of developing concomitant colorectal cancer.

How Can Diagnostic Accuracy be Improved?

It is possible that new molecular markers in conjunc- tion with histologic dysplasia will improve the sensitivity of the surveillance and biopsy approach. Can- cer surveillance may also be improved by better selection of patients for inclusion in surveillance programmes, using markers other than dysplasia to predict cancer. Molecular genetic research may produce better ways of selecting patients at greatest risk [27].

It may also produce a better premalignant marker than dysplasia and it can help in distinguishing colitic dysplasia from other entities. Chromoendoscopy can improve the detection of dysplasia and may also be helpful in distinguishing colitic neoplasia from non-colitic neoplasia. Dye spraying of the colonic mucosa during colonoscopy (chromoendoscopy) combined with high-resolution colonoscopy using a magnifying colonoscope is another measure used to increase the detection rate of neoplastic lesions in patients with colitis [28, 29]. Many of these new methods may be promising but no technique has yet been convincing or has entered into routine clinical practice.

Last but not the least, it should be emphasised that, apart from a highly skilled endoscopist and histopathologist, for a program of surveillance to be efficient, it also has to rely on both the physician’s and patient’s compliance. There should be regular call back for all participating patients so that no patients are lost in follow-up.

Even then, cancer surveillance does not totally eliminate the risk of cancer. Despite many successful results, one must question as to whether colonoscopy surveillance, arguably efficacious for special clinic populations, is truly effective even in the community at large. In other words, will the results obtained from careful clinical trials – produced under ideal conditions – be reproducible when deployed in routine clinical practice? For example, it has been esti- mated, as a “best-case” scenario, that colonoscopic surveillance may decrease the incidence of cancer from 7–8% down to 0.5–1% and the author claims that no actual program is likely to enjoy such success.

Therefore, “at risk” ulcerative colitis patients who are averse to take this cancer risk and who cannot accept the imperfect and in many respects inconvenient nature of colonoscopic surveillance, should be recommended for prophylactic colectomy.

(5)

Prophylactic Surgery: What Are the Options?

Panproctocolectomy and Ileostomy (PI)

For prevention of cancer development in patients with ulcerative proctocolitis, complete removal of all potential malignant colorectal mucosa has to be done. Before the advent of the surgical options of our days, panproctocolectomy with construction of a conventional ileostomy was the standard procedure for cancer prophylaxis in patients with longstanding ulcerative colitis. Such an operation eliminates further risks of colorectal cancer and no surveillance should be needed. Although cancer in the ileostomy in these patients has been demonstrated to occur many years after surgery, this is probably a different state of affairs [31]. Bowel metaplasia may occur where gut contents come into regular contact with the squamous epithelium of the skin, which is an important step in the development of these rare tumours, as is also true in some cases of adenocarcinoma of the oesophagus. Ileostomy cancer develops at the mucocutaneous junction of the ileostomy, and chronic irritation caused by trauma and/or chemical agents from stoma appliances or adhesives may be factors in its unclear etiology.

Colectomy and Ileorectal Anastomosis (IRA)

Conflicting results have been presented as regards the indications for colectomy with IRA for ulcerative colitis. Although many surgeons today are still reluc- tant to use the technique, emphasising not only the persistent cancer risk but also the poor function [32–34], others consider the operation a viable alternative when used selectively in patients without signs of mucosal dysplasia and whose rectum is not severe- ly affected by inflammation or fibrosis [35–37]. The colectomy and IRA procedure for a condition that almost invariably involves an inflamed rectum cer- tainly seems illogical. Apart from poor function, there is a significant risk of cancer development in the chronically inflamed mucosa. It is often argued that a substantial proportion of IRA patients – maybe even half of them – will have their rectum excised eventually due to persistent or relapsing proctitis [38]. However, in many cases, the proctitis often set- tles spontaneously or after local treatment or recurs periodically. Thus, the patient may enjoy reasonably good general health and bowel function. Therefore, even if 40–50% of the patients will ultimately require rectal excision, roughly half of the patients will con- tinue to enjoy a satisfactory result and many of those who finally fail, have been able to postpone a major

operation (IPAA) or an abdominal stoma for several years. Therefore, the time “bought” by IRA will get many young people through their formative years of education, allowing them to plan for a family and a professional career. However, the long-term risk of cancer in the rectal stump is the main strong argu- ment that has been put forward against the use of this operation – a risk that increases with the duration of the disease and with the passage of time after the colectomy. The cumulative probability of cancer development approaches 5 and 15% after a 20 and 30-year observation, respectively [34, 39]. A very important predictive factor is the presence of severe dysplasia in the rectum or carcinoma of the colon at the time of surgery [40]. Bearing this in mind, the risk of rectal cancer should therefore be low in patients with a short antecedent disease history and in those without mucosal dysplasia in the colon/rectum. As regular well-designed colonoscopic surveillance is considered justified for the control of patients with longstanding panproctocolitis – in whom the cancer risk is 3–4 times higher – similar guidelines [19] should be quite appropriate and safe for surveillance of patients with an IRA. In other words, by using a meticulous follow-up system for patients with IRA, it should be possible to identify patients who are at a particular risk and urge them to undergo prophylactic rectal excision.

The controversy regarding colectomy and IRA and the place of this procedure in the treatment of UC has been characterised in the past by personal prejudices and overzealous condemnation of the technique by several of the more prominent experts on colorectal surgery [38]. Employed on a selective basis, IRA should a safe procedure with low mortality and morbidity and good prospects for success as a prophylactic procedure in many patients with longstanding ulcerative colitis. However, before surgery, the patients must be fully informed of any inconven- iences and risks associated with the procedure, and, most importantly, they must be prepared to submit to lifelong endoscopy surveillance.

Restorative Proctocolectomy: Ileo Pouch-Anal Anastomosis (IPAA)

The interest in IRA may well have declined with the advent of proctocolectomy and the ileo pouch-anal anastomosis (IPAA). IPAA or restorative proctocolectomy, i.e. construction of a reservoir of distal ileum and an ileo pouch-anal anastomosis is the current most popular option for surgical treatment of ulcerative proctocolitis. There is no stoma or need for an external bag and the normal route of defeca- tion is preserved, i.e. a normal body image. It has

(6)

become the first choice operation to be recommended around the world. In the conventional technique, colectomy is combined with endoanal mucous proc- tectomy and the ileal pouch is hand-sewn to the pectinate line. In the currently most popular technique the abdominal dissection is carried out down to the levator muscle, the rectum is severed at this level and the ileal pouch is connected to the rectal stump by a stapling device. In analogy to the tradi- tional total proctocolectomy procedure, it has been considered a curative and cancer-prophylactic procedure since all diseased mucosae are completely removed. At first sight, restorative proctocolectomy therefore seems to be an unmistakable opportunity;

however, recent results imply that the IPAA procedure may not be the panacea it was thought to be. It is a demanding operation with a high potential for complications even in an experienced surgeon’s hands and the functional results are sometimes far from perfect [41]; even its place as a cancer-prophylactic procedure may in fact be in question. An increasing number of cancers have been reported in these patients and the incidence is expected to rise as the length of follow-up increases [42].

Chronic inflammation in the ileal mucosa (pouchitis) is a frequent complication in continent ileostomy and has proved subsequently to be so even in the pelvic pouch (IPAA)[43]. A case of adenocarcinoma in the continent ileostomy [44] and sporadic reports of dysplasia in the ileal pouch mucosa, have currently appeared in the literature, suggesting that the mor- phological transformation of the ileal pouch mucosa might eventually result in cellular dysplasia and eventually carcinoma [45, 46]. The atrophic colon- like mucosa in the ileal pouch is hypothetically considered a potentially premalignant condition with risk of subsequent development of advanced neoplastic transformation. It has been suggested that dysplasia and aneuploidy, as demonstrated by these authors, reflect a different pathway of an atrophic mucosa-dysplasia carcinoma sequence. However, the results from long-term studies, both on continent ileostomy patients and subsequently on IPAA patients, are reassuring [47, 48]. The overall incidence of mucosal dysplasia in the ileal pouch mucosa proved to be low and no case of high-grade dysplasia or carcinoma was observed. Considering an observation time of an average of 30 years in these studies, and the comparatively large series of patients, these results imply that dysplastic and neoplastic transformation within the ileal pouch mucosa is extremely rare regardless of the type of adaptation and the risk for epithelial dysplasia in the ileal pouch mucosa to progress into cancer seems to be very low.

The published reports on cancer that develops in the IPAA patients operated on for ulcerative colitis,

reflect a quite different issue, however [42, 48]. Irre- spective of the technique used, IPAA leaves residual rectal mucosa behind and dysplasia in these rectal mucosal remnants with subsequent cancer development has proved to be a procedural risk, reflecting the continuous risk of malignant transformation in the chronically inflamed rectal mucosa. Thus, it has been demonstrated that even after a careful macro- scopically complete mucosectomy, islands of remnants of rectal mucosa are left behind in about 20%

of the cases [49], and in the alternative technique where the ileal pouch is stapled to the top of the anal canal, varying amounts of rectal mucosa as well as the anal transitional zone mucosa remain preserved.

The rectal stump may even include part of the lower rectum in technically demanding cases. Thus, there is a potential risk for malignant development in the islands of mucosa remaining between the muscular cuff and ileal mucosa or in the retained mucosa in the anal transitional zone, the so-called residual epithelial cuff after the stapled technique. Moreover, it is convincingly demonstrated that this risk is impend- ing in patients with a long history of antecedent ulcerative colitis, and with the diagnosis of dysplasia or cancer in the operative specimen at the time of colectomy [42, 48]). Therefore, although there are reports suggesting that an IPAA is a successful surgical approach for ulcerative colitis patients with coex- isting colorectal cancer [50], it is doubtful if such an approach should be recommended.

Although some colorectal surgeons may question the need for routine surveillance for cancer in the IPAA patients [51], these observations imply that despite the fact that the cancer risk after IPAA may well be less than after the IRA procedure, a similar endoscopy surveillance is still justified. A close follow-up in all IPAA patients should be emphasised, with special attention focussed on those with a long antecedent disease history and those with dysplasia or cancer in the original specimen. In this context, it should be emphasised that the dysplasia or early cancer that arises from the residual rectal tissue in the muscular cuff after mucosectomy may not be easily detectable and endoscopy surveillance with deep random biopsies of the anal canal mucosa should be recommended.

References

1. Devroede GJ, Taylor WF, Sauer WG, Jackman RJ and Stickler GB (1971) Cancer risk and life expectancy of children with ulcerative colitis. N Engl J Med 285 1:17–21

2. Devroede G, Taylor WF (1976) On calculating cancer risk and survival of ulcerative colitis patients with the life table method. Gastroenterology 71(3):505–509

(7)

3. Eaden JA, Abrams KR and Mayberry JF (2001) The risk of colorectal cancer in ulcerative colitis: a meta- analysis. Gut 48(4):526–535

4. Macdougall IP (1964) The cancer risk in ulcerative colitis. Lancet 19:655–658

5. De Dombal FT, Watts JM, Watkinson G et al (1966) Local complications of ulcerative colitis: stricture, pseudopolyposis and carcinoma of colon and rectum.

BMJ 1:1442

6. Hulten L, Kewenter J, Ahren C (1972) Precancer and carcinoma in chronic ulcerative colitis: a histopatho- logical and clinical investigation. Scand J Gastroen- terol 7(7):663–669

7. Shetty K, Rybicki L, Brzezinski A et al (1999) The risk for cancer or dysplasia in ulcerative colitis patients with primary sclerosing cholangitis. Am J Gastroen- terol 94(6):1643–1649

8. Gyde SN, Prior P, Allan RN et al (1988) Colorectal cancer in ulcerative colitis: a cohort study of primary referrals from three centres. Gut 29(2):206–217 9. Edwards FC, Truelove SC (1964) The course and prog-

nosis of ulcerative colitis. III. Complications. Gut 32:1–22

10. Edling NP, Eklof O (1961) Distribution of malignancy in ulcerative colitis. Gastroenterology 41:465–466 11. Hulten L, Kewenter J, Kock NG (1971) The long–term

results of partial resection of the large bowel for intes- tinal carcinomas complicating ulcerative colitis. Scand J Gastroenterol 6(7):601–604

12. Slaney G, Brooke BN (1959) Cancer in ulcerative colitis. Lancet 2:694–698

13. Texter EC Jr (1957) The natural history of ulcerative colitis. J Chronic Dis 5(3):347–369

14. Carleson RFB, Philipsson J (1962) Ulcerative colitis: a follow-up investigation of a 20-year primary material.

Acta Med Scand Dec 172:647–656

15. Dennis C KK (1961) Cancer risk in ulcerative colitis:

formidability per patient year of late disease. Surgery 50:568–574

16. Itzkowitz SH, Harpaz N (2004) Diagnosis and manage- ment of dysplasia in patients with inflammatory bowel diseases. Gastroenterology 126(6):1634–1648

17. Chambers WM, Warren BF, Jewell DP et al (2005) Can- cer surveillance in ulcerative colitis. Br J Surg 92(8):928–936

18. Mpofu C, Watson AJ, Rhodes JM (2004) Strategies for detecting colon cancer and/or dysplasia in patients with inflammatory bowel disease. Cochrane Database Syst Rev 2:CD000279

19. Eaden JA, Mayberry JF (2002) Guidelines for screening and surveillance of asymptomatic colorectal cancer in patients with inflammatory bowel disease. Gut 51(Suppl 5):10–12

20. Fireman Z, Wagner G, Weissman J et al (2001) Preva- lence of short-segment Barrett’s epithelium. Dig Liver Dis 33(4):322–325

21. Ullman T, Croog V, Harpaz N, Sachar D and Itzkowitz S (2003) Progression of flat low-grade dysplasia to advanced neoplasia in patients with ulcerative colitis.

Gastroenterology 125(5):1311–1319

22. Lim CH, Axon AT (2003) Low-grade dysplasia: non- surgical treatment. Inflamm Bowel Dis 9(4):270–272 23. Eaden J, Abrams K, McKay H et al (2001) Inter-observ-

er variation between general and specialist gastrointestinal pathologists when grading dysplasia in ulcerative colitis. J Pathol 194(2):152–157

24. Karlen P, Kornfeld D, Brostrom O et al (1998) Is colonoscopic surveillance reducing colorectal cancer mortality in ulcerative colitis? A population based case control study. Gut 42(5):711–724

25. Lynch DA, Lobo AJ, Sobala GM et al (1993) Failure of colonoscopic surveillance in ulcerative colitis. Gut 34(8):1075–1080

26. Winawer S, Fletcher R, Rex D et al (2003) Colorectal cancer screening and surveillance: clinical guidelines and rationale-update based on new evidence. Gas- troenterology 124(2):544–560

27. Taylor BA, Pemberton JH, Carpenter HA et al (1992) Dysplasia in chronic ulcerative colitis: implications for colonoscopic surveillance. Dis Colon Rectum 35(10):950–956

28. Itzkowitz SH, Marshall A, Kornbluth A et al (1995) Sialosyl-Tn antigen: initial report of a new marker of malignant progression in long-standing ulcerative colitis. Gastroenterology 109(2):490–497

29. Kiesslich R, Jung M, DiSario JA et al (2004) Perspec- tives of chromo and magnifying endoscopy: How, how much, when, and whom should we stain? J Clin Gas- troenterol 38(1):7–13

30. Kawamura YJ, Togashi K, Sasaki J, Konishi F (2005) Acetic acid spray in colonoscopy: an alternative to chromoendoscopy. Gut 54(2):313

31. Provenzale D, Kowdley KV, Arora S, Wong JB (1995) Prophylactic colectomy or surveillance for chronic ulcerative colitis? A decision analysis. Gastroenterolo- gy 109(4):1188–1196

32. Quah HM, Samad A, Maw A (2005) Ileostomy carcinomas a review: the latent risk after colectomy for ulcerative colitis and familial adenomatous polyposis. Col- orectal Dis 7(6):538–544

33. Albrechtsen D, Bergan A, Gjone E et al (1981) Elective surgery for ulcerative colitis: colectomy in 158 patients. Scand J Gastroenterol 16(6):825–831 34. Parc R, Legrand M, Frileux P et al (1989) Comparative

clinical results of ileal-pouch anal anastomosis and ileorectal anastomosis in ulcerative colitis. Hepatogas- troenterology 36(4):235–239

35. Pastore RL, Wolff BG, Hodge D (1997) Total abdominal colectomy and ileorectal anastomosis for inflammatory bowel disease. Dis Colon Rectum 40(12):1455–1464 36. Khubchandani IT, Sandfort MR, Rosen L et al (1989) Current status of ileorectal anastomosis for inflammatory bowel disease. Dis Colon Rectum 32(5):400–403 37. Leijonmarck CE, Lofberg R, Ost A, Hellers G (1990)

Long–term results of ileorectal anastomosis in ulcerative colitis in Stockholm County. Dis Colon Rectum 3:195–200

38. Tonelli FBG et al (1989) Continenza e modificazioni della pressione del canale anale e della distensibilita rettale dopo ileorettoanastomosi. Chirurgia 2:289–296 39. Goligher JC (1984) Surgery of the anus, rectum and

colon. Bailiere, London

40. Baker WN, Glass RE, Ritchie JK, Aylett SO (1978) Can- cer of the rectum following colectomy and ileorectal anastomosis for ulcerative colitis. Br J Surg 65(12):862–868 41. Morson BC, Pang LS (1967) Rectal biopsy as an aid to

(8)

cancer control in ulcerative colitis. Gut 8(5):423–434 42. Delaini GG, Colucci G, Hulten L (2005) The ileoanal

pouch procedure in the long-term perspective: a criti- cal review. Tech Coloproctol 9:187–192

43. Lee SW, Sonoda T, Milsom JW (2005) Three cases of adenocarcinoma following restorative proctocolectomy with hand-sewn anastomosis for ulcerative colitis:

a review of reported cases in the literature. Colorectal Dis 7(6):591–597

44. Svaninger G, Nordgren S, Oresland T, Hulten L (1993) Incidence and characteristics of pouchitis in the Kock continent ileostomy and the pelvic pouch. Scand J Gastroenterol 28(8):695–700

45. Cox CL, Butts DR, Roberts MP et al (1997) Develop- ment of invasive adenocarcinoma in a long-standing Kock continent ileostomy: report of a case. Dis Colon Rectum 40(4):500–503

46. Setti-Carraro P, Ritchie JK, Wilkinson KH et al (1994) The first 10 years’ experience of restorative proctocolectomy for ulcerative colitis. Gut 35(8):1070–1075 47. Gullberg K, Stahlberg D, Liljeqvist L et al (1997) Neo-

plastic transformation of the pelvic pouch mucosa in

patients with ulcerative colitis. Gastroenterology 112(5):1487–1492

48. Hulten L, Willen R, Nilsson O et al (2002) Mucosal assessment for dysplasia and cancer in the ileal pouch mucosa in patients operated on for ulcerative colitis: a 30-year follow-up study. Dis Colon Rectum 45 4:448–452

49. Borjesson L, Willen R, Haboubi N et al (2004) The risk of dysplasia and cancer in the ileal pouch mucosa after restorative proctocolectomy for ulcerative proctocolitis is low: a long-term term follow-up study. Colorec- tal Dis 6(6):494–498

50. O’Connell PR, Pemberton JH, Weiland LH et al (1987) Does rectal mucosa regenerate after ileoanal anastomosis? Dis Colon Rectum 30(1):1–5

51. Remzi FH, Preen M (2003) Rectal cancer and ulcerative colitis: Does it change the therapeutic approach?

Colorectal Dis 5(5):483–5

52. Herline AJ, Meisinger LL, Rusin LC et al (2003) Is routine pouch surveillance for dysplasia indicated for ileoanal pouches? Dis Colon Rectum 46(2):156–159