Cancer Risk

Patients who have chronic IBD, whether classified as ulcerative colitis or Crohn’s colitis, are at greater than average risk of developing CRC . The reported risk varies considerably, however, with older studies generally showing a very high risk (perhaps reflecting referral bias) and recent studies suggesting a much lower risk. A recent meta-analysis of 116 studies found the overall prevalence of CRC in ulcerative colitis to be 3%, with a cumulative risk of CRC of 2% at 10 years, 8% at 20 years, and 18% at 30 years . There also is considerable geographic variation, with higher rates of cancer in the United States and United Kingdom and lower rates in Scandinavia . The mean time from the diagnosis of colitis to the diagnosis of cancer is 17 years, with a mean age at diagnosis of cancer of 51 years for men and 54 for women . Although most studies demonstrate an increased risk of cancer in IBD, this finding is not universal. Indeed, two recent population studies from Denmark and the Mayo Clinic failed to show any increased risk of cancer in patients who had ulcerative colitis compared with a control population . Furthermore, the reported mortality from the colon cancer is variable. One study suggested that colon cancer accounts for one third of deaths in ulcerative colitis , whereas another study found that this cancer accounts for only one sixth of deaths .

Crohn’s colitis, particularly with extensive colonic involvement, carries a similar risk of developing CRC . CRC associated with Crohn’s disease and CRC associated with ulcerative colitis seem to share unusual features . Both groups tend to have cancer diagnosed after 8 years of IBD and frequently have multifocal cancer with an aggressive (signet ring or mucinous) histology. The cancers in Crohn’s colitis were confined to areas of diseased mucosa. A recent population-based study from Olmsted County, Minnesota showed only a modestly increased risk (standardized incidence ratio, 1.9) for CRC in Crohn’s disease but a marked (40-fold) increased risk for small intestinal cancer .

Pathogenesis

The precise pathogenesis of CRC in IBD is unclear. The pathogenesis is believed to involve chronic inflammation leading to increased cell proliferation with subsequent development of dysplasia . This postulated mechanism is different from the gene deletion sequence that occurs in sporadic, non-IBD carcinomas . A detailed discussion of the molecular mechanisms of neoplasia is beyond the scope of this article and has been reviewed elsewhere . The progression to dysplasia and cancer in ulcerative colitis may involve genomic instability with chromosomal breakage caused by shortened telomeres; patients who have IBD may have a mutator phenotype that predisposes them to colon cancer . Colonoscopic surveillance to prevent colon cancer in IBD is based on the concept of stepwise progression from chronic inflammation through low-grade dysplasia (LGD) and high-grade dysplasia to cancer.

Risk Factors

Risk factors for CRC in patients who have IBD are

• 
  

  Duration of disease

  
  
• 
  

  Extent of colitis

  
  
• 
  

  Primary sclerosing cholangitis

  
  
• 
  

  Family history of CRC

  
  
• 
  

  Inflammatory pseudopolyps

  
  
• 
  

  Degree of histologic inflammation

The duration of disease and extent of colitis are well-established risk factors. Other important clinical risk factors for neoplasia in ulcerative colitis have been identified recently. A recent case-controlled study of 188 patients who had ulcerative colitis–related colon cancer showed that the presence of inflammatory pseudopolyps was an independent risk factor (odds ratio [OR], 2.5), whereas cigarette smoking, the use of aspirin, nonsteroidal anti-inflammatory drugs, or 5-acetyl salicylic acid (5-ASA), and a history of one or two surveillance colonoscopies decreased the risk . More severe histologic inflammation of the colon also has been recognized recently as an important risk factor . The coexistence of primary sclerosing cholangitis (PSC) confers an approximately fivefold increase in CRC risk . A meta-analysis of 11 studies incorporating 16,844 patients who had ulcerative colitis, including 560 who had concomitant PSC, reported an approximately fivefold increase in the CRC risk associated with PSC . The mechanism of this increase in PSC is unknown, but altered bile salt composition in the colonic lumen has been suggested as a mechanism . This increased risk of cancer has led to recommendations that patients who have IBD and PSC undergo more intensive surveillance colonoscopy beginning at the time of diagnosis (see guidelines presented later). In addition, patients who have PSC and IBD have an increased risk of cholangiocarcinoma . Ulcerative colitis associated with ileal inflammation (“backwash ileitis”) seems to be a risk factor for colon cancer .

The clinical characteristics of cancers associated with IBD are somewhat different from those of sporadic CRC. These cancers occur at a younger age (mean age, 43 years) , are more frequently multifocal , and exhibit more aggressive histology . Despite these unfavorable prognostic factors, a recent case-controlled study of 290 patients who had IBD-related CRC showed that the 5-year survival rates were virtually identical (56% versus 57%) to those of patients who had non–IBD-related CRC .

Surveillance Strategies (Secondary Prevention)

Colonoscopy with removal of elevated polypoid precursor lesions (adenomatous polyps) is the most common prevention strategy for sporadic colon cancer. In IBD-related cancers, however, the precursor dysplastic lesion often is flat and may not be readily evident endoscopically. Thus, the primary strategy for preventing CRC in patients who have IBD is surveillance colonoscopy with multiple random biopsies throughout the colon to detect dysplasia, with subsequent proctocolectomy for patients harboring dysplastic lesions in the hopes of preventing progression to cancer or of removing cancer at an earlier, potentially curable, stage.

In a retrospective cohort analysis of 91 screened subjects versus 95 unscreened controls, screening improved overall survival, but the survival benefit was not related to increased cancer-related survival . In contrast, a study of 41 patients who had ulcerative colitis showed that patients undergoing colonoscopic surveillance had cancers diagnosed at an earlier Dukes stage and had an improved 5-year survival (77.2% versus 36.3%) . In a population-based, nested, case-control study of 142 patients who had ulcerative colitis who died from colon cancer, one surveillance colonoscopy was associated with a decreased risk of cancer death (OR, 0.29), and two or more colonoscopies were associated with a further risk reduction . A recent report of a surveillance program involving 600 patients, with up to 40 years of follow-up, reported that the cancer incidence was lower than expected (10.8% at 40 years) and that the 5-year survival rate for those who had cancer was very high (73%) . Indeed, only 2% of the surveillance group died from CRC during the 40-year study period. One half of the cancers, however, were thought to be interval cancers that potentially were missed on the preceding colonoscopy, and this surveillance program did not uniformly prevent advanced cancer.

In a recent Cochrane review, the authors found no direct evidence that colonoscopy prolonged survival in patients who had IBD, although patients in surveillance programs tended to have cancers diagnosed earlier with a correspondingly better prognosis . They could not exclude lead-time bias as a contributing factor for this benefit. On the basis of indirect evidence, however, the authors concluded that surveillance seemed to decrease mortality from CRC and that surveillance was cost effective.

Despite the lack of conclusive evidence of benefit, surveillance colonoscopy with biopsy currently is recommended by the major gastrointestinal societies worldwide, including the American Gastroenterology Association, the American Society for Gastrointestinal Endoscopy, the American College of Gastroenterology, the British Society of Gastroenterology, and the American Society of Colon and Rectal Surgeons. A recent consensus conference of the Crohn’s and Colitis Foundation of America also endorsed this recommendation .

Dysplasia

Surveillance strategies critically depend on the detection and accurate histologic classification of dysplasia, defined as unequivocal neoplastic transformation of mucosa, and subsequent proctocolectomy to reduce the incidence and mortality from CRC in patients who have IBD. The precise histologic definitions of low-, high-, and indeterminate-grade dysplasia have been reviewed recently and are beyond the scope of this article . Interobserver variability in the diagnosis of dysplasia is common and. may justify outside pathologic review of biopsies initially read as containing dysplasia . Pathologists generally agree when there is no dysplasia or high-grade dysplasia but often disagree in cases of indeterminate- or low-grade dysplasia .

High-grade Dysplasia and Dysplasia-associated Lesions or Masses

High-grade dysplasia, when confirmed independently by two experienced pathologists, generally warrants proctocolectomy because of the high (42%–62%) risk of concurrent cancer . The finding of any degree of dysplasia in the setting of a grossly visible mass lesion, called a “dysplasia-associated lesion or mass” (DALM), is generally a strong indication for immediate proctocolectomy because up to 50% of these patients have frank cancer at surgery .

Low-grade Dysplasia

The management of LGD is controversial. Many experts recommend immediate colectomy, but others favor intense colonoscopic surveillance . Patients who have LGD have an approximately 20% risk of harboring concurrent cancer and a 50% risk of progressing to advanced dysplasia with time . In a retrospective review of patients who had LGD, one half developed advanced neoplasia (defined as cancer, DALM, or high-grade dysplasia) during a mean follow-up of 32 months .

Not all studies show this association between LGD and advanced neoplasia, however. Among 60 patients who had LGD followed for a mean of 10 years, no patient developed cancer, although 11 patients developed DALMs . In a retrospective cohort study of 29 patients who had LGD followed for 10 years, 10% developed high-grade dysplasia or cancer compared with 4% of controls, a difference that was not statistically significant .

Rubin and Turner suggest that multifocal LGD is an indication for colectomy but that unifocal LGD can be managed by either colectomy or aggressive colonoscopic surveillance. Other experts suggest that unifocal and multifocal LGD have the same cancer risk and that a policy of increased endoscopic surveillance is unwise because patients may develop advanced (node-positive) CRC despite intensive colonoscopic surveillance . Both unifocal and multifocal LGD were found to progress to high-grade dysplasia or CRC in 50% at 5 years. A recently published meta-analysis of 20 surveillance studies of patients who had LGD found that LGD was associated with a ninefold increased risk of developing cancer and a 12-fold risk of developing high-grade dysplasia or cancer . LGD was associated with concurrent CRC in 22% and subsequent cancer in 7.9%.

Table 1 lists the recommendations for management of dysplasia promulgated by various societies. Proctocolectomy is the preferred management strategy for any patient who has IBD and high-grade dysplasia and for any grade of DALM. The management of LGD remains controversial. Most experts recommend immediate colectomy because of the high risk of concurrent or subsequent cancer . Patients choosing intensive colonoscopic surveillance (eg, at 6-month intervals) should be advised of the risk of developing advanced CRC despite this intensive surveillance. Furthermore, reliance on detection of dysplasia alone to predict CRC cannot protect all patients, because 26% of patients who have carcinoma identified at surgery do not harbor concomitant dysplasia .

Dysplasia-associated Lesions or Masses Versus Adenoma-like Masses

Patients who have IBD can develop polypoid lesions that resemble sporadic adenomas, which have been called “adenoma-like masses” (ALMs). In two longitudinal studies, patients who had IBD with colonoscopically resected polyps and who had no evidence of dysplasia adjacent to the polyp or elsewhere in the colon did not develop dysplasia or cancer during 4 years of follow-up . In a subsequent 7-year follow-up of one of these studies, 58% developed further ALMs, an incidence that was similar to the incidence of sporadic adenomas in the control group of patients who did not have ulcerative colitis, and only one study patient (who had concomitant PSC) developed carcinoma . Thus, pedunculated polyps occurring within areas of colitis or elsewhere in the colon can be managed safely by colonoscopic polypectomy without colectomy, provided biopsies of adjacent mucosa reveal no dysplasia .

Guidelines for Colonoscopic Surveillance in Irritable Bowel Disease

Table 2 lists the surveillance approaches currently recommended by various national and international societies . A surveillance program also should include periodic office visits with physical examination and routine laboratory testing. Surveillance colonoscopy should be considered starting after 7 or 8 years of disease. If possible it should be performed when the colitis is in clinical remission to minimize any misinterpretation of inflammatory change as dysplasia. To maximize detection of neoplasia, jumbo biopsy forceps should be used, two to four biopsies should be obtained every 10 cm for a minimum of 33 biopsies, and additional targeted biopsies should be performed for any raised lesions (suspected DALMs) or strictures . Strictures in the setting of ulcerative colitis should be considered neoplastic until proven otherwise, and strictures in Crohn’s disease should be biopsied or brushed to exclude malignancy.

Gastroenterologists do not seem to adhere strictly to these guidelines . A questionnaire survey of British gastroenterologists showed that all consultants recommended surveillance colonoscopy, but only 4% recommended colectomy for LGD, and, surprisingly, only 53% recommended colectomy for high-grade dysplasia . A questionnaire survey found that 80% of members of the American Gastroenterology Association recommend surveillance colonoscopy beginning at 8 to 10 years of disease, but only half reported performing at least 31 biopsy specimens per colonoscopy, and only 40% recommended immediate colectomy for patients who have LGD . Despite published guidelines, gastroenterologists need further education about performance standards for surveillance colonoscopy. These two surveys also demonstrate a lack of consensus among practicing gastroenterologists regarding management of LGD.

Advanced Endoscopic Techniques

Although standard colonoscopy with 30 or more random biopsies is the reference standard for surveillance, it is imperfect. It requires extra time to obtain and process all the biopsy specimens. The yield of dysplasia per biopsy is low. Hurlstone and colleagues reported only 0.14% of approximately 12,000 nontargeted biopsies showed intraepithelial neoplasia (dysplasia), and Rutter and colleagues found no instances of intraepithelial neoplasia among 2904 biopsies in 100 patients. Several methods to enhance the efficiency of colonoscopic detection of intraepithelial neoplasia in IBD have been developed, including chromoendoscopy, with or without magnification endoscopy, narrow-band imaging, and optical coherence tomography .

Chromoendoscopy, the most widely studied technique, involves the topical application of dyes, most commonly indigo carmine or methylene blue, to enhance mucosal detail and to facilitate detection of abnormalities that should undergo targeted biopsy . Several studies have shown improved yields for intraepithelial neoplasia in patients who have ulcerative colitis undergoing chromoendoscopy with targeted biopsies compared with conventional colonoscopy with nontargeted biopsies. In a study of 263 patients who had ulcerative colitis randomly assigned to either conventional colonoscopy or chromoendoscopy with methylene blue, the yield for intraepithelial neoplasia was threefold higher in the chromoendoscopy group . In a similar study of 100 patients undergoing back-to-back conventional colonoscopy versus chromoendoscopy with indigo carmine, Rutter and colleagues found no dysplasia by conventional nontargeted biopsies compared with nine dysplastic lesions in the indigo-carmine group and found that chromoendoscopy required many fewer biopsies.

The use of high-magnification colonoscopes in conjunction with chromoendoscopy further enhances the yield for intraepithelial neoplasia and can allow preliminary classification of lesions as neoplastic or non-neoplastic based on chromoendoscopic appearance using a pit-pattern classification . In a recent controlled study of 161 patients who had ulcerative colitis randomly assigned to conventional surveillance colonoscopy versus chromoendoscopy with methylene blue and endomicroscopy, fivefold more neoplastic lesions were found using chromoendoscopy and endomicroscopy than with conventional colonoscopy .

Emerging experimental colonoscopic techniques for enhanced neoplasia detection include narrow-band imaging, optical coherence tomography, fluorescence endoscopy, and confocal laser endomicroscopy . More clinical studies of these techniques are needed. Furthermore, most of these techniques require specialized colonoscopes and additional endoscopic training and are more time consuming than conventional colonoscopy; these requirements increase the cost of colonoscopic surveillance.

Primary Prevention Strategies

Patients who have IBD are at increased risk of death from colon cancer because current surveillance strategies are imperfect in detecting early cancer or dysplasia. Pharmacotherapy to decrease this risk is an attractive option. Several drugs have been tried, but direct proof of efficacy for any primary prevention strategy is lacking because of the long required duration of exposure to a potential agent, the relatively low risk of CRC, and the difficult of eliminating confounding factors. The available data come primarily from retrospective observational studies comparing the incidence of cancer in patients using a medication and in those not using this medication.

5-Acetyl Salicylic Acid Compounds

In a nested case-control study of 3112 patients, including 102 cases of colon cancer, sulfasalazine administered for at least 3 months significantly decreased the risk of CRC (relative risk, 0.34) . In a similar case-control study of 102 cases of CRC complicating IBD, regular use of 5-ASA reduced the cancer risk by 75% . In this study, the risk of CRC was reduced dramatically (by 81%) in patients administered more than 1.2 g/d of mesalamine, but simply visiting a doctor more than twice a year had an even greater protective effect (OR, 0.16). Another study of 100 patients who had IBD and CRC also showed significant risk reductions with regular use of 5-ASA . Some studies have not shown this protective effect. A case-controlled retrospective study of 25 cases of IBD and cancer showed no relationship between mesalamine use or dosage and cancer risk . In a similar study of 364 patients who had IBD and cancer, using data extracted from two large insurance claims databases, the use of 5-ASA at any dose or duration for the year preceding the diagnosis of cancer did not reduce the cancer risk . A meta-analysis of nine observational studies, including 334 cases of CRC complicating IBD, showed a decrease in the risk of CRC among patients receiving 5-aminosalicylates (OR, 0.51) . In conclusion, 5-ASA seems to be beneficial in preventing cancer and is useful as a maintenance therapy to prevent flares of colitis with minimal toxicity .

Ursodeoxycholic Acid

Ursodeoxycholic acid (UDCA) is used commonly to treat PSC. One study of 59 patients who had ulcerative colitis and PSC showed a significantly lower incidence of colonic dysplasia in those administered UDCA (OR, 0.18) . Similar results were seen in a study of 42 patients who had PSC and ulcerative colitis in which patients receiving UDCA had a much lower incidence of developing dysplasia or colon cancer (OR, 0.26) . The use of UDCA in patients who had ulcerative colitis and PSC therefore is reasonable. Whether this apparent benefit extends to patients who have ulcerative colitis but do not have PSC is unknown. The mechanism of this apparent chemoprotective effect also is unknown.

Folic Acid

Low folate levels have been identified as a risk factor for sporadic CRC , and patients who have IBD have an increased risk of folic acid deficiency because of decreased intake, decreased intestinal absorption, or the use of concomitant medications such as sulfasalazine that interfere with folic acid metabolism. Studies in patients who have IBD given supplemental folic acid have shown a statistically nonsignificant trend toward a reduced risk of cancer and dysplasia .

In summary, 5-ASA compounds seem to decrease the risk of dysplasia and cancer in patients who have IBD. In patients who have PSC and ulcerative colitis, UDCA seems to have a beneficial effect. The data on the benefits of folic acid supplementation for cancer prevention in IBD are weak.

Personal History of Colon Cancer

The recurrence of colon cancer recurrence after curative resection may result from missed synchronous lesions, new metachronous lesions, or anastomotic recurrence. In a prospective study of patients undergoing colonoscopy either preoperatively or within 6 months postoperatively, synchronous cancers occurred in 5% and adenomatous polyps occurred in 28% of patients . Many of these cancers occurred at a site distant from the original malignancy and would not have been included in the original curative resection. Passman and colleagues similarly reported a 3.3% incidence of synchronous cancers, with most being stage I or II. Metachronous colon cancers occur at a higher frequency in patients who had prior colon cancer than in the general population. In one study, 42 metachronous cancers developed during a total of 15,345 person-years of follow-up for a cumulative incidence of 1.5% at 5 years . More than one half of metachronous lesions were found within 2 years of the original colonic resection, suggesting that some of these apparently metachronous lesions actually are missed synchronous lesions . Patients undergoing annual surveillance colonoscopy after curative resection were found to have a 5% risk of anastomotic recurrence and a 2% risk of metachronous cancers after 6 years . Of the metachronous lesions, 80% were found within 2 years of follow-up, and all these were resectable for cure, whereas none in the anastomotic recurrent group underwent curative resection. The benefit of colonoscopy in the first 2 years lies primarily in detecting metachronous cancers rather than anastomotic recurrence . The optimal interval for surveillance colonoscopy after the first year is debated, but several studies have not shown a survival benefit from annual follow-up colonoscopies .

Rectal cancer has a higher incidence of local recurrence than proximal cancers (20%–30% versus 2%–4%), and surgical excision with neoadjuvant chemoradiation is known to lower recurrence rates significantly (by about 10%) . Frequent proctoscopic evaluations or rectal endoscopic ultrasounds are recommended after low anterior resection for rectal cancer . Current guidelines for surveillance colonoscopy after curative surgery for colon cancer from the American Cancer Society and the United States Multi-Society Task Force on Colorectal Cancer are listed in Table 3 .

Personal History of Adenomatous Polyps

Patients who have a history of adenomatous polyps are at increased risk of subsequently developing colonic adenomas and colon cancer . Risk factors for subsequent adenomas include villous histology , multiplicity of polyps and increased polyp size . Right-sided adenomas may carry a higher risk than left-sided adenomas . Lesions found on index colonoscopy are categorized as having low risk for subsequent neoplasia (one or two tubular adenomas with LGD) or high risk (more than three adenomas of any size or a single adenoma larger than 1 cm, villous histology, or high-grade dysplasia) . The United States Multi-Society Task Force on Colorectal Cancer recommends follow-up colonoscopy at 3 years for patients who have high-risk lesions . If the follow-up colonoscopy demonstrates no polyps, the surveillance interval can be extended to 5 years. If low-risk lesions are detected on the index colonoscopy, follow-up may be delayed from 5 to 10 years . The current recommendations of the United States Multi-Society Task Force on Colorectal Cancer and the American Cancer Society are shown in Table 4 .

Family History of Adenomatous Polyps or Cancer

Several studies have documented an increased risk of colorectal neoplasia in patients who have a family history of colon cancer or adenomatous polyps . The risk is increased further when two or more family members are affected or when the neoplasia was diagnosed before the age of 60 years . Two recent meta-analyses have confirmed these findings . For persons with a first-degree relative who had colon cancer or two first-degree relatives who had colon cancer, the current American Gastroenterology Association guidelines ( Table 5 ) recommend screening colonoscopy every 5 years beginning at age 40 or 10 years younger than the age at diagnosis of the relative affected with CRC . For individuals with a first-degree relative who had adenomatous polyps diagnosed before the age of 60 years or two second-degree relatives who had colon cancer, standard average-risk guidelines apply, but screening is recommended beginning at age 40 years.