AGA Technical Review on the Diagnosis and Management of Colorectal Neoplasia in Inflammatory Bowel Disease
Article Outline
- Learning Objectives
- Are Patients With IBD at Increased Risk for Developing CRC?
- Are There Well-Substantiated Factors Other Than Dysplasia That Increase or Decrease the Risk of CRC in IBD?
- What Is the Natural History of Dysplasia?
- Should Colectomy Be Performed for Raised Dysplasia?
- Should Colectomy Be Performed for Flat Dysplasia?
- Is There Sufficient Rationale for Performing Surveillance Colonoscopy in Patients With IBD?
- How Should Surveillance Colonoscopy Be Performed?
- What Role Do the Newer Imaging Techniques Play in Identifying and Managing Dysplasia?
- Should Chemopreventive Agents Be Used to Lower the Risk of Developing Dysplasia or CRC in IBD?
- Should Molecular Markers Be Applied to Help Stratify Patients Into Low-Risk and High-Risk Groups?
- Areas of Future Research
- Acknowledgments
- References
- References (Online Only)
- Copyright
Abbreviations used in this paper: AMACR, α-methylacyl-CoA racemase, AZA, azathioprine, CI, confidence interval, CRC, colorectal carcinoma, DALM, dysplasia-associated lesion or mass, HGD, high-grade dysplasia, LGD, low-grade dysplasia, 6-MP, 6-mercaptopurine, MSI, microsatellite instability, OR, odds ratio, PSC, primary sclerosing cholangitis, RR, relative risk, SIR, standardized incidence ratio, STn, sialyl-Tn antigen, USPSTF, US Preventive Services Task Force
Podcast interview: www.gastro.org/gastropodcast.
Learning Objectives
This article has an accompanying continuing medical education activity on page e12. Upon completetion of reading this article, successful learners will be able to:
Colorectal carcinoma (CRC) complicating ulcerative colitis (UC) was first recognized in 1925 by Crohn and Rosenberg,1 but it was not until 1948 that Warren and Sommers reported CRC in a patient with Crohn's disease (CD).2 There has been much dispute regarding the magnitude of risk in both of these conditions. For many years it was believed that the risk in CD was insignificant. However, it is now recognized that the risk of developing CRC is equivalent, in both conditions given a similar extent and duration of disease.
Inflammatory bowel disease (IBD) is relatively rare in the general population. Consequently <1% of all cases of CRC are attributable to IBD. However, it remains 1 of the 3 high-risk conditions predisposing to CRC, along with familial adenomatous polyposis and Lynch Syndrome. Patients have up to a 1 in 5 chance of developing CRC after 30 years of disease.3 Thus, it is an important issue for both the patient and the physician. The risk is not equivalent for all patients and depends on a number of factors. This necessitates an individualized and sensible approach to surveillance in patients with IBD.
Are Patients With IBD at Increased Risk for Developing CRC?
UC
Patients with long-term UC have an increased risk of CRC, but the magnitude has been difficult to estimate. A number of factors have rendered the magnitude difficult to assess. First, a direct comparison between studies is difficult because of inconsistent methods used to calculate risk. Some studies reported the cumulative risk of developing CRC in a given population of patients with IBD, but unfortunately, many assume that all subjects have the same risk. Other studies have calculated the risk of CRC in IBD cohorts as a standardized incidence ratio (SIR) compared with a control population. These estimates can be adjusted for age and gender but do not provide information on the lifetime risk.
Second, inherent selection biases have affected earlier studies. Most studies were from tertiary referral centers. Case reports and population studies from these centers included patients with more severe recalcitrant disease, and also those who had been referred already with a diagnosis of cancer. In addition, risks were related to the hospital population rather than the population of the host community. More recently, population-based studies that covered defined geographical areas lean more toward conservative, but more accurate, risk estimates. However, some did not distinguish extent of disease, which has led to less informative conclusions. Studies that do stratify the data on key variables should be viewed as the gold standard.
Third, practices regarding treatment of UC differ across continents. Some centers have an aggressive policy with respect to use of aminosalicylates and/or early colectomy. Recent research has emerged that suggests a protective role for aminosalicylates in the prevention of CRC. Centers that adopted this policy of mesalamine prophylaxis long ago may well have modified the risk to their patients. In centers with high colectomy rates, lower cancer incidence rates would be expected because the procedure virtually eliminates the risk. Consequently, it is not surprising that the risk of CRC has been reported to be as low as 1.4% at 18 years in a Scandinavian cohort study of 783 patients,4 and as high as 34% after 30 years of disease in a tertiary referral center study of 267 patients.5
In an attempt to determine the overall risk, a meta-analysis of studies that reported colonic cancer risk in UC has been conducted.3 Initially, the meta-analysis identified 194 studies. Of these, 116 met the inclusion criteria from which a minimum amount of data (the number of patients and cancers detected) could be extracted. Nineteen of these studies reported stratified data (ie, reported the rates of cancer at 10, 20, and 30 years of disease). The meta-analysis showed that the risk is 2% at 10 years, 8% at 20 years, and 18% at 30 years of disease. The St Mark's group, in the United Kingdom, has since reported further data from their 30-year surveillance program.6 They have reported similar, although slightly lower, cumulative incidence rates of cancer and dysplasia of 7.7% at 20 years and 15.8% at 30 years. More recent population-based studies, such as data from Jess et al,7 have suggested that although the risk is increased in patients with extensive colitis (SIR, 2.4; 95% confidence interval [CI], 0.6–6.0), the risk has decreased over time. Several others studies have reported a lower relative risk (RR) of developing CRC (RR ranging from 1.8 to 2.8). These include studies from Canada, Italy, and Hungary and have been summarized in a review article by Loftus.8
CD
Until recently, the risk of CRC in CD was unclear. Early studies included patients with colonic resections; some included patients with isolated small bowel disease, and others did not adjust for the duration of disease.9, 10 The risk reported from studies that represent “at-risk populations” (ie, those with long-standing, unresected, and extensive colonic CD) provide estimates of increased risk.10, 11 For instance, in a large population-based survey of 1655 patients by Ekbom et al,11 a RR of 2.5 (95% CI, 1.3–4.3) for CRC was reported. However, for the 830 patients with Crohn's colitis, the RR was 5.6 (95% CI, 2.1–12.2). A recent study from Denmark provided contradictory results. Jess et al12 extended the study by Munkholm et al10 by reexamining the same group of patients with a longer follow-up period (17 years). The risk of CRC was not increased in the total group of patients or in patients with only colonic CD (standardized mortality ratio, 1.64; 95% CI, 0.2–5.92). However, 2 factors need to be considered. The cumulative colectomy rate remained 20% after 20 years of disease, and long-term maintenance therapy with mesalamine drugs has been practiced in this region for decades. These agents are believed to have chemopreventive properties and, thus, may have reduced the incidence of cancer.13, 14, 15 The same group conducted a meta-analysis of 6 studies to estimate the risk of CRC in CD.16 The pooled SIR for CRC was significantly increased (SIR, 1.9; 95% CI, 1.4–2.5), as was the risk of CRC independently (SIR, 2.5; 95% CI, 1.7–3.5). This was a meta-analysis of population-based studies, and 3 of the 6 reports were performed in Scandinavian countries. A further meta-analysis of 12 studies showed an overall CRC RR in CD patients of 2.5 (95% CI, 1.3–4.7). However, for patients with colonic disease, the RR increased to 4.5 (95% CI, 1.3–14.9).17
Comparing the Risk in UC and CD
Without adjusting for potential biases, the largest population-based study, from Manitoba, Canada, demonstrated a risk of CRC in patients with CD (RR, 2.64; 95% CI, 1.69–4.12) equal to patients with UC (RR, 2.75; 95% CI, 1.91–3.97).18 In fact, absolute cumulative CRC frequencies for CD and UC have been shown to be nearly identical: 8% for UC and 7% for CD after 20 years of disease.19 The latest data from Olmstead County, Minnesota, did not reach statistical significance.7 However, this study found a SIR of 2.4 for patients with pancolitis UC (95% CI, 0.6–6.0), and a SIR of 1.9 for those with CD (95% CI, 0.7–4.1).
Therefore, it is now accepted that the risk of cancer is equivalent in both conditions.11 In a study of 28 patients with CD-associated CRC and 52 with UC-associated CRC, the age at diagnosis of cancer (CD, 54 years; UC, 43 years), the duration of IBD before cancer (CD, 15 years; UC, 18 years), the multiplicity (CD, 11%; UC, 12%) and distribution of cancers, the presence of dysplasia (CD, 73%; UC, 79%), and the overall 5-year survival rates were similar (CD, 46%; UC, 50%).20 Crohn's colitis should raise the same concerns regarding the risk of developing cancer as does UC.
Are There Well-Substantiated Factors Other Than Dysplasia That Increase or Decrease the Risk of CRC in IBD?
Disease Duration
The increasing risk of CRC with disease duration in patients with UC has been demonstrated in the meta-analysis and surveillance program data previously mentioned.3, 6 An elevated RR is appreciable after 8 to 10 years of disease, which is the time at which regular colonoscopic surveillance should commence. A recent study from The Netherlands has suggested that cancers will be missed if surveillance is commenced at 8 to 10 years for patients with pancolitis, and 15 to 20 years for patients with left-sided disease because 9% to 15% of cancers in their study occurred before this time.21 However, the vast majority of studies show that the incidence is very low at 10 years of disease, and may be decreasing.3, 6, 7, 8 Thus, commencing surveillance prior to 8 years of disease increases the cost of a surveillance program with very little benefit.
Previous guidelines have recommended that the surveillance interval should be shortened commensurate with increasing duration of disease. Recent data suggest that it may not be necessary to intensify surveillance with increasing duration of disease, although this remains controversial.6
Regarding CD, in a hospital-based study from Copenhagen of 373 patients,10 patients with CD and long disease duration had a RR of 4.8 (P = .01); in long-standing cases in which surgery was not performed, the RR increased to 8.3 (P = .005). Similarly, Gillen et al conducted a hospital-based study in Birmingham, England.22 They assessed 281 patients with CD, 125 of whom had extensive Crohn's colitis. All 9 cancers in this series occurred after 12 years of disease; 6 of the 8 cases were diagnosed after 20 years of disease. Overall the RR was 3.4 (95% CI, 1.5–6.7), but this increased to 18.2 (95% CI, 7.8–35.8) in patients with long-standing CD and intact colon.
Thus, duration of disease is recognized to be one of the most important risk factors for development of CRC in patients with IBD.
Anatomic Extent of Disease
Extent of disease is another major risk factor. It is defined by the most extensive disease identified histologically or endoscopically at any time during the patient's illness. Data related to the risk of left-sided colitis should be viewed with caution because of variability in the definition of left-sided colitis, and varying methods used to determine extent of disease between different reports. For instance, the definition of left-sided colitis included patients with disease up to the splenic flexure in some reports, but in others, it included patients with disease up to the hepatic flexure. In addition, many studies were performed in the precolonoscopic era, relying on single- or double-contrast barium enemas to assess extent of disease.
Most cancers arise in patients with subtotal or pancolitis (which is generally defined as extension beyond the hepatic flexure). There is general agreement that there is little or no increased risk associated with proctitis or proctosigmoiditis, whereas left-sided colitis carries an intermediate risk of cancer.23, 24 For example, in a combined British and Swedish study,23 the excess risk observed in patients with pancolitis was 19.2. However, for those with left-sided disease, the risk was 2.8. Similarly, in a separate population-based Swedish cohort study,24 the RR for patients with pancolitis, left-sided colitis, and proctitis was 14.8 (95% CI, 11.4–18.9), 2.8 (95% CI, 1.6–4.4), and 1.7 (95% CI, 0.8–3.2), respectively. The development of cancer in patients with left-sided colitis is not as frequent as in patients with pancolitis during the first 2 decades of disease, but the incidence in these 2 groups is virtually equal by the fourth decade of disease.5, 23, 25
Heuschen et al evaluated backwash ileitis (the presence of mucosal inflammation proximal to the ileocecal valve) in UC.26 The group prospectively analyzed 590 consecutive patients with UC who had a restorative proctocolectomy. They found that the rate of cancer was more than 3 times higher in patients who had pancolitis and backwash ileitis compared to patients without ileitis (29% vs 9%). In addition, the presence of backwash ileitis was associated with multiple tumors (45% vs 24%). However, that study did not use strict pathologic criteria for backwash ileitis and may have included patients with CD in the analysis. In contrast, in a more recent study of 200 consecutive patients with clinically and pathologically confirmed UC, and in whom CD was excluded, the presence of ileitis was not associated with an increased rate of dysplasia or cancer in the colectomy specimens. Only one patient with ileitis (3%) and 4 control patients (2%) had carcinoma in their colonic resection specimens.27 Thus, at the present time, there is insufficient evidence that backwash ileitis is a risk factor for CRC in IBD.
Primary Sclerosing Cholangitis
A small subset (approximately 2%–5%) of patients with UC also develop primary sclerosing cholangitis (PSC). In 1992, Broome et al first reported an increased risk of CRC in patients with PSC; in a further case-control study, the absolute cumulative risk of developing CRC or dysplasia in patients with UC and PSC was 9% after 10 years of colitis, 31% after 20 years, and 50% after 25 years, compared with rates of 2%, 5%, and 10% in UC controls matched for duration and disease extent (P < .001).28 This association has been supported by studies from other centers as well.29, 30, 31, 32, 33, 34 In contrast, a case-control study by Loftus et al did not find a significantly increased RR of CRC in 178 patients with both PSC and UC.35 A further case-control study from the Mayo Clinic36 of 171 patients with UC found no association between PSC and CRC (odds ratio [OR], 1.23; 95% CI, 0.62–2.42). A meta-analysis was conducted by Soetikno et al to determine the overall risk of CRC in patients with both PSC and UC.37 Twenty-two pertinent studies were found, but only 11 met the inclusion criteria. The analysis indicated an approximate 4-fold increased risk of colorectal neoplasia in patients with PSC and UC, compared to patients with UC alone.
An increased risk of CRC following orthotopic liver transplantation for PSC in patients with UC has also been documented, with the incidence ranging from 5.6% to 11.1%38, 39 or approximately 1% per person per year.40 A recent study from the United Kingdom evaluated 100 patients with both PSC and IBD (92 UC, 8 CD) who underwent liver transplantation.41 Eighty-three of these patients had an intact colon. Their results are even more alarming, with a cumulative risk of developing CRC of 14% and 17% after 5 and 10 years posttransplantation, respectively. In addition, patients with PSC may have had subclinical UC for many years prior to diagnosis. For these reasons, it is highly recommended that patients with UC (or Crohn's colitis) and PSC undergo annual colonoscopy beginning at the time of diagnosis of PSC, and continue indefinitely, even after liver transplantation.
Family History of Sporadic CRC
A positive family history of sporadic CRC is associated with a 2- to 3-fold increased risk of CRC in the general population. This has also been found to be a risk factor in patients with UC. A case-control study from the Mayo Clinic of 297 patients found that a positive family history of sporadic CRC was twice as common in patients with colitis and CRC compared to UC controls matched for extent and duration of colitis.42 Askling et al also found that a positive family history of CRC was associated with a more than 2-fold increased risk of CRC in patients with IBD (RR, 2.5; 95% CI, 1.4–4.4).43 Patients with IBD and a first-degree relative diagnosed with CRC before the age of 50 years had an even higher risk (RR, 9.2; 95% CI, 3.7–23). In another study of more than 30,000 IBD cases from the same research group, relatives of patients with both IBD and CRC had an 80% increased risk of CRC.44 However, first-degree relatives of patients with only IBD were not at increased risk for CRC (SIR, 0.9; 95% CI, 0.82–0.97). The abundance of evidence suggests that both genetic and acquired factors are important for the development of CRC in patients with IBD.
Age of Onset of IBD
There is some debate as to whether early age of onset of colitis is related to CRC risk. Younger patients have a longer potential life span, and, thus, higher risks may reflect longer duration of disease.
The first study to report a higher risk in children followed 396 patients, all aged 14 years and younger, who were first seen at the Mayo Clinic between 1919 and 1965.45 They showed a 3% cancer incidence during the first decade of disease, and a 20% incidence during each of the second and third decades of disease. This study probably included an element of referral bias, but these findings have now been confirmed by others.23, 24 Ekbom et al24 found that the cumulative risk of CRC in patients with extensive colitis, after 35 years of follow-up, was 40% in patients in whom the disease started before the age of 15 years, and 25% in patients who developed colitis between 15 and 39 years of age. Markowitz et al have also reported the results of a colonoscopic screening program in young patients with either UC and CD (UC, 18; CD, 17).46 The mean age of patients was 21 years at the time of colonoscopy, and the mean duration of disease was 11 years. Seven patients had aneuploidy, 2 had dysplasia, and one patient had Dukes' C carcinoma at the age of 21 years, and also had UC for 14 years.
Early age of onset of disease is not a consistently reported risk factor for CRC in IBD. Greenstein et al5 calculated age-specific incidences per 1000 patient-years, for patients with universal or left-sided colitis. The investigators estimated an incidence of 3.6 per 1000 patient-years in pancolitic patients who were 10 to 19 years of age at onset of disease, compared to 12.7 per 1000 patient-years in patients who were between 30 and 39 years of age at onset. The apparent increased risk of CRC in patients older than 40 years of age may reflect some contribution from the general age-related risk of developing sporadic CRC. Support for this view comes from a report19 showing that the interval between onset of disease and development of cancer is similar in young compared with older patients.
Similarly, a study of colonoscopic surveillance in patients with CD did not suggest an additional risk of neoplasia in those with early-onset disease.47 The study included 259 patients with disease duration exceeding 8 years who had at least one third of the colon involved. One third of patients had a segmental resection before screening. Dysplasia was seen in 16 patients (6.2%). Forty-seven patients in the study had disease onset before 15 years of age. The mean duration of disease in this subgroup was 22 years. Only 2 patients in this subgroup had low-grade dysplasia (LGD; 4.3%).
The low incidence of CRC in young patients in the general population will undoubtedly influence calculations of SIRs. Weedon et al9 conducted a study at the Mayo Clinic of 449 patients with CD diagnosed before the age of 21 years. Only 8 cancers were detected, but this resulted in a large RR (26.6). A subgroup analysis of younger patients with CD (onset at younger than 30 years of age) enrolled into a large population-based Swedish study found similar results.11 The SIR was 9.5 for all younger patients. The RR increased to 20.9 in patients in that cohort with colonic disease. Similar results were found by Gillen et al, who also reported a RR of 57.2 in patients with extensive colonic CD, and early age of disease onset.22
Although controversial, children diagnosed with IBD have at least the same rate of developing CRC during the first several decades of disease as patients diagnosed between the ages of 21 and 40 years. Thus it is advisable to have an active policy toward surveillance in this group. Surveillance should be performed as frequently in children as it is in adults, and should be based on the duration of illness, not chronological age.
Severity of Endoscopic and Histologic Inflammation
Although it is likely that development of CRC is related to the underlying inflammatory process, case-control studies that analyzed the severity and frequency of symptomatic colitis attacks in patients with UC did not confirm inflammation as a risk factor.13, 15 The reasons may be 2-fold. First, patients with severe disease unsuccessfully treated with medical therapy are more likely to have a colectomy earlier in the course of their disease. Second, technical difficulties involved with tracking disease severity (by clinical or histologic data) in retrospective studies on large populations are quite formidable. Evidence is mounting to suggest that mesalamine reduces the risk of development of dysplasia and CRC in IBD,13, 14, 15 although it is not yet clear whether this is by reducing inflammation or if it is due to some other type of antineoplastic effect (see the section on use of chemopreventive agents). In addition, the risk of CRC is related to both increasing disease duration and extent of disease. Paradoxically though, if there were a simple relationship between inflammation and cancer, a higher incidence of rectal cancer in patients with proctitis would be expected, which does not seem to be the case.
Recent data from a case-control study of 68 patients with UC which defined inflammation at a colonoscopic and histologic level has indicated that greater degrees of colonoscopic (OR, 2.5; P < .001) or histologically active (OR, 5.1; P < .001) inflammation are associated with an increased risk of CRC.48 In that study, the degree of histologic inflammation was assessed semiquantitatively on a 5-point scale (normal, inactive, mild, moderate, or severe). This variable was shown to be positively correlated with an increased risk of neoplasia in both univariate and multivariate analyses, with ORs of 5.13 and 4.7, respectively. In a cohort study of 418 patients with colitis, Gupta et al analyzed the degree of histologic inflammation in mucosal biopsy specimens using a 4-point scale (normal or inactive, mild, moderate, or severe) and correlated this with the development of neoplasia on follow-up.49 In that study, 65 patients developed neoplasia (LGD, HGD, or carcinoma) but only 17 developed “advanced” neoplasia, defined as either HGD or cancer. By both univariate and multivariate analyses, patients' overall inflammation score was positively associated with the development of advanced neoplasia (hazard ratio, 3.0; 95% CI, 1.4–6.3). However, no significant association was noted between the degree of inflammation and the development of “any neoplasia” (which included the category of LGD). In a third study reported in abstract form, colonic inflammatory scores were evaluated in 56 patients with UC who also had dysplasia and CRC, and 90 control patients matched for extent and duration of disease and age. The mean inflammatory activity was higher in the cases with neoplasia (P = .013), corresponding to an OR of 1.71 (95% CI, 1.15–3.18) in the univariate analysis. Multivariate analysis, adjusting for a family history of CRC, smoking status, and mesalamine and immunomodulator use, also showed an increased risk of dysplasia and CRC with increased inflammatory activity (OR, 2.73; 95% CI, 1.44–5.18; P = .002).50
None of these studies compared the occurrence of neoplasia in patients with histologically normal mucosa to those with chronic inactive colitis. Nevertheless, CRC may develop in areas of mucosa in which “active” colitis has undergone remission to a normal or chronic inactive state (ie, areas in which the mucosa reveals histologic findings of inactive colitis, such as crypt architectural distortion or Paneth cell metaplasia). Of course, lack of endoscopic involvement of a portion of colon in which dysplasia and/or cancer has developed does not mean that the area was devoid of inflammation at a prior time in the patient's disease. For instance, Mathy et al reviewed histologic sections of 30 colectomy specimens from patients with UC-associated dysplasia or CRC. They found that CRC may arise in endoscopically normal, but histologically involved, areas of colon.51 In that study, neoplasia was not detected in areas of colon that were not involved with the inflammatory process. Thus, histologic involvement, rather than endoscopic (macroscopic) appearance of disease, is a better determinant of extent of disease with regard to evaluating cancer risk. For example, a patient with endoscopic disease up to 10 cm, but with chronic or active histologic inflammation up to 40 cm, should be considered to have left-sided colitis, not proctitis, for the purpose of endoscopic surveillance.
In CD, there is a convincing association between inflammation and cancer at various sites. Small bowel adenocarcinoma is extremely rare, but in Copenhagen a population-based study of 374 patients11 demonstrated a more than 60-fold risk (OR, 66.7; 95% CI, 18.1–170.7). Investigators have also noticed a higher than expected frequency of malignant neoplasms in perianal fistulae, excluded segments of bowel, and strictures in patients with CD.52, 53
Colonic strictures in patients with Crohn's colitis are a consequence of the transmural inflammatory process, and by itself, should not raise suspicion of neoplasia. However, in patients with UC, strictures are considered a risk factor for CRC. A retrospective study by Lashner et al of patients with UC in an IBD registry found that 3.2% (15 patients) had a true stricture.54 These were identified at 13.3 ± 9.9 years following the diagnosis of UC. Eleven patients had multiple strictures principally located in the left colon. Of the 15 patients, 11 had dysplasia and 2 had cancer. Ultimately, 6 patients had carcinoma detected at colonoscopy or colectomy (3 modified Dukes' stage A, 1 stage B, and 2 stage D). All cancers were located at the site of a stricture. Data from Rutter et al55 also highlight a higher than expected frequency of malignant neoplasms in patients with strictures (OR, 5.7; 95% CI, 1.7–18.9). Similar results have been found by other investigators.56, 57 In addition, the data from Rutter et al reveal an increased risk of cancer in patients with a shortened tubular colon (OR, 28.4; 95% CI, 1.6–512.2) as a result of ongoing inflammation. A study of patients with UC from Mount Sinai Hospital in New York elucidated 3 factors associated with malignancy in strictures: (1) appearance late in the course of UC, (2) location proximal to the splenic flexure, and (3) symptomatic large bowel obstruction.58 In that study, cancers associated with a stricture were more likely to be at an advanced stage compared with cancers that did not result in strictures. A stricture and foreshortened colon should, therefore, be considered strong risk factors for cancer, requiring intensive colonoscopic surveillance.
The colonoscopic appearance of the colon is also important, particularly with respect to raised lesions. DALMs (dysplasia associated lesions or masses) and inflammatory polyps are the main type of raised lesions. DALMs are discussed later in this article. Colonoscopically, neoplastic lesions should be differentiated from inflammatory lesions, preferably by biopsies.55, 59, 60, 61 Studies from Rutter et al55 and Velayos et al61 have shown that the presence of postinflammatory pseudopolyps doubles the risk of CRC (OR of 2.14 [95% CI, 1.24–3.7] and 2.5 [95% CI, 1.4–4.6], respectively). The presence of pseudopolyps likely indicates prior severe inflammation. Rather than the pseudopolyp itself being the origin of the cancer, a more likely explanation of this phenomenon is that it is more difficult to discern a dysplastic lesion in a field of pseudopolyps. Surveillance has reduced effectiveness in the context of multiple pseudopolyps.
Pharmacologic Therapy
The use of pharmacologic therapy to treat flares of IBD and to maintain remission may modulate the risk of developing dysplasia and CRC. This topic is reviewed later in the section on chemoprevention.
A summary of the risk factors and their absolute risks and RRs associated with the development of dysplasia and CRC in patients with IBD are shown in Table 1.
Table 1. Risk Factors for Development of Dysplasia and CRC in IBD
| Risk factor | Absolute risk | RR (95% CI) |
|---|---|---|
| Disease duration (UC) | CRC at 10 y = 2%3 | 2.4 (0.6–6.0)7 |
| CRC at 20 y = 8% | 2.8(1.91–3.97)18 | |
| CRC at 30 y = 18% | ||
| CRC/dysplasia at 20 y = 8%6 | ||
| CRC/dysplasia at 30 y = 16% | ||
| CRC at 20 y = 2.5% | ||
| CRC at 30 y = 7.6% | ||
| Disease duration (CD) | ND | 2.5(1.7–3.5)16 |
| 2.6(1.7–4.1)18 | ||
| 4.5(1.3–14.9)17 | ||
| Extent of colonic involvement | ND | |
| Pancolitis | 14.8(11.4–18.9)24 | |
| Left-sided disease | 2.8(1.6–4.4) | |
| Proctitis | 1.7(0.8–3.2) | |
| Presence of PSC | CRC at 10 y = 9%28 | 4.8(3.9–6.4)37 |
| CRC at 20 y = 31% | ||
| CRC at 25 y = 50% | ||
| Family history of sporadic CRC | ND | |
| In a first-degree relative older than 50 y | 2.5(1.4–4.4)43 | |
| In a first-degree relative younger than 50 y | 9.2(3.7–23.0) | |
| Young age at disease onset | ND | |
| Disease commencing at age younger than 15 y | 40%24 | |
| Disease commencing at age older than 15–39 y | 25% | |
| Inflammation | 2.5(1.5–4.4)48 | |
| Colonoscopic | ND | 5.1(2.7–11.1)48 |
| Histologic | 3.0(1.4–6.3)49 | |
| Backwash ileitis | ND | ND |
| Presence of a stricture | 3.2%54 | 5.7(1.7–18.9)55 |
| 5.1%150 | ||
| Presence of a shortened colon | ND | 28.4(1.6–512.2)55 |
| Presence of postinflammatory pseudopolyps | ND | 2.1 (1.2–3.7)552.5 (1.4–4.6)61 |
What Is the Natural History of Dysplasia?
Correct Interpretation of Dysplasia
To understand the natural history of dysplasia, the nuances involved in the correct interpretation of dysplasia must first be appreciated. In both CD and UC, carcinoma develops through an inflammation/dysplasia/carcinoma sequence.62, 63 At present, dysplasia, which is defined as unequivocal neoplastic epithelium confined to the basement membrane, is the best and most reliable marker of an increased risk of malignancy in patients with IBD.64, 65 Dysplasia is present in more than 90% of UC cases with carcinoma, occurs in any portion of the colon, but often parallels the location of cancer, and may occur as an isolated focus but is more often multiple or, rarely, diffuse.66, 67, 68
In CD, there is also an increased risk of dysplasia and adenocarcinoma in excluded segments of bowel and in the small intestine.69 In general, the pathologic features of dysplasia in CD are similar to those in UC.47, 63, 68 Reported frequencies of dysplasia in patients with CD and CRC range from 40% to 100%.52, 63, 70 Similar to UC, dysplasia in CD occurs more often in areas adjacent to, rather that distant from, the primary tumor mass. However, studies are not uniform in the frequency with which they report dysplasia distant from adenocarcinoma, which ranges from 27% to 100%.63 Nevertheless, these numbers mirror, for the most part, the data in UC.
Cancer may develop in patients without any evidence of dysplasia or in patients with only LGD.66, 71, 72, 73, 74 In other instances, carcinoma may develop within mucosa that contains extremely well-differentiated, non–dysplastic-appearing epithelium without the traditional morphologic features of dysplasia, as described further below.75, 76 Regardless of the location of the primary tumor, IBD-related neoplasia always develops in areas of chronic or chronic active inflammation.65, 75, 77
Classification of Dysplasia
There are 2 general gross patterns of dysplasia in IBD, commonly referred to as flat and elevated.60, 78, 79 For the purpose of this evidence-based review, flat dysplasia refers to endoscopically undetectable lesions, whereas raised dysplasia refers to endoscopically detectable lesions. It should be noted, however, that sometimes the term “flat dysplasia” has been used to describe endoscopically detectable, but only slightly raised, lesions. Raised lesions are often referred to by the acronym “DALM.”59 There is inconsistency in the published literature with regard to the criteria used to designate a particular dysplastic lesion as a DALM.59, 64, 74, 76, 79, 80, 81, 82, 83 Furthermore, although it is presumed that most, if not all, raised dysplastic lesions begin as flat endoscopically undetectable areas of dysplasia, this has never been investigated thoroughly. In fact, rarely, invasive cancers in IBD may appear as endoscopically undetectable lesions.
Microscopically, dysplastic changes in UC are separated into 3 morphologic categories: negative (regenerating epithelium), indefinite, and positive for dysplasia (low and high grade).75 This microscopic classification system has formed the basis for most retrospective and prospective studies that have evaluated the biologic characteristics, natural history, and treatment of IBD-related precursor lesions. This is a classification system based primarily on observations made by dedicated gastrointestinal pathologists with extensive experience related to IBD.
In 1998, an alternative classification system for dysplasia in the gastrointestinal tract was developed with the hope that it would standardize, and help resolve, differences in interpretation and terminology between Western and Japanese pathologists.84 This new classification system was developed in Vienna (termed the “Vienna classification”) and uses 5 diagnostic categories as outlined in Table 2. It is used by pathologists in many European countries, and in Japan, but has yet to be commonly accepted in the United States. In brief, the Vienna classification system84 is similar to the one proposed by Riddell et al in 198375 except that the former uses the term “noninvasive neoplasia” instead of “low- or high-grade dysplasia” and also uses the term “suspicious for invasive carcinoma” for neoplastic lesions that have some cytologic and/or architectural features of invasion but in which the invasive component cannot be demonstrated unequivocally by histologic evaluation. Differences among Western and Japanese pathologists relate to interpretation of nuclear features in neoplastic lesions.85 However, there are no clinically important differences between these 2 classification systems.
Table 2. Comparison of Vienna and Riddell's Classification of Dysplasia in IBD
| Vienna | Riddell |
|---|---|
| 1. Negative for neoplasia/dysplasia | Negative for dysplasia |
| 2. Indefinite for neoplasia/dysplasia | Indefinite for dysplasia |
| 3. Noninvasive low-grade neoplasia (low-grade adenoma/dysplasia) | LGD |
| 4. Noninvasive high-grade neoplasia | HGD |
| 4.1 High-grade adenoma/dysplasia | |
| 4.2 Noninvasive carcinoma (carcinoma in situ) | |
| 4.3 Suspicious of invasive carcinoma | |
| 5. Invasive neoplasia | Adenocarcinomaa |
| 5.1 Intramucosal adenocarcinoma | Intramucosal |
| 5.2 Submucosal carcinoma or beyond | Invasive |
aNot included in the system by Riddell et al. |
Interobserver Variability in the Morphologic Diagnosis of Dysplasia
As noted previously, cancer in IBD develops on a progressive (continuous) scale rather than in discrete intervals. As a result, there is a significant degree of variability in interpretation of the grade of dysplasia even among experienced gastrointestinal pathologists.75, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97 Particularly difficult diagnostic areas include differentiation of marked regenerative changes from true LGD, and distinction between HGD and early invasive carcinoma. Since the study by Riddell et al in 1983, several interobserver studies have shown only moderate levels of agreement in the interpretation of dysplasia in IBD.87, 88, 89, 90, 97 In general, levels of agreement are highest for the category of HGD and for biopsy specimens considered negative for dysplasia but are lowest for biopsy specimens in the indefinite and LGD categories. For instance, in an interobserver variability study by Dixon et al that included 100 cases, 4 diagnostic categories, and 6 participating pathologists, the overall level of agreement among all pairs of observers was considered fair (κ = 0.41).90 Similarly, in a study by Melville et al that included 5 experienced pathologists, all of whom analyzed 207 specimens from 85 patients with UC for no dysplasia, indefinite for dysplasia, LGD, HGD, or carcinoma, the κ values were generally poor, varying between 0.21 and 0.45 for each pair of observers.89 Therefore, more precise criteria and specific guidelines are needed to increase the level of consistency and agreement among pathologists in interpretation of dysplasia in IBD. The limitations of morphology-based dysplasia have led to interest in finding other more objective markers of an increased risk of CRC in patients with IBD.
Adjunctive Markers Useful in the Histologic Differential Diagnosis of Dysplasia
Due to the high level of interobserver variability in the diagnosis of dysplasia in IBD, investigators have sought to identify molecular and nonmolecular markers that may aid in the distinction of truly dysplastic from regenerating epithelium in IBD.91, 92, 93, 94, 95, 96, 98, 99, 100, 101, 102, 103, 104 Some of the markers that have been investigated include proliferating cell nuclear antigen or Ki-67 (markers of cell proliferation), p53, sucrase isomaltase, glut-1, α-methylacyl-CoA racemase (AMACR), and sialyl-Tn antigen (STn). All of these studies used the morphology-based dysplasia grading system as the primary method (gold standard) of comparing antibody expression patterns. Of the nonmolecular markers, several studies have shown that true dysplasia expresses markers of cell proliferation, such as Ki-67, at higher levels of the crypt and in the surface epithelium compared with mucosa that only show epithelial regenerative changes.91, 92, 93, 98 However, due to significant overlap, pathologists do not generally use this antibody for this differential diagnosis. P53 has been evaluated in several studies and is discussed further in the section on molecular markers. STn antigen is a mucin-associated antigen detected by immunohistochemistry and expressed in up to 90% of CRCs but only rarely in normal colonic mucosa. In retrospective studies in UC, STn expression was shown to correlate with the dysplasia-carcinoma sequence, was expressed in affected colitic mucosa as well, and occurred earlier than dysplasia. However, the sensitivity and specificity of this biomarker are low.
One promising pathology biomarker is AMACR. In a recently published retrospective study, the presence of AMACR positivity, a peptide shown to be expressed in a high proportion of prostatic intraepithelial neoplasia and prostatic carcinomas, was not expressed in any UC foci considered negative for dysplasia but was shown to be significantly increased in foci of LGD (96%), HGD (80%), and adenocarcinoma (71%).102 Only 14% of indefinite for dysplasia foci were focally and weakly positive for AMACR and, in fact, one of these patients revealed dysplasia upon follow-up. Thus, AMACR is a new technique to help diagnose dysplasia in IBD in diagnostically difficult cases. It has a high degree of specificity (nearly 100%) and a slightly lower degree of sensitivity that ranges from 96% in LGD to 80% in HGD in the study by Dorer and Odze.102 Given that HGD is not usually difficult to differentiate from regeneration, a combination of histologic evaluation and AMACR immunostaining (particularly for indefinite and low-grade lesions) is useful for biopsy assessment of dysplasia in patients with IBD.
Unconventional (Nonadenomatous) Patterns of Dysplasia
Rarely, dysplasia in IBD may show unusual cytologic or architectural changes that, on occasion, may even be difficult to recognize as neoplastic.76, 86, 105 For instance, dysplasia may demonstrate a villous growth pattern, being composed of elongated and distended mucinous epithelium without significant cytologic atypia.76 This type of neoplastic alteration is termed “villous dysplasia” or “villous hypermucinous mucosa”76, 86 (Figure 1A and B). In a retrospective cohort study of villous dysplasia by Rubio et al, villous changes were detected in 70% of UC cases with carcinoma in contrast to 0% in biopsy specimens from 20 patients without carcinoma.76 In another study by Andersen et al, villous hypermucinous mucosa in long-standing UC was associated with a high prevalence rate of k-ras mutations (61%) in comparison with conventional “adenomatous” LGD (4%).86 In one other recent study, some patients with IBD developed a very well-differentiated “tubuloglandular” carcinoma in association with a very low-grade form of precursor LGD.106
Figure 1.
(A) High-power photomicrograph of a biopsy specimen from a patient with UC showing serrated LGD. The contour of the luminal epithelium shows infolding and slight serration, but the nuclei are predominantly basally located with little evidence of stratification and only a mild degree of nuclear atypia. (B) High-power photomicrograph of the surface mucosa from a patient with UC and an underlying moderately differentiated adenocarcinoma (not shown). The surface mucosa shows hypermucinous villiform epithelium without significant cytologic atypia of the cell nuclei.
Not uncommonly, dysplastic epithelium in IBD may appear serrated, thereby mimicking some of the features of conventional sporadic hyperplastic polyps or serrated adenomas.75, 105, 107 In one recent case series by Srivastava et al, 3 patients with IBD developed multiple hyperplastic and serrated polyps, some with a dysplastic appearance, reminiscent of hyperplastic/serrated polyposis. Two of the patients had (or developed) cancer.107 Other cases have been described that show a mixture of villous, serrated, and traditional adenomatous dysplasia in the same patient. In one retrospective case-control study by Rubio et al, the histologic phenotype of dysplasia juxtaposing CRCs showed a villous pattern in 52.3%, a serrated pattern in 28.9%, a tubular pattern in 5.3%, and a “mixed” pattern in 13.2%.76 It remains unclear whether serrated and/or villous dysplasia in IBD evolves through a genetic and biologic pathway distinct from conventional adenomatous dysplasia and, thus, is an area in need of further investigation.
Natural History of Dysplasia
Understanding the natural history of dysplasia is important for determining the outcome and success of surveillance. It has long been assumed that colon carcinogenesis in IBD follows a sequential progression from inflammation to dysplasia (low grade then high grade) and, finally, invasive cancer. Although this model is conceptually useful and serves as a reasonable paradigm, it is by no means absolute. There are instances in which patients who have undergone regular colonoscopic surveillance developed CRC without any prior history of dysplasia.73 Also, at the time of colectomy for CRC, concurrent dysplasia may be absent in up to 25% of cases.108, 109, 110 Similarly, not all cases of LGD progress through a stage of detectable HGD before developing cancer.66, 71, 111 This disconcerting fact may be explained, at least in part, by the recent recognition of a type of colitis-associated CRC, termed low-grade tubuloglandular adenocarcinoma. This type of invasive cancer seems to arise directly from LGD and is characterized by the presence of deceptively benign-appearing glands that lack a desmoplastic stromal reaction in the invasive component. This tumor may account for up to 10% of all CRC cases in IBD.106 These examples highlight the limitations in predicting the natural history of dysplasia in UC. Nevertheless, until newer, more reliable markers of cancer risk are identified, clinical management relies mainly on the endoscopic and histologic identification of dysplasia in mucosal biopsy specimens of the colon.
There are several general factors that influence our understanding of the “natural history” of dysplasia (Table 3). These include (1) correct histologic interpretation of dysplasia, (2) prevalent versus incident dysplasia, (3) recurrent dysplasia, and (4) focality of dysplasia.
Table 3. Factors Affecting the Natural History of Dysplasia
| Correct histologic interpretation of dysplasia |
| Prevalent vs incident dysplasia |
| Recurrent dysplasia |
| Focality of dysplasia |
Needless to say, correct pathologic interpretation of dysplasia influences the subsequent “natural history” of neoplasia in any given patient. The importance of accurate histologic interpretation with respect to subsequent clinical outcome is highlighted by the long-term prospective surveillance study from St Mark's Hospital, which began enrolling patients in 1971.108 In that study, all available biopsy specimens showing dysplasia (n = 249) and a sample of biopsy specimens showing no dysplasia (n = 52) were blindly reviewed using the 1983 IBD Morphology Study Group criteria, and outcomes of patients were analyzed according to both the original and the revised histologic diagnosis. Among patients in whom LGD was diagnosed in flat mucosa during any examination using the original diagnostic scheme, the cumulative 5-year rate of progression to HGD or CRC was 16%. By excluding cases that were deemed nondysplastic, the revised cumulative 5-year rate of progression to HGD or CRC increased to 54%. Importantly, of the 73 biopsy specimens originally interpreted as LGD before 1983, only 5 were considered to be definitely dysplastic on later review; of the 34 biopsy specimens interpreted as LGD, only 17 (50%) were considered truly dysplastic upon re-review.
Indefinite dysplasia is a histologic diagnosis with its own natural history. When indefinite for dysplasia is detected on initial colonoscopy, progression to advanced neoplasia occurs in approximately 13% to 18%.6, 64 When indefinite for dysplasia was found at some time during surveillance, the progression rate was 28%.64 In a study from Mount Sinai Hospital in New York, the 5-year rate of progression to either HGD or CRC was 1.1%, 9%, and 45% in patients who initially had no dysplasia, indefinite for dysplasia, and LGD, respectively.112
Overall, determination of the grade of dysplasia is performed on the basis of a combination of microscopic features, including architectural alterations of the crypts and cytologic abnormalities of the dysplastic cells. Reactive (regenerating) epithelium, located in areas of inactive or active colitis, is categorized as negative for dysplasia. Regenerative features are often prominent in the presence of fresh erosions or ulcerations. Unfortunately, regenerating epithelium may acquire some cytologic features that mimic dysplasia and may, on occasion, cause difficulty in pathologic differentiation from true dysplasia. In this circumstance, the category of indefinite for dysplasia is used, which implies that definite distinction between markedly reactive epithelium and truly dysplastic epithelium cannot be established with complete certainty on the basis of the tissue available for analysis. For instance, regenerating epithelial cell nuclei are often enlarged and more variable in size, show nuclear stratification, vesicular nuclei, prominent nucleoli, increased hyperchromatism, and more abundant mitoses, features that overlap with true LGD. Although surface maturation, which corresponds to the progressive acquisition of cytoplasmic mucin in cells located close to, and on, the luminal surface, is a feature commonly assumed to represent evidence in favor of regeneration, true dysplasia may also, rarely, show evidence of surface maturation as well.76
Cases are best considered indefinite for dysplasia when the cytologic and architectural features approach that of LGD but, due to the presence of active inflammation or ulceration in the area, the observer is uncertain as to the neoplastic status of the tissue. Other potential reasons why pathologists may use the category of indefinite for dysplasia include instances in which epithelium shows unusual growth patterns or has processing or sectioning artifact that makes interpretation difficult. In the original standardized classification for dysplasia proposed in 1983, the indefinite for dysplasia category was further subdivided into “probably negative” or “probably positive.”75 This subdivision has not been used among expert gastrointestinal pathologists in recent years.
True dysplasia is categorized as low or high grade based on the degree of cytologic and architectural atypia of the crypts and surface epithelium (Figure 2A–C). By consensus, the category of HGD includes carcinoma in situ, a term not recommended for use in gastrointestinal biopsy pathology.68 Microscopically, the most common type of dysplasia in IBD has morphologic characteristics similar to non–IBD-associated sporadic adenomas. The crypts may be tubular or villous in contour, show excess budding, and show an increased number per unit area, rendering a crowded appearance to the mucosa. The cytologic features of LGD include nuclear enlargement, increased nuclear to cytoplasmic ratio, hyperchromasia, pleomorphism, and increased mitoses, both typical and atypical. In LGD, the nuclei are mainly limited to the basal half of the cell cytoplasm. Some cases of dysplasia appear villous in configuration and show preservation, or even an increased amount, of cytoplasmic mucin.76, 86, 105 HGD is distinguished from LGD primarily based on the degree of architectural and cytologic aberration, but there is no well-defined cutoff point separating these 2 types of lesions. In general, HGD demonstrates all of the features of LGD but to a more severe degree. HGD is characterized by prominent nuclear stratification and nuclei that are typically larger in size and may contain a more “open” nuclear chromatin pattern compared with LGD. Mitoses, both typical and atypical, are more frequent and are usually present in the upper levels of the crypts and even in the surface epithelium. Other characteristic changes of HGD include marked hyperchromatism, crowding, pleomorphism and loss of polarity of the nuclei, and architectural aberrations such as a back-to-back gland pattern and cribriforming of the crypts. In general, the features of the most atypical portion of the mucosa determine the overall grade of dysplasia in any particular biopsy sample. However, the precise number of high-grade dysplastic crypts that are necessary to upgrade a particular biopsy specimen from low to high grade has never been investigated and varies among expert gastrointestinal pathologists. In contrast to dysplasia in which neoplastic cells are confined to the basement membrane, carcinoma is defined by the presence of cells or glands that have penetrated into the lamina propria and/or submucosa. Single cell infiltration, small gland infiltration, or large dysplastic crypts with irregular jagged contours or a complex cribriform gland pattern are different patterns of invasion commonly seen in IBD. Necrosis, hemorrhage, ulceration, and desmoplasia are common changes seen in carcinoma as well.
Figure 2.
(A) High-power photomicrograph of LGD in UC. The mucosa is slightly villiform in contour and contains dysplastic columnar cells with enlarged pencil-shaped nuclei, slight nuclear stratification, increased mitoses, dystrophic goblet cells, and slight loss of nuclear polarity. (B) In contrast to panel A, this biopsy specimen shows an increased degree of architectural and cytologic atypia consistent with HGD. The glands show a back-to-back configuration and are lined by highly dysplastic cells with full-thickness nuclear stratification, marked loss of nuclear polarity, increased numbers of mitoses, some atypical, and dystrophic goblet cells. (C) In this biopsy specimen from a patient with UC, the degree of architectural and cytologic atypia is consistent with intramucosal adenocarcinoma. There is a complex arrangement of back-to-back glands, with cribriforming of the luminal epithelium, and an infiltrative appearance of the glands in the lamina propria. The cells show a high N/C ratio and marked loss of polarity.
There may be an important difference in natural history depending on whether dysplasia is prevalent or incident. Patients who demonstrate prevalent dysplasia (that which is found on the initial, screening colonoscopy) have shown a higher rate of progression to CRC compared to patients whose dysplasia is detected during the course of surveillance colonoscopy (incident dysplasia). A review of 10 prospective studies reported that when LGD was found at initial surveillance colonoscopy (prevalent LGD), HGD or CRC developed in 16 of 55 patients (29%) at some time during further follow-up, with CRC developing in 7 (13%).64 However, if LGD was found during surveillance (incident LGD), only 33 of 204 patients (16%) progressed to either HGD or CRC, with 17 (8%) progressing to CRC. For comparison, when no dysplasia is found on initial colonoscopy, the rate of subsequent progression to CRC ranges from 1.1% to 3.1%.64, 108, 111, 113
Recurrent dysplasiaMany clinicians are reluctant to recommend colectomy for LGD unless it is detected on at least one other subsequent surveillance colonoscopy. This raises the question as to whether colitis-associated dysplasia can spontaneously regress. Certainly, the finding of recurrent LGD raises concern that the colon is at increased risk for HGD/CRC, and although some studies have used this as an indication for colectomy,108 others have not.114 Given the limitations of detecting LGD, it is conceivable that even with careful surveillance, recurrent LGD may be missed. Thus, after detection of LGD on one examination, does not finding LGD on subsequent examinations offer any reassurance that the colon is truly at a lower risk for neoplasia? Few studies have addressed this issue. Woolrich et al retrospectively reviewed their series of patients who underwent surveillance colonoscopy for UC.73 Among 22 patients who had LGD on initial colonoscopy after 7 years of UC, 7 (31%) continued to have dysplasia (4 LGD, 2 HGD, 1 CRC) on their second surveillance colonoscopy within 2 years, but 15 (69%) had no dysplasia on their second colonoscopy. Of the latter group, 6 (40%) still went on to develop subsequent dysplasia (3 recurrent LGD, 3 CRC at 5–10 years without intervening HGD). Ullman also reported several patients with flat LGD who had a second surveillance examination negative for dysplasia, yet HGD or CRC subsequently developed upon follow-up.71 In contrast, a Swedish prospective surveillance study, in which progression rates from LGD to HGD or cancer were extremely low, reported that 16 of 60 patients (27%) with LGD had no further LGD despite a mean of 6 follow-up colonoscopies.114
Focality of dysplasiaIt is sometimes casually assumed that a colon manifesting only one area of dysplasia (unifocal dysplasia) would be at a lower risk for developing CRC than if more than one dysplastic focus (multifocal dysplasia) is detected. Unifocal dysplasia is more common than multifocal dysplasia.60, 115 Few studies have reported the outcome of patients based on the focality of dysplasia. In general, if multifocal dysplasia is detected, even on one colonoscopic examination, (and especially if any biopsy site manifests HGD), colectomy is often recommended. However, the natural history of multifocal LGD is not known. A recent study specifically addressed focality of LGD, and found that the overall 53% 5-year progression rate of flat LGD to either HGD or CRC was nearly identical among the 39 patients with unifocal LGD and the 7 patients with multifocal LGD.116 Until more data are acquired on this issue, unifocal flat LGD should not be discounted.
Should Colectomy Be Performed for Raised Dysplasia?
Gross Features of Raised Dysplasia
There is lack of consistency in the literature with regard to the criteria and methods used to designate raised, endoscopically visible, dysplastic lesions as DALMs.60, 78, 79 For instance, some studies categorize dysplastic lesions as flat only if they are endoscopically undetectable, whereas others include visible plaque-like or velvety discolored areas of mucosa in this category as well. There is also discordance with regard to the definitions used (including the criteria for dysplasia), the gross features and terms used to describe DALMs, and the clinical and endoscopic features of the lesions. Some studies are based on endoscopic biopsy specimens, whereas others use resection specimens to define the gross characteristics of dysplastic lesions. Finally, few studies actually include representative photographs of DALMs. As a result, there is variability in the reported frequencies of cancer associated with DALMs between different studies, ranging from 38% to 84%.64, 79
Regardless, recent studies suggest that raised (endoscopically visible) dysplastic lesions in IBD may be broadly separated into those that appear similar to non–IBD-related sporadic adenomas, herein referred to as “adenoma-like,” and those that do not resemble adenomas, which are referred to as “non–adenoma-like.”117, 118, 119 Biopsy specimens of non–adenoma-like DALMs may represent the surface of an invasive adenocarcinoma.59 Based on the widely accepted premise that carcinogenesis in UC and CD occurs exclusively in areas of mucosa involved with the inflammatory process, it is assumed, and accepted, by most investigators that dysplastic lesions that occur proximal to (in UC) or in between areas of involved colon (in CD) may justifiably be considered sporadic in origin and thus unrelated to the underlying IBD.81 This classification system is useful, both pathologically and clinically, because it helps separate DALMs into 2 groups with a different natural history, risk of malignancy, and treatment (see further in the following text).117, 120, 121
Grossly adenoma-like DALMs represent well-circumscribed, smooth or papillary, nonnecrotic, sessile, or pedunculated polyps that, similar to sporadic adenomas (Figure 3A and B), are usually readily amenable to removal by routine endoscopic methods (Table 4).79, 117 Other synonyms used to describe this lesion include adenoma-like polyp, adenoma-like dysplastic polyp, polypoid dysplasia, and adenoma-like mass. Non–adenoma-like lesions (Figure 4A and B) include velvety patches, plaques, irregular bumps and nodules, wart-like thickenings, stricturing lesions, and broad-based masses.59, 60, 74, 80, 82 Non–adenoma-like DALMs are not usually amenable to removal by colonoscopic polypectomy. Unfortunately, there are little data on the incidence of cancer as it relates to the different gross subtypes of non–adenoma-like DALMs, and this is an area in need of further research.
Figure 3.
(A and B) Endoscopic images of adenoma-like DALMs. Images provided courtesy of Jerome D. Waye, MD.
Table 4. Gross Features of DALMs
| Adenoma-like (endoscopically resectable) | Non–adenoma-likea (endoscopically nonresectable) |
|---|---|
| Sessile/pedunculated | Usuallysessile(broadbased) |
| Well circumscribed | Poorlycircumscribed |
| Smooth surface | Irregularsurface |
| Visible borders | Indistinctborder |
| Nonulcerated | Ulceration/necrosis |
| No stricture | Stricture |
| No mucosal tethering | Tethering |
aIncludes velvety patches, plaques, nodules, bumps, and wart-like thickening. |
Figure 4.
(A and B) Endoscopic images of non–adenoma-like DALMs. Image 4B provided courtesy of Jerome D. Waye, MD.
The published data are conflicting with regard to the prevalence of flat and raised dysplasia (DALMs) in patients with UC.59, 60, 74, 80, 82, 83, 120, 122 One recent retrospective endoscopic correlative study by Rutter et al evaluated 56 patients with UC who developed dysplasia (either flat or raised) in the course of a 14-year surveillance program conducted at St Mark's Hospital in London.60 A total of 110 neoplastic areas were detected in 56 patients, of which 77.3% were visible at colonoscopy. More specifically, 74 of the visible lesions (87%) were “polypoid” (adenoma-like), 4 were described as having an “irregular” outline, and one was described as a “plaque.” In addition, 6 were described as macroscopic cancers although without further gross description. Thus, the results of this study suggested that most dysplastic lesions in UC are endoscopically visible and that the majority of visible lesions are, in fact, well-circumscribed adenoma-like polyps rather than irregular adenoma-like lesions.
However, in a study by Toruner et al of 635 patients with IBD who underwent surveillance colonoscopy between January 2002 and November 2003, 24 patients (3.8%) had flat dysplasia, 12 (1.9%) had IBD-related polypoid dysplasia, and 28 (4.4%) had sporadic tubular adenomas.121 In this study, 40% of dysplasia was flat, not raised.
There are also conflicting data regarding the percentage of raised dysplastic lesions (DALMs) that are adenoma-like versus non–adenoma-like.60, 64, 74, 80, 82, 83 In fact, the data in most studies are difficult to interpret because lesions are usually referred to as “polypoid” without further gross description. For instance, in one study by Blackstone et al and one retrospective study by Butt et al, the prevalence rate of “polypoid” dysplasia in their UC patient cohorts was only 42% and 28%, respectively, of raised macroscopically detectable dysplastic lesions.59, 82 Although the prevalence of sporadic adenomas increases with age in the general non-IBD population, patient age, by itself, is not a reliable enough criterion in order to judge whether a dysplastic lesion is sporadic or colitis associated.
Microscopic Features of Raised Dysplasia (DALMs)
Non–adenoma-like and adenoma-like DALMs are differentiated on the basis of their gross (endoscopic) features because histologically, and particularly on biopsy analysis, both types of lesions may appear identical.81 Both types of DALMs are typically composed of a tubular, tubulovillous, or villous proliferation of “adenomatous” epithelium containing dysplastic columnar cells (Figure 5A–C). Dysplastic columnar cells show pencil-shaped, basally located nuclei, stratification, clumped chromatin, multiple small nucleoli, mucin depletion, increased mitosis (both typical and atypical), nuclear hyperchromaticity, and loss of cell polarity.81, 123, 124 A variable degree of acute and chronic inflammation, and ulceration may be present as well. Biopsy specimens from non–adenoma-like DALMs may only show fragments of dysplastic epithelium, which, in many instances, simply represents the surface of an underlying carcinoma. In this circumstance, biopsy specimens from deeper portions of the lesion may reveal the characteristic features of invasive adenocarcinoma.
Figure 5.
(A) Medium-power view of an adenoma-like DALM occurring in a patient with UC. This lesion was located within an area of established chronic colitis. However, it was smooth and well-circumscribed grossly. Microscopically it shows the typical features of a sporadic tubular adenoma. There was no evidence of flat dysplasia elsewhere in the colon from this patient. This patient was treated with polypectomy and continued surveillance and has been negative for dysplasia after 8 years of follow-up. (B) High-power photomicrograph of a well-circumscribed adenoma-like polyp within an area of chronic UC. This lesion shows a mixture of low-grade dysplastic glands and nondysplastic glands at the surface of the polyp. There is also an increased degree of inflammation in the lamina propria. This lesion was interpreted as an adenoma-like polypoid area of dysplasia in UC because of the histologic features of the lesions and the association with flat LGD in areas surrounding the polypoid lesion. (C) Low-power view of a tissue section from a non–adenoma-like DALM in UC. The section shows elongated finger-like “papillary” projections of predominantly low-grade, dysplastic columnar epithelium. This lesion represented the surface of a non–adenoma-like DALM, which was sessile, broad based, and showed an irregular surface contour. After pathologic evaluation of the resection specimen, a well-differentiated adenocarcinoma with infiltration into the submucosa was evident.
The microscopic features of adenoma-like DALMs are similar to those of sporadic adenomas unrelated to IBD.81 There have been several retrospective case-control studies designed specifically to evaluate features that may help differentiate UC-related adenoma-like DALMs from sporadic adenomas by microscopic, immunohistochemical, or molecular methods.81, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132 For instance, in one retrospective case-control study by Torres et al, 89 adenoma-like DALMs from 59 patients with UC were evaluated morphologically.81 These investigators utilized strict endoscopic criteria to select a homogeneous population of DALMs that resembled sporadic adenomas. Patients with UC-related adenoma-like DALMs had a statistically significant longer duration of disease (greater than 10 years), a higher proportion of polyps with tubulovillous, or villous, architecture, a mixture of normal and dysplastic epithelium at the surface of the polyps, and a higher degree of mononuclear inflammation in the lamina propria compared with non–UC-related sporadic adenomas. In addition, the median age of UC patients was 48 years compared with 63.5 years for non-UC patients with sporadic adenomas.
The presence of stalk dysplasia has been suggested, anecdotally, to represent a feature indicative of an IBD-related polypoid dysplastic lesion rather than a sporadic adenoma. However, in the study by Torres et al, both UC-related adenoma-like DALMs and sporadic adenomas had a similar prevalence rate of stalk dysplasia. However, patients with UC-related adenoma-like DALMs may show dysplasia in flat mucosa adjacent to, or distant from, the dysplastic polyp, which is rare in patients with UC and sporadic adenomas.81
Attempts have also been made to separate UC-related adenoma-like DALMs from sporadic adenomas by immunohistochemical or molecular methods.125, 126, 127, 128, 129, 130, 131, 132 However, the studies that evaluated these parameters were all retrospective in design and lacked consistency in the criteria used for selection of cases into each of the 2 diagnostic categories. Fortunately, distinction between adenoma-like DALMs and sporadic adenomas is no longer clinically relevant because both types of lesions are treated equally well with endoscopic resection (see the following section on treatment).
Management of Raised Dysplasia
Perhaps the most important recent advance in the management of dysplasia in IBD is the appreciation that polypoid dysplastic lesions can often be managed more conservatively than flat dysplasia. There is compelling recent data to suggest that patients with UC and an adenoma-like DALM may be treated adequately by polypectomy and continued surveillance regardless of the underlying pathogenesis (whether UC related or sporadic) of the lesion.117, 118, 122, 133 In a recent long-term prospective follow-up study of 34 patients with UC who had a polypectomy and continued surveillance for adenoma-like DALMs in UC, 20 of 34 patients (59%), overall, developed at least one further adenoma-like DALM on follow-up, but only one patient developed flat LGD and only one other developed adenocarcinoma after initial polypectomy. Most importantly, there were no significant differences in the incidence of polyp formation on follow-up between patients with UC and an adenoma-like DALM (62.5%) and patients with UC and a known sporadic adenoma (50%), or between either of these 2 UC patient groups and a non-UC sporadic adenoma control group (49%). This study represents the long-term outcome results of a previously reported study by the same research group that included, on average, about 3 years of follow-up.119 Another prospective cohort study by Rubin et al showed results that were similar to those of the study by Odze et al.118 In their study, the outcome of 30 patients with UC, and 18 patients with CD, with 70 dysplastic polyps that resembled “adenomas” endoscopically and pathologically, all of which were treated by polypectomy and continued surveillance (with a mean follow-up of 4.1 years), was evaluated. Most patients (52%) did not develop any further polyps, and none developed either flat dysplasia or adenocarcinoma. Finally, in one recent follow-up study (mean; 6.0 years) of adenoma-like lesions in 148 patients with UC, of the 87 patients treated with polypectomy, only 4.6% developed dysplasia on follow-up (2 of which were ultimately diagnosed with carcinoma), which supports polypectomy as a valid course of treatment in these patients.133
As a result of these studies, a treatment algorithm for patients with UC and either an adenoma-like or a non–adenoma-like DALM59, 68, 74, 82, 83 has been proposed and is outlined in Figure 6. There is a high association of cancer with non–adenoma-like DALMs that are considered to be endoscopically unresectable, ranging from 38% to 83%. For this reason, it is recommended that patients with UC and an endoscopically unresectable, non–adenoma-like DALM, regardless of the grade of dysplasia detected on biopsy analysis, should undergo a colectomy because of the high association with metachronous, or synchronous, carcinoma. (US Preventive Services Task Force [USPSTF] Grade A: High certainty that the magnitude of net benefits is substantial. Patients with IBD with a non–adenoma-like DALM should be treated with colectomy.)
Figure 6.
Proposed management scheme for DALMs in IBD.79
In contrast, adenoma-like DALMs located outside, or proximal, to areas of known colitis may be assumed to be sporadic in origin and thus treated conservatively by polypectomy and continued surveillance.134 Similarly, adenoma-like DALMs located within areas of known colitis may also be treated conservatively by polypectomy and continued surveillance if the lesion has been excised completely, shows an absence of dysplasia at the margins of the specimen, and there is no evidence of flat dysplasia elsewhere in the colon, either adjacent to, or distant from, the polypoid lesion.117, 118, 134 These recommendations apply to patients with UC regardless of their age, duration, or extent of colitis.79, 117, 118, 134 (USPSTF Grade A: High certainty that the magnitude of net benefits is substantial. Patients with IBD with an adenoma-like DALM and no evidence of flat dysplasia elsewhere in the colon can be managed safely by polypectomy and continued surveillance.) Although a scar may be detectable after colonoscopic resection, it is advisable to tattoo the mucosa adjacent to any suspicious lesion that is resected endoscopically for ease of future identification. Some referral centers with significant experience in the field of colitis-associated dysplasia have published preliminary reports on the use of endoscopic mucosal resection for non–adenoma-like dysplastic lesions.135
Should Colectomy Be Performed for Flat Dysplasia?
As mentioned in the previous section, if a raised dysplastic lesion can be resected endoscopically, regardless of whether the dysplasia is low or high grade, the patient can be managed with continued surveillance colonoscopy and does not necessarily require a colectomy. This management is therefore analogous to that of a sporadic adenoma in the general population. However, if flat dysplasia is encountered in colitic mucosa, especially if the lesion cannot be completely removed by endoscopy, serious consideration should be given to colectomy. This is particularly true for HGD. Studies indicate that at the time flat HGD dysplasia is discovered, CRC may already be present in 42% to 67% of cases. These results are based on 10 of 24 patients (42%) with HGD identified in a review of 10 prospective surveillance trials,64 8 of 12 patients (67%) with HGD managed in a long-term prospective trial from St Mark's Hospital in London,108 and 4 of 6 patients (67%) with HGD followed up in a 23-year surveillance program in Tokyo, Japan.136 An update of the St Mark's experience demonstrated that 5 of 11 patients (45%) who had an immediate colectomy for HGD revealed CRC.6 Although the absolute number of patients in these studies is small, the aggregate experience has been of enough concern to warrant colectomy in most instances. (USPSTF Grade A: There is high certainty that colectomy for flat HGD treats undiagnosed synchronous cancer and prevents metachronous cancer.) If colectomy is not performed when HGD is first detected, what is the likelihood that CRC will develop on further follow-up? Since most patients with HGD have already undergone colectomy, data on the risk of developing CRC in nonoperated patients with HGD is rather limited. The review of 10 prospective surveillance trials found that 15 of 47 patients (32%) with HGD developed CRC upon further follow-up.64 The updated St Mark's surveillance data indicated that 2 of 8 patients (25%) with HGD who did not have immediate colectomy, progressed to CRC.6 Although several of these studies were conducted before it was fully appreciated that polypoid dysplasia, can sometimes be managed conservatively with endoscopic resection alone, the immediate and subsequent risk of CRC is large enough to warrant a recommendation for colectomy.
It has been more difficult to achieve a consensus among expert clinicians on the optimal management of patients with flat LGD. In part, this is because not every study that reported the outcome of LGD distinguished between raised and flat LGD. Similar to flat HGD, there is a chance that synchronous CRC is present in the colon of a patient with flat LGD. Although the rate of synchronous CRC is lower for LGD than HGD, in general, it is still considerable. A study from Mount Sinai Hospital in New York specifically evaluated the fate of flat LGD and found that 3 of 11 patients (27%) who underwent colectomy within 6 months of diagnosis of flat LGD had an unexpected finding of HGD or CRC in their resection specimens.71 In other studies that did not distinguish flat from polypoid LGD, 3 of 16 patients (19%) with LGD who underwent immediate colectomy had synchronous CRC.64 Likewise, an update of the St Mark's data indicates that 2 of 10 patients (20%) with LGD (gross morphology not specified) had CRC on immediate colectomy, a rate of CRC that did not decline (7/36; 19%) if colectomy was performed after a period of follow-up.6
A recent meta-analysis of 20 surveillance studies analyzed the cancer risk of 477 patients with flat LGD and 31 patients with LGD in a DALM that study, the criteria for LGD was not specified.137 The average duration of UC was 12 years, with an average of 3.6 colonoscopies per patient and 18 biopsies per colonoscopy. The incidence of CRC was 14 per 1000 patient-years. For HGD and/or CRC, it was 30 per 1000 patient-years. The positive predictive value of flat LGD was 22% for synchronous CRC and 36% for synchronous HGD and CRC. The positive predictive value for progression to HGD and CRC was 14.6%. If LGD was detected in a DALM, the rates of synchronous and metachronous cancer were higher. Overall, when LGD is detected on surveillance, there is a 9-fold increased risk of developing CRC, and a 12-fold risk of developing HGD or CRC.137
Even if we assume that there is an approximate 22% rate of concurrent CRC, and a 36% rate of concurrent HGD and CRC when only flat LGD is found, many physicians and patients prefer to avoid immediate colectomy and opt for continued, albeit closer, surveillance. If immediate colectomy is not performed for LGD, what is the subsequent rate of progression? The meta-analysis by Thomas et al indicated that the positive predictive value is approximately 14.6% for progression of flat LGD to HGD and/or CRC.137 However, significant variability is present between studies. Studies from the St Mark's Hospital surveillance program, initially reported a 5-year progression rate of LGD to advanced neoplasia of 54%,108 although an updated 30-year follow-up analysis refined this rate to 23%.6 Likewise, 2 retrospective series from the Mayo Clinic116 and Mount Sinai Hospital67 carefully selected cases of flat LGD using specific histologic criteria and excluded raised LGD. They reported 5-year progression rates to advanced neoplasia of 33% and 53%, respectively. In the latter study, cancers detected at the time of immediate colectomy (within 6 months of the diagnosis of flat LGD) were at an earlier stage compared to cancers detected after a period of expectant management. A report, in abstract form, from the University of Washington, where an average of 44 biopsy specimens per colonoscopy were obtained and patients with LGD (morphology not designated) were followed for 3- to 12-month intervals, progression to HGD or CRC occurred in 9 of 40 patients (22%) over a mean of 5 years.138 In contrast, other prospective studies have reported lower rates of progression in patients with LGD. The Leeds group found only a 3% initial, and 10% subsequent, rate of progression to CRC during a 10-year follow-up period. Since these rates were not significantly higher than the 0.8% and 3% progression rates among patients without dysplasia, they concluded that LGD is not associated with a higher risk of CRC.139 A study from Huddinge, Sweden, observed a 35% rate of progression to HGD or CRC during 20 years of follow-up.111 Likewise, a prospective study from the Karolinska Institute in Sweden found no progression to CRC, and only 2 cases of progression to HGD, over a 10-year period.114 Investigators in Tokyo described progression to HGD or CRC in 22% of 9 patients with LGD over a 23-year surveillance period.136 Thus, progression of LGD to advanced neoplasia ranges from as low as 0% to 3% over 10 years, to 35% to 54% over 5 years. Furthermore, a summary of 8 studies revealed that after a diagnosis of LGD was made, subsequent surveillance, with an average of 4.3 colonoscopies per patient, detected more HGD lesions (n = 47) than CRCs (n = 18) or DALMs (n = 8).137 This may be important if one considers the goal of surveillance to be the prevention of mortality from CRC rather than the detection of HGD. Thus, the decision to undergo colectomy versus continued surveillance in patients with flat LGD should be individualized and discussed at length between the patient and the colorectal surgeon (See note in Acknowledgment section).140 A recent medical decision analysis of 2 simulated cohorts of 10,000 patients with long-standing UC and newly diagnosed unifocal flat LGD reported that immediate colectomy dominated over enhanced surveillance, and yielded higher quality-adjusted life-years and lower costs.141 (USPSTF Grade Insufficient: The current evidence is insufficient to assess the balance of benefits and harms of colectomy for flat LGD.)
Is There Sufficient Rationale for Performing Surveillance Colonoscopy in Patients With IBD?
Randomized controlled trials have not been performed to prove that surveillance colonoscopy is effective. However, a large number of case series have suggested a benefit of surveillance colonoscopy.13, 72, 83, 113, 142 Three case-control studies have examined this issue. In a population-based, nested case-control study of 142 patients with UC (derived from a study population of 4664 patients with UC) from Stockholm, Sweden, 2 of 40 patients with UC and CRC and 18 of 102 controls had undergone at least one surveillance colonoscopy (RR, 0.29; 95% CI, 0.06–1.31). Twelve controls, but only one patient with UC, had undergone 2 or more surveillance colonoscopies (RR, 0.22; 95% CI, 0.03–1.74). Although not statistically significant, the investigators of that study suggested that frequent colonoscopy protects against CRC.143 Lashner et al, in 1990, found that 4 of 91 patients who underwent surveillance died of CRC, compared to 2 of 95 patients who did not undergo surveillance (RR, 2.09; 95% CI, 0.39–11.12). Colectomy was less common in the surveillance group (33 vs 51; P < .05). It was performed, on average, 4 years later (after 10 years of disease) in the surveillance group.54 Finally, Choi et al examined 41 patients who developed CRC between 1974 and 1991.144 In this outcome study, 19 patients who underwent colonoscopic surveillance presented with a significantly earlier stage of cancer compared to 22 patients who did not participate in a colonoscopic surveillance program (P = .039). The 5-year survival rate was 77.2% in the surveillance group and 36.3% in the nonsurveillance group. In the Cochrane pooled data analysis of these 3 studies, 8 of 110 patients in the surveillance group died of CRC compared to 13 of 117 patients in the nonsurveillance group (RR, 0.81; 95% CI, 0.17–3.83).145 The Cochrane analysis concluded the following:
There is no clear evidence that surveillance colonoscopy prolongs survival in patients with extensive colitis. There is evidence that cancers tend to be detected at an earlier stage in patients who are undergoing surveillance, and these patients have a correspondingly better prognosis, but lead-time bias may contribute substantially to this apparent benefit. There is indirect evidence that surveillance is likely to be effective at reducing the risk of death from IBD-associated and indirect evidence that it may be acceptably cost-effective.145
There are little data on the effectiveness of surveillance in Crohn's colitis. One study that followed 259 patients with extensive Crohn's colitis found that 7% of patients had dysplasia or cancer upon screening colonoscopy, and an additional 14% developed dysplasia or cancer on surveillance examinations.146 Despite the lack of data from randomized controlled trials, surveillance colonoscopy is currently considered the standard of care. Many groups have previously provided recommendations for implementation of surveillance colonoscopy.68, 140, 145, 147, 148, 149 (USPSTF Grade B: There is moderate certainty that surveillance colonoscopy results in at least moderate reduction in CRC in patients with IBD. Despite the lack of randomized controlled trials, surveillance colonoscopy is recommended in patients with IBD at increased risk for developing CRC. Patients with extensive UC and CD of the colon are most likely to benefit from a surveillance program.)
Several factors may adversely affect the success of surveillance colonoscopy (Table 5). First, the onus of detecting dysplasia rests with the endoscopist. Unlike sporadic adenomas, which are typically single, well-visualized, and often readily resectable lesions, many factors make endoscopic detection and management of colitis-associated dysplastic lesions problematic (discussed previously). These issues should be conveyed to the patient prior to embarking on a program of surveillance, so that there is an understanding of the limitations of surveillance colonoscopy. While an older notion held that most dysplasia in colitic colons was invisible by endoscopy, it is becoming increasingly clear that, in fact, most dysplasia in IBD is visible. In a retrospective study, even after excluding polypoid cancers and apparent tubular adenomas, approximately two thirds of dysplastic lesions in patients with UC were visible.60 Nevertheless, because a sizable minority of dysplastic lesions in IBD colons is invisible with standard endoscopes, chromoendoscopy, and newer high-definition colonoscopes, are likely to enhance detection of dysplastic lesions.
Table 5. Factors That Influence the Success of Surveillance Colonoscopy
| Endoscopic recognition of dysplasia |
| Endoscopic resectability and completeness of resection |
| Adequacy of mucosal sampling |
| Interfering anatomic factors: strictures; pseudopolyps |
| Patient acceptance/compliance |
Second, the concept of “endoscopic resectability” is becoming increasingly important, and is now being factored into surveillance decision making, with the advent of more sensitive techniques such as chromoendoscopy (see the following text). As noted previously, in patients with an adenoma-like DALM and no dysplasia elsewhere in the colon, endoscopic polypectomy and continued colonoscopic surveillance are considered adequate therapy. This, of course, assumes that the entire polypoid lesion can be endoscopically resected. If at any point a dysplastic lesion is deemed endoscopically unresectable, surveillance should be discontinued in favor of colectomy.
Third, the success of surveillance also depends on the adequacy of mucosal sampling. Detection of dysplasia may be affected by the number of biopsy specimens obtained at colonoscopy. It has been estimated that 33 and 64 biopsy specimens are required to detect dysplasia with 90% and 95% probabilities, respectively.150 This has served as the basis for surveillance practice recommendations.140 Few studies have directly addressed the outcome of patients undergoing high-biopsy versus low-biopsy surveillance. Surveys of practicing gastroenterologists in the United States151 and the United Kingdom152 indicate that few practitioners take this many biopsy specimens. In fact, extensive biopsies of the colonic mucosa evaluates only a tiny fraction of the colon surface area. This has spawned the development of new imaging strategies such as chromoendoscopy and narrow band imaging designed to generate “targeted” rather than random biopsies.
Fourth, the frequency of surveillance may influence the success of a surveillance program. Although the precise interval to perform surveillance colonoscopies has not been rigorously studied, in some cases important pathology has been detected within 2 years of a negative surveillance colonoscopy. In the classic St Mark's UC surveillance study, in which patients underwent colonoscopy every 2 years (or sooner if dysplasia was found), 7 interval cancers developed between 10 and 28 months after a colonoscopy that was reportedly free of dysplasia or cancer.108 A study from Leeds, England, described one patient with Dukes' C cancer and 3 patients with HGD, that developed 1 to 2 years after a colonoscopy without dysplasia.139
Fifth, anatomic factors, such as strictures and inflammatory pseudopolyps may interfere with the ability to detect or sample colonic neoplasms. A case-control study of patients with UC from St Mark's Hospital demonstrated that inflammatory polyps and colonic strictures increased the risk of CRC by 2-fold and 4-fold, respectively, compared to individuals without these abnormalities.55 In fact, patients with macroscopically normal colons had a significantly lower risk of CRC (OR, 0.38; 95% CI, 0.19–0.73). Strictures in UC are of concern, not only because they make passage of colonoscopes difficult, but because approximately 24% may be malignant.58 Of particular concern are strictures located proximal to the splenic flexure in patients with long-duration UC, particularly if symptoms of bowel obstruction are present. In contrast, strictures of the colon are more common in patients with Crohn's colitis. Among a series of hospitalized patients with CD, 6.8% of patients with a stricture had malignancy, compared to 0.7% without strictures.153 In that series, compared to patients with benign strictures, malignant strictures tended to be shorter in length, and were associated with longer disease duration. Use of a small-caliber (pediatric) colonoscope increased the rate of accomplishing a complete colonoscopy by 16% to 33% in patients with Crohn's colitis, and in several instances, this technique was responsible for finding dysplasia or cancer.47
Finally, and most importantly, patient acceptance is crucial to the success of surveillance. If patients drop out of surveillance, or are delinquent in following up, this may have negative consequences.108, 111 In a case-control study by Eaden of 102 cases of CRC in UC, with matched controls, hospital physician visits more than twice a year was associated with a decreased risk of developing CRC (OR, 0.16; 95% CI, 0.04–0.60).13
How Should Surveillance Colonoscopy Be Performed?
Patients with ulcerative proctitis or ulcerative proctosigmoiditis are not at increased risk for IBD-related CRC and thus may be managed on the basis of average-risk recommendations. However, all patients should undergo a screening colonoscopy 8 years after the onset of symptoms with multiple biopsies throughout the colon to assess the true microscopic extent of disease. When performing surveillance colonoscopy, the most proximal extent of disease detected histologically at any point in time should define the patient's true extent of disease.
The most appropriate technique of surveillance colonoscopy has never been subjected to a randomized clinical trial. Nevertheless, a series of recommendations have been promulgated for the most effective performance of surveillance colonoscopy.68, 140, 147, 154 Similar to any colonoscopic surveillance examination, every effort should be made to maximize cleansing of the bowel to allow for a thorough examination of the colonic mucosa. Careful inspection of the mucosa along with a sufficient number of biopsy specimens should be obtained from all anatomic segments of the colon. In a study by Rubin et al, 33 biopsy specimens were needed to detect dysplasia with 90% probability if dysplasia is present.150 In that study, at least 64 biopsy specimens were needed to reach a 95% probability of detecting dysplasia. Four quadrant biopsy specimens should be obtained from appoximately every 10 cm of colon. Ideally, specimens of each anatomic segment should be submitted in a separate specimen jar, rather than pooling several segments, to avoid confusion regarding the location of a dysplastic area that might need to be monitored. Some experts recommend that more biopsy specimens be obtained from the rectosigmoid given that there is an increased risk of cancer in this area in patients with UC.73, 155, 156 Since the presence of active inflammation may make it difficult for some pathologists to distinguish reactive atypia from dysplasia, it is also recommended that surveillance colonoscopy be performed, ideally when the patient is in clinical remission.154
The Crohn's & Colitis Foundation of America consensus conference recommendations140 for colonoscopic surveillance in patients with UC are listed as follows:
The Crohn's & Colitis Foundation of America consensus conference recommendations for colonoscopic surveillance in patients with CD of the colon are listed as follows:
Recommendations from the British Society of Gastroenterology published in 2002 are similar but not identical to the Crohn's & Colitis Foundation of America consensus conference recommendations.154, 157 The British Society of Gastroenterology plans to update their guidelines in 2010 to incorporate duration and extent of disease, clinical risk factors (family history of CRC, PSC), and endoscopic and histologic findings to stratify patients into low- and high-risk categories (personal communication, Jayne Eaden, MD, November 2009). In the new recommendations, the interval between surveillance examinations is dependent on each individual's personal risk factors. In patients with a previous history of PSC, ongoing active inflammation, previous history of dysplasia or strictures, and strong family history of bowel cancer, annual surveillance is recommended. An intermediate cohort will include patients with postinflammatory polyps and a family history of bowel cancer later in life, for which surveillance examinations every 3 years will be recommended. Patients with quiescent disease and no other risk factors will be recommended for surveillance every 5 years.158 The British Society of Gastroenterology guidelines recommend chromoendoscopy and targeted biopsies as the preferred endoscopic method of surveillance. Alternatively, 2 to 4 random biopsies from every 10 cm of colon should be taken. The following summarizes the British Society of Gastroenterology guidelines:
The 2004 American College of Gastroenterology guidelines recommend that after 8 to 10 years of colitis, annual or biannual surveillance colonoscopy, with multiple biopsies at regular intervals, should be performed in patients with either left-sided colitis or pancolitis.147 Patients with proctitis and proctosigmoiditis are not considered at increased risk for cancer and thus are recommended to have average-risk surveillance.
The recommendations of the authors of the present AGA Institute Technical Review for surveillance colonoscopy in IBD are summarized as follows:
Management Based on Histologic Findings
Colectomy is recommended for patients with flat HGD confirmed by an expert gastrointestinal pathologist.92, 124, 137, 138 In contrast, in patients with biopsy specimens considered indefinite for dysplasia, guidelines suggest surveillance colonoscopy between 3 to 12 months.92, 124, 138 Multifocal LGD is a stronger indication for colectomy. Although controversial, there is evidence to suggest that patients with flat, unifocal, LGD should also be considered for colectomy.140, 154 The optimal surveillance interval for patients with flat LGD is unknown, but 3 to 6 months is usually recommended for initial surveillance examinations. Likewise, whether the finding of flat LGD on serial colonoscopies portends a higher risk of synchronous or metachronous CRC compared to the finding of flat LGD on a single examination has not been clearly defined. In this situation, it is important to educate the patient regarding the possibility of an interval cancer. Management of patients with polypoid dysplasia has been reviewed in detail elsewhere in this report. Figure 7 summarizes our recommendations for the management of flat and raised dysplasia.
Figure 7.
Recommendations for the management of flat and raised dysplasia. Modified and reprinted with permission from Elsevier.68
What Role Do the Newer Imaging Techniques Play in Identifying and Managing Dysplasia?
Newer techniques are needed to facilitate identification of neoplastic lesions in patients with IBD. Chromoendoscopy has been proposed as a method to increase the yield of detecting dysplasia on surveillance colonoscopy and is considered a technique easily applicable to clinical practice. Chromoendoscopy has been more commonly used in Europe and Asia than in the United States to identify nonpolypoid flat and depressed neoplastic lesions in the colon.159 Chromoendoscopy has 2 main advantages57: (1) it improves detection of subtle colonic lesions, which raises the sensitivity of the endoscopic examination, and (2) once a lesion is detected, its chromoendoscopic appearance can help improve lesion characterization, which increases the specificity of the endoscopic examination. The use of a magnifying colonoscope may further improve the sensitivity and specificity of chromoendoscopy. For instance, crypt architecture can be categorized by evaluating the pit pattern, which aides differentiation between neoplastic and nonneoplastic changes, and enables the performance of targeted biopsies.160 Neoplastic changes are characterized by an irregular, tubular, or villous crypt architecture staining pattern. Stellar or regular round pits are associated with nonneoplastic changes. Several dyes have been used in chromoendoscopy. These include absorptive and contrast stains. Indigo carmine neither reacts with nor is absorbed by colonic mucosa, but rather it pools in mucosal grooves. Alterations in the appearance signify a disruption of the normal colonic mucosa. Methylene blue is a vital stain that is rapidly absorbed by normal colonic mucosa but poorly absorbed by dysplastic or inflamed tissue. In addition, chromoendoscopy has also been shown to provide a more accurate evaluation of extent of disease and the degree of inflammatory activity.161
Several studies have evaluated the use of chromoendoscopy as an adjunctive method to diagnose dysplasia or cancer.60, 161, 162, 163, 164 In a study by Kiesslich et al, 165 patients with long-standing UC were randomized to conventional colonoscopy or colonoscopy with chromoendoscopy using 0.1% methylene blue.161 In the chromoendoscopy group, there was better correlation between the endoscopic assessment of both the degree (P = .0002) and the extent (89% vs. 52%; P < .0001) of colonic inflammation compared with conventional colonoscopy. In addition, more targeted biopsies were possible with chromoendoscopy, and significantly more dysplasia was detected (32 vs 10; P = .003). The sensitivity and specificity for differentiation of nonneoplastic from neoplastic lesions were 93%. In a second “back-to-back” colonoscopy study, 100 patients with long-standing UC underwent conventional colonoscopy with both random and directed biopsies, followed by spraying with indigo carmine and then directed biopsies.162 Withdrawal times were similar, at 10 minutes for colonoscopy with multiple biopsies and 11 minutes for colonoscopy with chromoendoscopy and directed biopsies. In this study, there was no dysplasia in 2904 nontargeted biopsy specimens with 43 mucosal abnormalities in 20 patients in the pre–dye spray patients, of which 2 were dysplastic. After dye spraying, an additional 114 abnormalities were detected in 55 patients, of which 7 were dysplastic. The investigators in this study concluded that careful mucosal examination, aided by pancolonic chromoendoscopy and targeted biopsies of suspicious lesions, may be a more effective surveillance method than routine endoscopy with multiple nontargeted biopsies. In a third nonrandomized study, a total of 350 patients with long-standing UC for more than 8 years underwent surveillance colonoscopy using indigo carmine and a magnification scope (high-magnification chromoscopic colonoscopy) applied when mucosal abnormalities were identified. These patients were compared with 350 disease duration– and disease extent–matched control patients who underwent conventional colonoscopic surveillance.163 Significantly more dysplastic lesions were detected in the magnification chromoscopy group compared with controls (69 vs. 24; P < .0001). Dysplasia was observed in 67 lesions, of which 53 (79%) were detected using magnification chromoscopy alone. Chromoscopy increased the number of flat dysplastic lesions detected compared with controls (P < .001). Twenty dysplastic lesions were detected in 12,850 nontargeted biopsy specimens in the high-magnification chromoscopic colonoscopy group (0.16%), in contrast to 49 dysplastic lesions detected from the 644 targeted biopsy specimens in the high-magnification chromoscopic colonoscopy group (8%). Of the 12,482 nontargeted biopsy specimens obtained from the control patients, 18 (0.14%) showed dysplasia. The median extubation time was significantly longer (P < .02) in the high-magnification chromoscopic colonoscopy group (24 minutes; range, 10–36 minutes) compared with controls (13 minutes; range, 8–24 minutes). The investigators in that study concluded that magnification chromoscopy improves detection of dysplasia in patients with UC. Furthermore, neoplastic and nonneoplastic mucosal changes can be predicted with high accuracy using magnification techniques. The only study from the United States analyzed 115 patients with IBD at high risk for dysplasia, many of whom had a history of dysplasia.164 Each patient underwent colonoscopy with random biopsy specimens obtained from all 4 quadrants every 10 cm, had biopsy specimens obtained from suspicious lesions, and also had biopsies performed on lesions detected by methylene blue chromoendoscopy. After dye spray, targeted biopsy specimens revealed significantly more patients with dysplasia (16 LGD, 1 HGD) compared with random biopsies (3 LGD) or targeted nondye spray (8 LGD, 1 HGD). Targeted biopsy specimens, either with and without dye spray, detected dysplasia in 20 patients compared with only 3 patients using the random biopsy approach. Although these data are encouraging, what is not known at this time is whether the detection of dysplasia by chromoendoscopic methods is associated with better clinical outcomes compared with white light endoscopy. Also, with improved white light technology (high-definition and magnification colonoscopes), it is not clear whether the enhanced detection of lesions by chromoendoscopy will remain greater than standard colonoscopy. Thus, at this time, normal white light colonoscopy, using standard or high-definition colonoscopes along with multiple colon biopsies, remains a reasonable method of surveillance for patients with IBD. However, chromoendoscopy with targeted biopsies is considered an acceptable alternative to white light endoscopy for endoscopists who have experience with this technique.
Newer endoscopic techniques are being explored to aid in the diagnosis of dysplasia in IBD, although none of these techniques has yet been rigorously studied. New techniques include narrow band imaging,165 fluorescence endoscopy,166, 167 optical coherence tomography,168 and confocal endomicroscopy.168, 169, 170 In a small study of fluorescence endoscopy in 37 patients with UC, Messmann et al reported that the sensitivity of fluorescence endoscopy for detection of dysplastic lesions was excellent and ranged from 87% (95% CI, 0.73–1.00) to 100% (95% CI, 1.00–1.00) after local sensitization.168 Another promising area is confocal laser endomicroscopy, which allows in vivo histology during endoscopy.169 Kiesslich et al randomized 161 patients with UC in remission to conventional colonoscopy or panchromoendoscopy using 0.1% methylene blue in conjunction with endomicroscopy to detect dysplasia or CRC.167 In the conventional colonoscopy group (n = 73), random biopsy examinations and targeted biopsy examinations were performed. In the endomicroscopy group (n = 80), circumscribed mucosal lesions were identified by chromoscopy and then evaluated for targeted biopsy by endomicroscopy. The primary outcome variable was histologic evidence of neoplasia. Using chromoendoscopy in conjunction with endomicroscopy (average examination time, 42 minutes), significantly more intraepithelial neoplasia could be detected (19 vs 4 cases; P = .007) than with standard colonoscopy (average examination time, 31 minutes). A total of 5580 confocal images from 134 circumscript lesions were compared with histologic results from 311 biopsy specimens. By using chromoscopy with endomicroscopy, 4.75-fold more neoplastic lesions were detected (P = .005) compared with conventional colonoscopy, and 50% fewer biopsy specimens (P = .008) were required. If biopsies were performed on only circumscribed lesions in the conventional colonoscopy group, the total number of biopsy specimens could have been reduced by more than 90%. The presence of neoplastic changes was predicted with high accuracy (94.7% sensitivity, 98.3% specificity, and 97.8% accuracy).
Should Chemopreventive Agents Be Used to Lower the Risk of Developing Dysplasia or CRC in IBD?
In the general population without IBD, it is well established that chemoprevention can reduce the incidence of adenomas and CRC.171 The notion that chemoprevention might reduce the incidence of colorectal neoplasia in IBD is particularly attractive because even though surveillance colonoscopy affords relative protection, some patients still develop CRC despite seemingly optimal surveillance. Curiously, some of the most effective chemopreventive agents for sporadic CRC are anti-inflammatory medicines (aspirin, nonsteroidal anti-inflammatory drugs, cyclooxygenase-2 inhibitors). Indeed, this has been interpreted as evidence that even sporadic CRC is somehow driven by pathways involved in inflammation. Because colonic neoplasia in IBD is even more certainly believed to arise from long-standing chronic inflammation, it stands to reason that patients who use anti-inflammatory medications for their IBD might be relatively protected from developing colonic neoplasia. While the findings of several studies support this concept, the evidence is conflicting due to differences in study design and details of medication use.
The chemopreventive agents that have been studied in IBD include aminosalicylates (mesalamine, mesalazine), corticosteroids, immunomodulators, folic acid, and ursodeoxycholic acid (UDCA) (Table 6). With the exception of 2 prospective studies of UDCA, all studies have been retrospective, using a case-control or rarely a cohort design, with a few studies drawing conclusions from population registries or pharmaceutical databases. Studies vary as to whether they include dysplasia along with CRC as the neoplastic end point. In many instances, the chemopreventive agent was not the primary variable studied. To be sure, in retrospective studies, it is quite difficult to account for changes in medication dosages over time, especially using a case-control study design. Any attempt to overcome these various biases using a more definitive prospective, randomized, placebo-controlled trial is not likely to occur because it would take several thousand patients to be monitored for many years and withholding standard anti-inflammatory treatments such as mesalamine or immunomodulators in a control group is not ethically feasible. Even a study design that does not withhold active medication but instead analyzes a dose-response relationship among users would require a large number of patients to be followed up for many years.172 Indeed, it has been estimated that more than 3000 patients without dysplasia would have to be followed up to detect a significant effect of mesalamine on preventing progression to neoplasia.112
Table 6. Chemoprevention: Risk of Developing CRC and/or Dysplasia in Patients With IBD
| Author, year | Design | Agent | OR or RR | 95% CI | P value |
|---|---|---|---|---|---|
| UDCA | |||||
| Tung et al, 2001173 | Case-control; PSC | UDCA (any vs none) | 0.18 | 0.05–0.61 | .005 |
| Pardi et al, 2003174 | Prospective; PSC | UDCA (median, 42 mo) | 0.26 | 0.06–0.99 | .034 |
| Wolf et al, 2005175 | Retrospective; PSC | UDCA (at least 6 mo vs never) | 0.56 | 0.25–1.30 | NS |
| Sjoqvist et al, 2004176 | Prospective | UDCA (500 mg twice daily vs placebo) | 0% vs 22% | NA | NS |
| Aminosalicylates | |||||
| Positive studies | |||||
| Lashner et al, 1989179 | Case-control | No sulfasalazine (sulfa allergy) | 11.71 | 1.65–83.10 | <.05 |
| Pinczowski et al, 199415 | Case-control | Sulfasalazine (1+ course >3 mo) | 0.38 | 0.20–0.69 | Significant |
| Moody et al, 199614 | Cohort | Sulfasalazine (compliant) | 0.10 | Not determined | <.001 |
| Eaden et al, 200013 | Case-control | Mesalamine (any vs none) | 0.25 | 0.13–0.48 | <.00001 |
| Mesalazine (<1.2 g/day) | 0.08 | 0.08–0.85 | .04 | ||
| Mesalazine (>1.2 g/day) | 0.09 | 0.03–0.28 | <.00001 | ||
| Sulfasalazine (<2 g/day) | 0.56 | 0.17–1.84 | NS | ||
| Sulfasalazine (>2 g/day) | 0.41 | 0.18–0.92 | .03 | ||
| Olsalazide, balsalazide | 0.40 | 0.04–3.58 | NS | ||
| Aspirin (yes vs no) | 0.80 | 0.21–2.98 | NS | ||
| VanStaa et al, 2005181 | Database; case-control | Mesalamine (irregular use) | Reference | — | — |
| Mesalamine (regular use) | 0.60 | 0.38–0.96 | Significant | ||
| Mesalazine (>30 Rx before) | 0.31 | 0.11–0.84 | Significant | ||
| Mesalazine (≥1.2 g/day) | 0.56 | 0.31–1.03 | NS | ||
| Sulfasalazine (>30 Rx before) | 0.77 | 0.37–1.60 | NS | ||
| Sulfasalazine (≥2 g/day) | 0.69 | 0.35–1.37 | NS | ||
| Rubin et al, 2006182 | Case-control | Mesalamine use (yes vs no) | 0.54 | 0.13–2.21 | NS |
| Mesalamine (>1.2 g/day) | 0.28 | 0.09–0.85 | .024 | ||
| Bansal and Sonnenberg, 1996180 | Database; nonsteroidal anti-inflammatory drug–associated diseases | Risk of CRC mortality | 0.51 | 0.28–0.94 | .03 |
| Risk of CRC incidence | 0.84 | 0.65–1.08 | NS | ||
| Negative studies | |||||
| Lashner et al, 1989179 | Case-control | Sulfasalazine use (not defined) | 1.50 | 0.43–5.19 | NS |
| Lashner et al, 1997186 | Cohort | Sulfasalazine (>6 mo) | 0.95 | 0.34–2.70 | NS |
| Mesalamine (>6 mo) | 0.88 | 0.21–3.73 | NS | ||
| Shetty et al, 199934 | Case-control; PSC | Sulfasalazine (>6 mo) | 1.11 | 0.87–1.42 | NS |
| Tung et al, 2001173 | Case-control; PSC | Sulfasalazine (not defined) | 3.30 | 0.8–14 | NS |
| Mesalamine (not defined) | 0.88 | 0.25–3.2 | NS | ||
| Lindberg et al, 2001183 | Cohort | Sulfasalazine (>6 mo vs <6 mo) | 34% vs 44% | NA | NS |
| Bernstein et al, 2003184 | Database; case-control | Mesalamine (>2 mo vs nonusers) | 1.46 | 0.58–3.73 | NS |
| Mesalamine (high-dose; long duration) | 0.48 | 0.07–3.25 | NS | ||
| Rutter et al, 200448 | Case-control | Mesalamine (never or <3 mo) | Reference | — | — |
| Mesalamine (≤10 yr) | 0.52 | 0.20–1.34 | NS | ||
| Mesalamine (>10 y) | 0.63 | 0.18–2.21 | NS | ||
| Sulfasalazine (never or <3 mo) | Reference | — | — | ||
| Sulfasalazine (≤10 y) | 1.59 | 0.53–4.79 | NS | ||
| Sulfasalazine (>10 y) | 1.36 | 0.48–3.81 | NS | ||
| Terdiman et al, 2007185 | Database; case-control | Mesalamine use in 12 mo before CRC diagnosis | 0.97 | 0.77–1.23 | NS |
| Gupta et al, 200749 | Cohort; retrospective | Mesalamine (use >4 mo) | 0.5 | 0.1–2.4 | NS |
| Ullman et al, 2008112 | Cohort; retrospective | Mesalamine (any vs none) | 0.7 | 0.2–2.44 | NS |
| Mesalamine (>2 g/day vs ≤2 g/day) | 0.77 | 0.22–2.67 | NS | ||
| Mesalamine (each 1 g/day increment vs none) | 0.92 | 0.58–1.47 | NS | ||
| Corticosteroids | |||||
| Eaden et al, 200013 | Case-control | Systemic corticosteroids (yes vs no) | 0.26 | 0.01–0.70 | .008 |
| Local corticosteroids (yes vs no) | 0.44 | 0.19–1.02 | NS | ||
| Lashner et al, 1989179 | Case-control | Prednisone use (not defined) | 0.43 | 0.11–1.68 | NS |
| Lashner et al, 1997186 | Cohort | Prednisone (>6 mo) | 1.52 | 0.55–4.16 | NS |
| Shetty et al, 199934 | Case-control; PSC | Prednisone (>6 mo) | 1.03 | 1.00–1.06 | NS |
| Tung et al, 2001173 | Case-control; PSC | Prednisone (not defined) | 0.50 | 0.16–1.5 | NS |
| Gupta et al, 200749 | Cohort; retrospective | Oral corticosteroids (use >4 mo) | 0.6 | 0.2–1.7 | NS |
| Immunomodulators | |||||
| Lashner et al, 1997186 | Cohort | AZA (>6 mo) | 1.12 | 0.26–4.77 | NS |
| Tung et al, 2001173 | Case-control; PSC | AZA (not defined) | 0.68 | 0.17–2.6 | NS |
| Methotrexate (not defined) | 2.7 | 0.23–31 | NS | ||
| Cyclosporine (not defined) | 0.57 | 0.15–2.2 | NS | ||
| Rutter et al, 200448 | Case-control | AZA (never) | Reference | — | — |
| AZA (1–5 y) | 0.47 | 0.09–2.43 | NS | ||
| AZA (>5 y) | 0.19 | 0.02–1.48 | NS | ||
| Matula et al, 2005192 | Cohort; retrospective | AZA, 6-MP | |||
| Gupta et al, 200749 | Cohort; retrospective | AZA, 6-MP (use >4 mo) | 0.8 | 0.3–2.7 | NS |
| Folate | |||||
| Lashner et al, 1989179 | Case-control | Folic acid (yes vs no) | 0.38 | 0.12–1.20 | NS |
| Lashner et al, 1997186 | Cohort | Folic acid (0.4 mg/day; >6 mo) | 0.76 | 0.36–1.61 | NS |
| Folic acid (1.0 mg/day; >6 mo) | 0.54 | 0.20–1.48 | NS | ||
| Rutter et al, 200448 | Case-control | Folic acid (yes vs no) | 0.67 | 0.07–6.41 | NS |
| Gupta et al, 200749 | Cohort; retrospective | Folic acid (use >4 mo) | 1.3 | 0.4–3.7 | NS |
| Statins | |||||
| Poynter et al, 2005198 | Case-control; sporadic CRC | Statin use for at least 5 y | 0.57 | 0.44–0.73 | <.05 |
| Case-control; IBD subset | Statin use for at least 5 y | 0.06 | 0.006–0.55 | <.05 |
UDCA
Patients with PSC and UC are among those at the highest risk for CRC. Most patients with PSC are treated with UDCA for their cholestatic liver disease. A retrospective cross-sectional study of patients with UC who have PSC demonstrated that use of UDCA was strongly associated with decreased incidence (OR, 0.18; 95% CI, 0.05–0.61) of colonic dysplasia173 (Table 6). This protective effect remained after adjusting for duration of colitis, age at onset of colitis, and sulfasalazine use. Likewise, in a prospective study of UDCA therapy in 52 patients with PSC followed up at the Mayo Clinic, use of UDCA was associated with a significant reduction (RR, 0.26; 95% CI, 0.06–0.92) in the development of dysplasia and cancer.174 Another study found no chemopreventive effect for dysplasia or CRC in patients with UC and PSC who were treated with UDCA for at least 6 months.34 Others have reported no chemopreventive effect for UDCA in patients with UC and PSC, although the point estimates favored a protective effect.175 While little is known about whether UDCA might be chemopreventive for patients with UC who do not have PSC, in a small study from Sweden, patients with LGD and/or DNA aneuploidy (most of whom did not have PSC) were randomized to receive UDCA (500 mg twice daily) or placebo for 2 years.176 Two of 9 patients receiving placebo progressed to HGD or a DALM requiring colectomy, whereas none of the 10 patients treated with UDCA progressed. Use of UDCA has been associated with a significant reduction in recurrent sporadic colonic adenomas with HGD.177 While its antineoplastic mechanism is not known, UDCA reduces the colonic concentration of potentially carcinogenic bile acids such as deoxycholic acid, inhibits ras gene mutations and cyclooxygenase-2 expression, and has antioxidant activity. Thus, UDCA appears to be a potentially useful chemopreventive agent worthy of further investigation in patients with IBD. (USPSTF Grade A: High certainty that the magnitude of net benefits is substantial. UDCA has demonstrated a significant reduction in CRC in patients with UC who also have PSC.)
Aminosalicylates
Aminosalicylates are an attractive class of potential chemopreventive agents because they are safe, relatively inexpensive, and already used for maintenance therapy by many patients with IBD on a long-term basis. Population-based studies in Scandinavia indicate a lack of increased risk of CRC compared with the general population that has been attributed not only to a more liberal colectomy rate but also to a more widespread use of aminosalicylates.12, 178
Several studies have observed a significant chemopreventive effect of mesalamine compounds (Table 6). Investigators in Leicestershire, England, reported an approximate 75% to 90% reduction in the incidence of CRC with the use of mesalamine agents.13 While mesalazine was more protective than sulfasalazine, both compounds demonstrated a dose-response effect, thereby supporting an apparent protective role (Table 6). Even after adjusting for other variables, mesalazine at doses >1.2 g/day reduced the risk of cancer by 81% (P = .006). Although that study has been criticized because cases were drawn from a different population than controls, other studies have also supported a protective role for mesalamine.14, 15, 179, 180, 181, 182 A case-control study from the University of Chicago, designed to examine the potential chemopreventive role of mesalamine, compared 26 patients with UC with dysplasia or CRC to 96 UC controls without neoplasia and reported that patients who took at least 1.2 g/day of mesalamine had a 72% risk reduction for dysplasia and CRC and that as the dose of mesalamine increased, the odds of neoplasia decreased.182 In addition, a study from the Mayo Clinic compared 188 UC cases with CRC to 188 UC controls without CRC to identify predictive and protective factors and found a 60% reduction in CRC for patients who took mesalamine for 1 to 5 years compared with <1 year.61 After adjusting for exposure to surveillance colonoscopy and other anti-inflammatory therapy, there was still a 50% reduction in CRC risk associated with 1 to 5 years of mesalamine use. This effect is comparable to 2 large population-based studies reporting substantial reductions in CRC among patients treated with mesalamine. A Swedish population-based cohort of 3112 patients with UC that investigated risk factors for developing CRC compared 102 cases of CRC with 196 matched controls without CRC.15 A significant 62% reduction in CRC was found among individuals who used sulfasalazine for at least 3 months. A study utilizing the UK General Practitioner Research Database to identify users of mesalamine medications who also had a history of IBD found that regular users of mesalamine had a 40% lower incidence of CRC, especially those prescribed mesalazine rather than sulfasalazine, and those given prescriptions over a longer period.181 Moody et al retrospectively analyzed a 10-year cohort of patients with UC (1972–1981) and noted that the crude proportion developing CRC was 3% (5/152 patients) in patients treated with mesalamine on a long-term basis compared with 31% (5/16 patients) in those who stopped treatment or could not comply.14 Indirect evidence comes from a case-control study of folate as a chemopreventive, in which patients who were allergic to sulfa (and therefore did not take sulfasalazine) were found to have an 11-fold greater risk of developing CRC.179 Another study utilizing the US Department of Veterans Affairs national database selected patients hospitalized between 1981 and 1993 who had a diagnosis of IBD (UC and CD) and those with CRC.180 Among patients with IBD, those with a history of a disease associated with the use of nonsteroidal anti-inflammatory drugs had a significant (49%) reduction in CRC mortality and a nonsignificant (16%) reduction in CRC incidence.
In contrast to these positive studies, there are other reports suggesting a lack of chemopreventive effect for mesalamine. A retrospective cohort study from Mount Sinai Hospital in New York examined the use of mesalamine among patients with no initial dysplasia on the rate of progression to either HGD or CRC.112 While the point estimates of the hazard ratio suggested a 30% reduction for any mesalamine use, a 23% reduction for a dose of >2 g/day, and an 8% reduction for each 1 g/day increase in dose of mesalamine, these did not reach statistical significance. A Swedish study of 143 patients with UC who underwent regular surveillance colonoscopies over a 20-year period before 1993 retrospectively analyzed the use of sulfasalazine or mesalamine treatment for at least 6 months.183 There was a slightly lower rate of developing dysplasia or CRC among users compared with nonusers (34% vs 44%), but this did not achieve statistical significance. Two studies using large registries also reported a lack of chemopreventive effect, but the duration of exposure to mesalamine was limited. A study using the Canadian Manitoba IBD Epidemiology Database found no chemopreventive effect for mesalamine, although there was a trend toward benefit with higher dose and longer use of mesalamine.184 An issue with that study is that it only analyzed mesalamine use within 2 years of the diagnosis of CRC, and compliance could not be assessed. Likewise, a study from the United States utilizing 2 large administrative claims databases found that exposure to mesalamine therapy of any dose or duration during the 12 months before diagnosis of CRC was not associated with a reduced risk of CRC, although a trend toward reduced risk was observed with an increasing number of prescriptions in the previous year.185 Other negative studies of mesalamine were not designed to study mesalamine use as the primary variable and either did not provide sufficient details regarding dose or duration34, 173, 179, 183, 186 or studied high-risk patients with PSC, where any possible effect of mesalamine may be hard to achieve.34, 173 Two recent studies indicate that more severe degrees of histologic inflammation correlate with increased risk of colonic dysplasia and CRC.48, 49 In both studies, use of mesalamine was not associated with a lowering of this risk (Table 6).
Velayos et al performed a meta-analysis in 2005 whereby pooled analysis demonstrated a preventive effect of mesalamine for CRC (OR, 0.51; 95% CI, 0.37–0.69) and for CRC plus dysplasia (OR, 0.51; 95% CI, 0.38–0.69).187 This benefit occurred with regular use or use of at least 1.2 g/day of mesalamine equivalents. Use of mesalamine was not significantly associated with a lower risk of dysplasia (OR, 1.18; 95% CI, 0.41–3.43), although only 2 studies evaluated this outcome. Most studies have noted that sulfasalazine appears to have less of an effect than mesalamine (Table 6). While this may be due to the inability of patients to tolerate higher doses of sulfasalazine, it may also relate to the fact that because sulfasalazine competitively inhibits folic acid absorption, folate depletion in these patients might counteract any apparent chemopreventive benefit from sulfasalazine.
If, indeed, mesalamine compounds are chemopreventive, where in the dysplasia-carcinoma sequence might they work? Only one study has attempted to address this question. Given the inherent difficulties of conducting a prospective interventional study, Ullman et al conducted a retrospective cohort study in which patients with UC were identified with no dysplasia, indefinite dysplasia, or flat LGD at a baseline time point in the past and then “followed” for the development of HGD or CRC while abstracting data on use of mesalamine (as well as surveillance colonoscopies)112 (Table 6). For patients who initially had no dysplasia, the hazard ratio for those taking high-dose mesalamine (>2 g/day) was 0.77 (95% CI, 0.22–2.67). Interestingly, among patients with initial indefinite dysplasia, none of the 27 patients taking high-dose mesalamine progressed to HGD/CRC, whereas 4 of 29 patients (13%) taking low-dose mesalamine did progress (P = .11). In the group with initial flat LGD, the dose of mesalamine had no effect on progression to HGD/CRC. These data suggest that if mesalamine has any chemopreventive effect, it may act early in the neoplastic progression, before the development of LGD.
Aminosalicylates may have several chemopreventive mechanisms. They can exert anti-inflammatory activity by activating peroxisome proliferator activated receptor γ, scavenging oxygen-derived free radicals, inhibiting lipoxygenase, and blocking nuclear factor κB.188, 189 Mesalazine has also been shown to promote apoptosis in sporadic CRC tissues but to have no such effect on adjacent normal mucosal tissue.190 Moreover, mesalazine improves DNA replication fidelity in CRC cells.191 (USPSTF Grade B: Moderate certainty that the magnitude of net benefits is moderate. Aminosalicylates are chemopreventive against CRC.)
Corticosteroids
If mesalamine compounds prevent colonic neoplasia by suppressing inflammation, it might be expected that other anti-inflammatory medications used to treat patients with IBD would also be protective against CRC. While not a primary variable in any study, corticosteroid use has been analyzed in several studies, most of which found no chemopreventive effect, although details regarding dose and duration are not available (Table 6). The Leicestershire group reported that systemic corticosteroids, and to a lesser extent topical corticosteroids, resulted in a significant CRC risk reduction.13 The Mayo Clinic study also found a significant reduction in CRC with oral prednisone use for more than 1 year.61 Whether or not this supports the notion that inflammation predisposes to colonic neoplasia, chronic corticosteroid use cannot be condoned for chemoprevention due to its significant toxicity. (USPSTF Grade D: High certainty that the magnitude of net benefits is negative. Oral or topical corticosteroids, while demonstrating antineoplastic effects in 2 studies, are associated with too many side effects to warrant use as chemopreventive agents.)
Immunomodulators and Biologic Agents
Long-term use of immunomodulators (particularly azathioprine [AZA] and 6-mercaptopurine [6-MP]) is a viable consideration for chemoprevention because these agents as already a cornerstone of maintenance therapy. Here again, little is known regarding purine analogues are chemopreventive agents because, with one recent exception, no study examined these agents as the primary variable for CRC risk reduction in IBD. The most robust study comes from Mount Sinai Hospital in New York, which analyzed a retrospective cohort of patients with UC who were initially free of dysplasia in whom the use of 6-MP or AZA (expressed as 6-MP equivalents) was examined using a statistical method that accounted for changes in dose over time.192 Regardless of whether 6-MP use was analyzed as any exposure, greater than or less than 25 mg/day, or average daily dose, there was no protective effect of 6-MP on progression to colorectal dysplasia or CRC. The St Mark's study on the effect of inflammation on developing colorectal neoplasia analyzed AZA exposure of less than 1 year, 1 to 5 years, or more than 5 years as a secondary variable and found a nonsignificant trend toward a protective effect of AZA use for 1 to 5 years (OR, 0.34; 95% CI, 0.09–1.25) and greater than 5 years (OR, 0.73; 95% CI, 0.30–1.78).48 A similar study from Mount Sinai Hospital investigating the role of histologic inflammation on neoplasia progression also showed no significant risk reduction of immunomodulators.49 In patients with UC and PSC, use of AZA (dose and duration not specified) was analyzed as a secondary variable and found to have no chemopreventive effect in this high-risk subset of patients.173 Likewise, the use of AZA for at least 6 months (dose and duration not specified) also had no effect on the risk of dysplasia or CRC.179
Likewise, other studies compared patients with IBD who were either users or nonusers of immunomodulators and looked at the rates of CRC in their patients. A case-control study of 86 patients who had received AZA and 180 matched controls from St Mark's Hospital found no protective effect of long-term AZA use among patients with extensive UC.193 Another study of 626 patients with UC and CD from Oxford, England, found no difference in the rate of CRC or HGD among long-term users of AZA, although doses were not given.194 Similarly, Korelitz et al tracked 550 patients who had received 6-MP between 1969 and 1997.195 Only 8 patients developed CRC (1.6%), a value not greater than that observed with long-standing colitis. Thus, while we may conclude that purine analogue immunomodulators do not appear to lower the risk of CRC in IBD, it is reassuring that they also do not increase this risk. Any possible chemopreventive effect of anti–tumor necrosis factor therapy has not been studied to date. (USPSTF Grade Insufficient: no recommendation, insufficient evidence to recommend for or against the use of thiopurines. AZA or 6-MP has not been consistently associated with lower rates of CRC.)
Folic Acid/Calcium/Multivitamins/Statins
In the setting of sporadic CRC, low folate intake has been associated with an increased risk of developing colorectal adenomas and carcinomas.196, 197 Patients with chronic IBD are predisposed to folate deficiency because of inadequate nutritional intake, excessive intestinal losses with active disease, and, in those taking sulfasalazine, reduced intestinal absorption due to competitive inhibition. Results of 4 studies, 2 of which analyzed folate use as the primary variable,179, 186 suggest a trend toward protection against CRC in folate users, but none of them demonstrated statistical significance (Table 6). Accurate data on folate use are difficult to ascertain because patients may take folate in various supplements without realizing it, and physicians often neglect to document folate use in medical records. There is insufficient evidence for or against calcium or multivitamins at this time. In a population-based case-control study of patients who had a diagnosis of CRC in northern Israel between 1998 and 2004, statin use for at least 5 years was associated with a modest reduction in CRC in the non-IBD population but a substantial 94% risk reduction among a subset of 55 patients with IBD198 (Table 6). Further studies are needed to verify whether statins may be chemopreventive in patients with IBD. (USPSTF Grade Insufficient: No recommendation, insufficient evidence to recommend for or against the use of supplements. Folic acid supplements, calcium, or multivitamins, have not been consistently associated with lower rates of CRC. Insufficient evidence to recommend for or against the use of statins as chemopreventive agents.)
Should Molecular Markers Be Applied to Help Stratify Patients Into Low-Risk and High-Risk Groups?
Knowledge about the molecular pathogenesis of colitis-associated cancer has been derived from studies of sporadic colon carcinogenesis. Not surprisingly, many of the molecular alterations responsible for sporadic CRC also play a role in colitis-associated cancer. The 2 major genetic pathways, chromosomal instability and microsatellite instability (MSI), occur with, roughly, the same frequency in sporadic CRC and colitis-associated cancer (chromosomal instability, 80%; MSI, 20%).199, 200 In addition, the CpG island methylation pathway, an epigenetic alteration, contributes to carcinogenesis in both sporadic colon carcinogenesis and colitis-associated cancer. A complete review of molecular genetic abnormalities in colitis-associated cancer is beyond the scope of this review; the reader is referred to other reports on this topic.201
To date, much of our understanding about the types of molecular alterations in colitis-associated cancer has been derived from cross-sectional studies that evaluated a marker of interest at only one point in time, from lesions that represent the pathologic spectrum of no dysplasia, indefinite for dysplasia, LGD, HGD, and cancer. In these studies, genetic alterations that demonstrated a preferential, or increased, expression in neoplastic tissues was considered a potentially useful marker of neoplasia. Thus, when a marker was expressed in nondysplastic tissue, from a patient who also had dysplasia or cancer elsewhere in the colon, the marker was considered to be expressed “earlier” than dysplasia.
From a clinical standpoint, however, a more accurate estimate of whether a marker is expressed “early” in carcinogenesis is to study patients longitudinally over time. Because patients with IBD typically undergo periodic surveillance colonoscopies with repeated tissue sampling, this provides the rather unique opportunity to test whether a marker correlates with subsequent development of dysplasia or cancer. To date, few studies have used this type of chronological study design. Four tissue-based markers, aneuploidy, p53, MSI, and the mucin-associated STn antigen, have been evaluated in a chronological context, and each has been shown to be a harbinger of subsequent risk of developing dysplasia or cancer. As noted previously, since the early 1980s, aneuploidy was detected more diffusely throughout the colon than dysplasia, and it usually (but not always) preceded or arose coincident with the initial detection of dysplasia.150, 202, 203, 204 Overexpression of p53 was observed in nondysplastic mucosa from patients with UC who later developed colitis-associated cancer.94, 101 Other investigators described 2 patients who had either colitis-associated cancer or HGD, both of whom had abnormal p53 expression in dysplastic lesions 2 years before the diagnosis.205 In a limited study, 4 patients with UC and high MSI CRCs manifested high MSI in nondysplastic biopsy specimens of chronic colitis anywhere from 2 to 12 years before the diagnosis of colitis-associated cancer.206 STn antigen expression appears to be more frequent in UC colons with dysplasia and is expressed earlier and independently from aneuploidy.103, 104
In addition to these tissue-based studies, some investigators have evaluated k-ras mutations in serum207 and p53 and k-ras mutations in colonic effluent,208 but to date, the results are preliminary. Because the DNA shed into stool should theoretically provide a more comprehensive sampling of abnormal cells than random pinch biopsy specimens or even biopsy specimens directed by chromoendoscopy, stool DNA testing could potentially enhance the management of patients with long-standing IBD who are at risk for developing CRC. Stool DNA testing, using a panel that incorporates markers of chromosomal instability and MSI, has already shown reasonable sensitivity and specificity for detecting sporadic CRCs.209 A newer version of the stool DNA test that incorporates a marker of gene methylation has demonstrated a sensitivity of 88% for sporadic CRC.210 It is currently unknown whether these stool DNA marker panels will be useful for detecting neoplasia in patients with IBD.
The potential for molecular profiling of patients is emerging in the field of IBD neoplasia. A recent study identified 699 transcripts that differed between benign mucosa and HGD and 242 transcripts that differed between benign mucosa and colitis-associated cancer, but additional studies are needed to confirm these findings and elucidate their relevance to the biology of neoplasia in UC.211
At present, there is no consensus as to how, or whether, any of these markers of cancer risk should be incorporated into clinical management of patients with long-standing IBD. Colectomy can hardly be recommended to a patient solely on the basis of marker positivity without evidence of dysplasia, even if the patient's tissue demonstrated marker positivity on several colonoscopies. Future studies should investigate whether molecular markers help to further stratify patient risk for colonic neoplasia in IBD. (USPSTF Grade Insufficient: No recommendation; insufficient evidence to recommend for or against the use of molecular markers. Molecular markers should not yet be applied to help stratify patients into low- and high-risk groups.)
Areas of Future Research
Research and patient care in the field of IBD-associated colorectal neoplasia has become more enlightened in the last decade. It has become clear that raised dysplasia may not be as ominous as previously believed, as long as complete endoscopic resection can be assured. It is also becoming apparent that dysplastic lesions in the colitic colon are, in fact, visible and that newer, more sensitive imaging methods such as chromoendoscopy can improve dysplasia detection rates. Just as the quality of the colonoscopic examination has become an important metric in gastroenterology practice, this is perhaps even more important in the IBD population, where the risk of cancer is higher. To better enable a comparison across studies, it will be important for future publications in this field to specify the nature of the dysplastic lesions encountered (eg, flat vs raised, endoscopic resectability, whether the lesions was found by special imaging techniques).
In addition, among the many unanswered questions regarding CRC risk in IBD, Table 7 provides a partial list of areas that merit further investigation.
Table 7. Areas Worthy of Future Research
| What is the outcome of surveillance colonoscopy in patients with Crohn's colitis? |
| How should we incorporate patient preferences and risk perception in surveillance programs? |
| What is the natural history of chromoendoscopically detected dysplasia? |
| How will newer white light technology compare with other image-enhancing techniques? |
| Do anti–tumor necrosis factor agents have chemopreventive properties? |
| Can molecular markers in tissue, blood, or stool enhance risk assessment? |
Acknowledgments
NOTE: Dr Robin McLeod, a member of the AGA Institute Medical Position Panel, believes that in view of the high-risk (19% to 27%) of underlying cancer, in most situations patients should be offered colectomy when there is flat LGD which has been confirmed by an experienced pathologist.
The authors would like to acknowledge the expert assistance of Sheila A. Agyeman.
References
- . The sigmoidoscopic picture of chronic ulcerative colitis (non-specific). Am J Med Sci. 1925;170:220–228
- . Cicatrizing enteritis as a pathological entity. Am J Pathol. 1948;24:475–501
- . The risk of colorectal cancer in ulcerative colitis: a meta-analysis. Gut. 2001;48:526–535
- . Long term prognosis in ulcerative colitis—based on results from a regional patient group from the county of Copenhagen. Gut. 1985;26:158–163
- Cancer in universal and left-sided ulcerative colitis: factors determining risk. Gastroenterology. 1979;77:290–294
- Thirty-year analysis of a colonoscopic surveillance program for neoplasia in ulcerative colitis. Gastroenterology. 2006;130:1030–1038
- Risk of intestinal cancer in inflammatory bowel disease: a population-based study from Oolmsted Ccounty, Minnesota. Gastroenterology. 2006;130:1039–1046
- . Epidemiology and risk factors for colorectal dysplasia and cancer in ulcerative colitis. Gastroenterol Clin North Am. 2006;35:517–531
- Crohn's disease and cancer. N Engl J Med. 1973;289:1099–1103
- Intestinal cancer risk and mortality in patients with Crohn's disease. Gastroenterology. 1993;105:1716–1723
- Increased risk of large-bowel cancer in Crohn's disease with colonic involvement. Lancet. 1990;336:357–359
- Intestinal and extra-intestinal cancer in Crohn's disease: follow-up of a population-based cohort in Copenhagen County, Denmark. Aliment Pharmacol Ther. 2004;19:287–293
- Colorectal cancer prevention in ulcerative colitis: a case-control study. Aliment Pharmacol Ther. 2000;14:145–153
- Long-term therapy with sulphasalazine protects against colorectal cancer in ulcerative colitis: a retrospective study of colorectal cancer risk and compliance with treatment in Leicestershire. Eur J Gastroenterol Hepatol. 1996;8:1179–1183
- Risk factors for colorectal cancer in patients with ulcerative colitis: a case-control study. Gastroenterology. 1994;107:117–120
- Increased risk of intestinal cancer in Crohn's disease: a meta-analysis of population-based cohort studies. Am J Gastroenterol. 2005;100:2724–2729
- . Meta-analysis: colorectal and small bowel cancer risk in patients with Crohn's disease. Aliment Pharmacol Ther. 2006;23:1097–1104
- Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer. 2001;91:854–862
- Ulcerative colitis and Crohn's disease: a comparison of the colorectal cancer risk in extensive colitis. Gut. 1994;35:1590–1592
- . Similarity of colorectal cancer in Crohn's disease and ulcerative colitis: implications for carcinogenesis and prevention. Gut. 1994;35:950–954
- High frequency of early colorectal cancer in inflammatory bowel disease. Gut. 2008;57:1246–1251
- Crohn's disease and colorectal cancer. Gut. 1994;35:651–655
- Colorectal cancer in ulcerative colitis: a cohort study of primary referrals from three centres. Gut. 1988;29:206–217
- Ulcerative colitis and colorectal cancer (A population-based study). N Engl J Med. 1990;323:1228–1233
- Colorectal cancer in ulcerative colitis (Influence of anatomical extent and age at onset on colitis-cancer interval). Gut. 1991;32:167–169
- Backwash ileitis is strongly associated with colorectal carcinoma in ulcerative colitis. Gastroenterology. 2001;120:841–847
- Pathologic features and clinical significance of “backwash” ileitis in ulcerative colitis. Am J Surg Pathol. 2005;29:1472–1481
- Primary sclerosing cholangitis and ulcerative colitis: evidence for increased neoplastic potential. Hepatology. 1995;22:1404–1408
- . Liver involvement in patients operated for ulcerative colitis, with special reference to the association of cholangitis with colorectal dysplasia and carcinoma. Int J Colorectal Dis. 2000;15:167–171
- Risk and natural history of colonic neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis. Gastroenterology. 1996;110:331–338
- . Pericholangitis and sclerosing cholangitis are risk factors for dysplasia and cancer in ulcerative colitis. Am J Gastroenterol. 1993;88:1174–1178
- . Is there an excess risk for colorectal cancer in patients with ulcerative colitis and concomitant primary sclerosing cholangitis? (A population based study). Gut. 1997;41:522–525
- The risk of cancer and dysplasia among ulcerative colitis patients with primary sclerosing cholangitis. Am J Gastroenterol. 1997;92:1285–1288
- The risk for cancer or dysplasia in ulcerative colitis patients with primary sclerosing cholangitis. Am J Gastroenterol. 1999;94:1643–1649
- Risk of colorectal neoplasia in patients with primary sclerosing cholangitis. Gastroenterology. 1996;110:432–440
- Primary sclerosing cholangitis and colorectal carcinoma in patients with chronic ulcerative colitis: a case-control study. Cancer. 1998;82:822–826
- Increased risk of colorectal neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis: a meta-analysis. Gastrointest Endosc. 2002;56:48–54
- Development of colon cancer after liver transplantation for primary sclerosing cholangitis associated with ulcerative colitis. Hepatology. 1990;11:477–480
- Increased risk of early colorectal neoplasms after hepatic transplant in patients with inflammatory bowel disease. Dis Colon Rectum. 1993;36:908–912
- Risk of colorectal neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis following orthotopic liver transplantation. Hepatology. 1998;27:685–690
- Colorectal cancer in patients with inflammatory bowel disease after liver transplantation for primary sclerosing cholangitis. Transplantation. 2003;75:1983–1988
- Familial predisposition for colorectal cancer in chronic ulcerative colitis: a case-control study. Gastroenterology. 1998;115:1079–1083
- Family history as a risk factor for colorectal cancer in inflammatory bowel disease. Gastroenterology. 2001;120:1356–1362
- Colorectal cancer rates among first-degree relatives of patients with inflammatory bowel disease: a population-based cohort study. Lancet. 2001;357:262–266
- Cancer risk and life expectancy of children with ulcerative colitis. N Engl J Med. 1971;285:17–21
- Endoscopic screening for dysplasia and mucosal aneuploidy in adolescents and young adults with childhood onset colitis. Am J Gastroenterol. 1997;92:2001–2006
- Screening and surveillance colonoscopy in chronic Crohn's colitis. Gastroenterology. 2001;120:820–826
- Severity of inflammation is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology. 2004;126(2):451–459
- Histologic inflammation is a risk factor for progression to colorectal neoplasia in ulcerative colitis: a cohort study. Gastroenterology. 2007;133:1099–1105quiz 1340–1341
- Increased inflammatory activity is an independent risk factor for dysplasia and colorectal cancer in ulcerative colitis: a case-control analysis with blinded prospective review of pathology. Gastroenterology. 2006;130(Suppl 2):A2
References (Online Only)
- Gross versus microscopic pancolitis and the occurrence of neoplasia in ulcerative colitis. Inflamm Bowel Dis. 2003;9:351–355
- Lower gastrointestinal malignancy in Crohn's disease. Gut. 1994;35:347–352
- Carcinoma arising in anorectal fistulas of Crohn's disease. Dis Colon Rectum. 1998;41:992–996
- Dysplasia and cancer complicating strictures in ulcerative colitis. Dig Dis Sci. 1990;35:349–352
- Cancer surveillance in longstanding ulcerative colitis: endoscopic appearances help predict cancer risk. Gut. 2004;53:1813–1816
- Adenocarcinoma in strictures of ulcerative colitis without antecedent dysplasia by colonoscopy. Am J Gastroenterol. 1994;89:119–122
- Endoscopic appearance of dysplasia in ulcerative colitis and the role of staining. Endoscopy. 2004;36:1109–1114
- . Benign and malignant colorectal strictures in ulcerative colitis. Gut. 1992;33:938–941
- Dysplasia-associated lesion or mass (DALM) detected by colonoscopy in long-standing ulcerative colitis: an indication for colectomy. Gastroenterology. 1981;80:366–374
- Most dysplasia in ulcerative colitis is visible at colonoscopy. Gastrointest Endosc. 2004;60:334–339
- Predictive and protective factors associated with colorectal cancer in ulcerative colitis: a case-control study. Gastroenterology. 2006;130:1941–1949
- . Cancer in inflammatory bowel disease (An evidence-based analysis and guide for physicians and patients). Gastroenterol Clin North Am. 2002;31:237–254
- . Dysplasia and colorectal cancer in Crohn's disease. J Clin Gastroenterol. 2003;36(5 Suppl):S75–S78discussion S94–S96
- . Are we telling patients the truth about surveillance colonoscopy in ulcerative colitis?. Lancet. 1994;343:71–74
- . Significance and detection of dysplasia in chronic colitis. Cancer. 1996;78:2261–2263
- . Colorectal cancer in idiopathic inflammatory bowel disease. Semin Diagn Pathol. 1996;13:339–357
- Progression of flat low-grade dysplasia to advanced neoplasia in patients with ulcerative colitis. Gastroenterology. 2003;125:1311–1319
- . Diagnosis and management of dysplasia in patients with inflammatory bowel diseases. Gastroenterology. 2004;126:1634–1648
- Intestinal adenocarcinoma in Crohn's disease: a report of 30 cases with a focus on coexisting dysplasia. Am J Surg Pathol. 1999;23:651–655
- . Colorectal carcinoma in patients with Crohn's disease. Gastroenterology. 1985;89:398–407
- . Hold the bone: good news for young women with IBD. Inflamm Bowel Dis. 2003;9:396–397
- Colonoscopic surveillance in long-standing total ulcerative colitis—a 15-year follow-up study. Gastroenterology. 1990;99:1021–1031
- . Surveillance in the routine management of ulcerative colitis: the predictive value of low-grade dysplasia. Gastroenterology. 1992;103:431–438
- Precancer and cancer in extensive ulcerative colitis: findings among 401 patients over 22 years. Gut. 1990;31:800–806
- Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol. 1983;14:931–968
- Villous dysplasia (An ominous histologic sign in colitic patients). Dis Colon Rectum. 1984;27:283–287
- . Premalignant and early malignant lesions in the gastrointestinal tract: definitions, terminology, and problems. Am J Gastroenterol. 1996;91:864–872
- . The world isn't flat. Gastrointest Endosc. 2004;60:426–427
- . Adenomas and adenoma-like DALMs in chronic ulcerative colitis: a clinical, pathological, and molecular review. Am J Gastroenterol. 1999;94:1746–1750
- . Inflammatory bowel disease. Semin Gastrointest Dis. 1996;7:94–104
- . Polypoid dysplasia and adenomas in inflammatory bowel disease: a clinical, pathologic, and follow-up study of 89 polyps from 59 patients. Am J Surg Pathol. 1998;22:275–284
- Macroscopic lesions in dysplasia and carcinoma complicating ulcerative colitis. Dig Dis Sci. 1983;28:18–26
- Surveillance for colonic carcinoma in ulcerative colitis. Gastroenterology. 1985;89:1342–1346
- The Vienna classification of gastrointestinal epithelial neoplasia. Gut. 2000;47:251–255
- Differences in the diagnostic criteria used by Japanese and Western pathologists to diagnose colorectal carcinoma. Cancer. 1998;82:60–69
- Villous, hypermucinous mucosa in long standing ulcerative colitis shows high frequency of K-ras mutations. Gut. 1999;45:686–692
- Interobserver variability in the diagnosis of ulcerative colitis-associated dysplasia by telepathology. Mod Pathol. 2002;15:379–386
- Inter-observer variation between general and specialist gastrointestinal pathologists when grading dysplasia in ulcerative colitis. J Pathol. 2001;194:152–157
- Observer study of the grading of dysplasia in ulcerative colitis: comparison with clinical outcome. Hum Pathol. 1989;20:1008–1014
- Observer variation in the assessment of dysplasia in ulcerative colitis. Histopathology. 1988;13:385–397
- The pattern of cell proliferation in neoplastic and nonneoplastic lesions of ulcerative colitis. Cancer. 1996;78:2307–2312
- Immunohistochemical assessment of Ki67 and p53 expression assists the diagnosis and grading of ulcerative colitis-related dysplasia. Histopathology. 2000;37:108–114
- Ki-67: a useful marker for the evaluation of dysplasia in ulcerative colitis. Mol Pathol. 1998;51:327–332
- . p53 immunohistochemistry of ulcerative colitis-associated with dysplasia and carcinoma. Pathol Int. 1999;49:858–868
- Glut-1 expression in dysplastic and regenerative lesions of the colon. Int J Mol Med. 2001;7:615–619
- A unique basal pattern of p53 expression in ulcerative colitis is associated with mutation in the p53 gene. Histopathology. 2001;39:482–492
- Can histopathologists reliably assess dysplasia in chronic inflammatory bowel disease?. J Clin Pathol. 1987;40:1282–1286
- High proliferative activity is associated with dysplasia in ulcerative colitis. Dis Colon Rectum. 2000;43(10 Suppl):S34–S39
- Sucrase-isomaltase expression in chronic ulcerative colitis and dysplasia. Hum Pathol. 1992;23:774–779
- p53 protein expression in ulcerative colitis-associated colorectal dysplasia and carcinoma. Hum Pathol. 1994;25:1069–1074
- Evaluation of the usefulness of testing for p53 mutations in colorectal cancer surveillance for ulcerative colitis. Am J Gastroenterol. 1999;94:456–462
- . AMACR immunostaining is useful in detecting dysplastic epithelium in Barrett's esophagus, ulcerative colitis, and Crohn's disease. Am J Surg Pathol. 2006;30:871–877
- Sialosyl-Tn antigen is prevalent and precedes dysplasia in ulcerative colitis: a retrospective case-control study. Gastroenterology. 1996;110:694–704
- Sialyl-Tn antigen as a marker of colon cancer risk in ulcerative colitis: relation to dysplasia and DNA aneuploidy. Gastroenterology. 1998;115:1395–1404
- Villous and serrated adenomatous growth bordering carcinomas in inflammatory bowel disease. Anticancer Res. 2000;20:4761–4764
- . Intestinal low-grade tubuloglandular adenocarcinoma in inflammatory bowel disease. Am J Surg Pathol. 2006;30:1022–1029
- Hyperplastic/serrated polyposis in inflammatory bowel disease: a case series of a previously undescribed entity. Am J Surg Pathol. 2008;32:296–303
- Factors affecting the outcome of endoscopic surveillance for cancer in ulcerative colitis. Gastroenterology. 1994;107:934–944
- Two distinct groups of colorectal cancer in inflammatory bowel disease. Inflamm Bowel Dis. 2009;15:9–16
- . Ulcerative colitis and colonic cancer (Problems in assessing the diagnostic usefulness of mucosal dysplasia). Dis Colon Rectum. 1985;28:383–388
- Twenty years' colonoscopic surveillance of patients with ulcerative colitis (Detection of dysplastic and malignant transformation). Scand J Gastroenterol. 1996;31:1195–1204
- Progression to colorectal neoplasia in ulcerative colitis: effect of 5-aminosalicylic acid. Clin Gastroenterol Hepatol. 2008;6:1225–1230
- . Cancer surveillance in ulcerative colitis. Gastroenterology. 1991;100:1241–1248
- Low-grade dysplasia in extensive, long-standing inflammatory bowel disease: a follow-up study. Dis Colon Rectum. 2002;45:615–620
- Colorectal neoplasia in Crohn's colitis: a retrospective comparative study with ulcerative colitis. Histopathology. 2007;50:574–583
- The fate of low grade dysplasia in ulcerative colitis. Am J Gastroenterol. 2002;97:922–927
- Long-term follow-up after polypectomy treatment for adenoma-like dysplastic lesions in ulcerative colitis. Clin Gastroenterol Hepatol. 2004;2:534–541
- Colonoscopic polypectomy in chronic colitis: conservative management after endoscopic resection of dysplastic polyps. Gastroenterology. 1999;117:1295–1300
- . Polypectomy may be adequate treatment for adenoma-like dysplastic lesions in chronic ulcerative colitis. Gastroenterology. 1999;117:1288–1294discussion 1488–1491
- Are dysplasia and colorectal cancer endoscopically visible in patients with ulcerative colitis?. Gastrointest Endosc. 2007;65:998–1004
- Endoscopic factors in the diagnosis of colorectal dysplasia in chronic inflammatory bowel disease. Inflamm Bowel Dis. 2005;11:428–434
- Is dysplasia visible during surveillance colonoscopy in patients with ulcerative colitis?. Scand J Gastroenterol. 2008;43:698–703
- . Differential diagnosis of adenomas and dysplastic lesions in patients with ulcerative colitis. Z Gastroenterol. 1993;31:653–656
- Differential diagnosis of dysplasia-associated lesion or mass and coincidental adenoma in ulcerative colitis. Dis Colon Rectum. 1998;41:322–327
- Distinction between dysplasia-associated lesion or mass (DALM) and adenoma in patients with ulcerative colitis. Hum Pathol. 2000;31:288–291
- Genetic alterations in chronic ulcerative colitis-associated adenoma-like DALMs are similar to non-colitic sporadic adenomas. Am J Surg Pathol. 2000;24:1209–1216
- Loss of heterozygosity of the von Hippel Lindau gene locus in polypoid dysplasia but not flat dysplasia in ulcerative colitis or sporadic adenomas. Hum Pathol. 1998;29:961–964
- Comparison of genetic alterations in colonic adenoma and ulcerative colitis-associated dysplasia and carcinoma. Hum Pathol. 1998;29:131–136
- The differentiation of true adenomas from colitis-associated dysplasia in ulcerative colitis: a comparative immunohistochemical study. Hum Pathol. 1999;30:898–905
- P53 and beta catenin expression in chronic ulcerative colitis–associated polypoid dysplasia and sporadic adenomas: an immunohistochemical study. Am J Surg Pathol. 1999;23:963–969
- Decreased expression of CD44, alpha-catenin, and deleted colon carcinoma and altered expression of beta-catenin in ulcerative colitis-associated dysplasia and carcinoma, as compared with sporadic colon neoplasms. Cancer. 2000;89:733–740
- Artificial neural networks distinguish among subtypes of neoplastic colorectal lesions. Gastroenterology. 2002;122:606–613
- . Sporadic adenoma in ulcerative colitis: endoscopic resection is an adequate treatment. Gut. 2006;55:1151–1155
- . Management of neoplastic polyps in inflammatory bowel disease. Inflamm Bowel Dis. 2003;9:260–266
- Endoscopic mucosal resection for flat neoplasia in chronic ulcerative colitis: can we change the endoscopic management paradigm?. Gut. 2007;56:838–846
- Earlier surveillance colonoscopy programme improves survival in patients with ulcerative colitis associated colorectal cancer: results of a 23-year surveillance programme in the Japanese population. Br J Cancer. 2003;89:1232–1236
- Meta-analysis: cancer risk of low-grade dysplasia in chronic ulcerative colitis. Aliment Pharmacol Ther. 2007;25:657–668
- A prospective study of low-grade dysplasia in ulcerative colitis using an extensive biopsy surveillance protocol (abstr). Gastroenterology. 2005;128:A566
- Ten year follow up of ulcerative colitis patients with and without low grade dysplasia. Gut. 2003;52:1127–1132
- . Consensus conference: colorectal cancer screening and surveillance in inflammatory bowel disease. Inflamm Bowel Dis. 2005;11:314–321
- Outcomes of obstetric hospitalizations among women with inflammatory bowel disease in the United States. Clin Gastroenterol Hepatol. 2009;7:329–334
- Colorectal cancer surveillance in patients with ulcerative colitis. Br J Surg. 1994;81:689–691
- Is colonoscopic surveillance reducing colorectal cancer mortality in ulcerative colitis? (A population based case control study). Gut. 1998;42:711–714
- Colonoscopic surveillance reduces mortality from colorectal cancer in ulcerative colitis. Gastroenterology. 1993;105:418–424
- . Strategies for detecting colon cancer and/or dysplasia in patients with inflammatory bowel disease. Cochrane Database Syst Rev. 2004;(2):CD000279
- Screening and surveillance colonoscopy in chronic Crohn's colitis: results of a surveillance program spanning 25 years. Clin Gastroenterol Hepatol. 2008;6:993–998quiz 953–954
- . Ulcerative colitis practice guidelines in adults (update): American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol. 2004;99:1371–1385
- . Diagnosis and management of dysplasia in patients with ulcerative colitis and Crohn's disease of the colon. Inflamm Bowel Dis. 2009;15:630–638
- Colorectal cancer screening and surveillance: clinical guidelines and rationale—Update based on new evidence. Gastroenterology. 2003;124:544–560
- DNA aneuploidy in colonic biopsies predicts future development of dysplasia in ulcerative colitis. Gastroenterology. 1992;103:1611–1620
- Physicians' perceptions of dysplasia and approaches to surveillance colonoscopy in ulcerative colitis. Am J Gastroenterol. 1995;90:2106–2114
- . How gastroenterologists screen for colonic cancer in ulcerative colitis: an analysis of performance. Gastrointest Endosc. 2000;51:123–128
- Malignant colorectal strictures in Crohn's disease. Am J Gastroenterol. 1991;86:882–885
- . Guidelines for screening and surveillance of asymptomatic colorectal cancer in patients with inflammatory bowel disease. Gut. 2002;51(Suppl 5):V10–V12
- . Predominance of rectosigmoid neoplasia in ulcerative colitis and its implication on cancer surveillance. Gastroenterology. 1993;104:666–667
- Clinicopathological characteristics of colorectal carcinoma complicating ulcerative colitis. Gut. 1994;35:1419–1423
- . Guidelines for the management of inflammatory bowel disease in adults. Gut. 2004;53(Suppl 5):V1–V16
- . Advances in IBD (Evolving protocols in colorectal cancer surveillance). Gastroenterology and Hepatology. 2008;4:114–116
- Nonpolypoid (flat and depressed) colorectal neoplasms. Gastroenterology. 2006;130:566–576
- Pit pattern in colorectal neoplasia: endoscopic magnifying view. Endoscopy. 2001;33:367–373
- Methylene blue-aided chromoendoscopy for the detection of intraepithelial neoplasia and colon cancer in ulcerative colitis. Gastroenterology. 2003;124:880–888
- Pancolonic indigo carmine dye spraying for the detection of dysplasia in ulcerative colitis. Gut. 2004;53:256–260
- Indigo carmine-assisted high-magnification chromoscopic colonoscopy for the detection and characterisation of intraepithelial neoplasia in ulcerative colitis: a prospective evaluation. Endoscopy. 2005;37:1186–1192
- Chromoendoscopy-targeted biopsies are superior to standard colonoscopic surveillance for detecting dysplasia in inflammatory bowel disease patients: a prospective endoscopic trial. Am J Gastroenterol. 2008;103:2342–2349
- Narrow-band imaging compared with conventional colonoscopy for the detection of dysplasia in patients with longstanding ulcerative colitis. Endoscopy. 2007;39:216–221
- Autofluorescence imaging colonoscopy for the diagnosis of dysplasia in ulcerative colitis. Inflamm Bowel Dis. 2007;13:640–641
- Fluorescence endoscopy for the detection of low and high grade dysplasia in ulcerative colitis using systemic or local 5-aminolaevulinic acid sensitisation. Gut. 2003;52:1003–1007
- . What new endoscopic imaging modalities will become important in the diagnosis of IBD. Inflamm Bowel Dis. 2008;14(Suppl 2):S172–S176
- Confocal laser endoscopy for diagnosing intraepithelial neoplasias and colorectal cancer in vivo. Gastroenterology. 2004;127:706–713
- Chromoscopy-guided endomicroscopy increases the diagnostic yield of intraepithelial neoplasia in ulcerative colitis. Gastroenterology. 2007;132:874–882
- . Chemoprevention of colorectal neoplasia: the potential for personalized medicine. Gastroenterology. 2008;134:1224–1237
- . Will a 5-ASA a day keep the cancer (and dysplasia) away?. Am J Gastroenterol. 2005;100:1354–1356
- Ursodiol use is associated with lower prevalence of colonic neoplasia in patients with ulcerative colitis and primary sclerosing cholangitis. Ann Intern Med. 2001;134:89–95
- Ursodeoxycholic acid as a chemopreventive agent in patients with ulcerative colitis and primary sclerosing cholangitis. Gastroenterology. 2003;124:889–893
- . The impact of ursodeoxycholic acid on cancer, dysplasia and mortality in ulcerative colitis patients with primary sclerosing cholangitis. Aliment Pharmacol Ther. 2005;22:783–788
- Ursodeoxycholic acid treatment in IBD-patients with colorectal dysplasia and/or DNA-aneuploidy: a prospective, double-blind, randomized controlled pilot study. Anticancer Res. 2004;24:3121–3127
- Phase III trial of ursodeoxycholic acid to prevent colorectal adenoma recurrence. J Natl Cancer Inst. 2005;97:846–853
- Colorectal carcinoma in ulcerative colitis is decreasing in Scandinavian countries. Anticancer Res. 2001;21:2921–2924
- Effect of folate supplementation on the incidence of dysplasia and cancer in chronic ulcerative colitis (A case-control study). Gastroenterology. 1989;97:255–259
- . Risk factors of colorectal cancer in inflammatory bowel disease. Am J Gastroenterol. 1996;91:44–48
- 5-Aminosalicylate use and colorectal cancer risk in inflammatory bowel disease: a large epidemiological study. Gut. 2005;54:1573–1578
- Aminosalicylate therapy in the prevention of dysplasia and colorectal cancer in ulcerative colitis. Clin Gastroenterol Hepatol. 2006;4:1346–1350
- . Proximal colorectal dysplasia or cancer in ulcerative colitis (The impact of primary sclerosing cholangitis and sulfasalazine: results from a 20-year surveillance study). Dis Colon Rectum. 2001;44:77–85
- Does the use of 5-aminosalicylates in inflammatory bowel disease prevent the development of colorectal cancer?. Am J Gastroenterol. 2003;98:2784–2788
- 5-Aminosalicylic acid therapy and the risk of colorectal cancer among patients with inflammatory bowel disease. Inflamm Bowel Dis. 2007;13:367–371
- The effect of folic acid supplementation on the risk for cancer or dysplasia in ulcerative colitis. Gastroenterology. 1997;112:29–32
- . Effect of 5-aminosalicylate use on colorectal cancer and dysplasia risk: a systematic review and metaanalysis of observational studies. Am J Gastroenterol. 2005;100:1345–1353
- . Review article: mechanisms of action of mesalazine in preventing colorectal carcinoma in inflammatory bowel disease. Aliment Pharmacol Ther. 2003;18(Suppl 2):10–14
- Intestinal antiinflammatory effect of 5-aminosalicylic acid is dependent on peroxisome proliferator-activated receptor-gamma. J Exp Med. 2005;201:1205–1215
- Mesalazine-induced apoptosis of colorectal cancer: on the verge of a new chemopreventive era?. Aliment Pharmacol Ther. 1999;13:1397–1402
- Mesalazine improves replication fidelity in cultured colorectal cells. Cancer Res. 2005;65:3993–3997
- Chemoprevention of colorectal neoplasia in ulcerative colitis: the effect of 6-mercaptopurine. Clin Gastroenterol Hepatol. 2005;3:1015–1021
- Long-term neoplasia risk after azathioprine treatment in inflammatory bowel disease. Lancet. 1994;343:1249–1252
- Long-term risk of malignancy after treatment of inflammatory bowel disease with azathioprine. Aliment Pharmacol Ther. 2002;16:1225–1232
- Malignant neoplasms subsequent to treatment of inflammatory bowel disease with 6-mercaptopurine. Am J Gastroenterol. 1999;94:3248–3253
- Folate intake and carcinogenesis of the colon and rectum. Int J Epidemiol. 1991;20:368–374
- Folate, methionine, and alcohol intake and risk of colorectal adenoma. J Natl Cancer Inst. 1993;85:875–884
- Statins and the risk of colorectal cancer. N Engl J Med. 2005;352:2184–2192
- Microsatellite instability in ulcerative colitis-associated colorectal dysplasias and cancers. Cancer Res. 1994;54:4841–4844
- Genomic instability is an early event during the progression pathway of ulcerative-colitis-related neoplasia. Am J Pathol. 1999;154:1825–1830
- . Molecular biology of dysplasia and cancer in inflammatory bowel disease. Gastroenterol Clin North Am. 2006;35:553–571
- DNA aneuploidy in ulcerative colitis: reproducibility, topographic distribution, and relation to dysplasia. Gastroenterology. 1992;102:1149–1154
- DNA aneuploidy and histologic dysplasia in long-standing ulcerative colitis (A 10-year follow-up study). Dis Colon Rectum. 1994;37:313–319discussion 319–320
- . DNA aneuploidy as a marker of premalignancy in surveillance of patients with ulcerative colitis. Br J Surg. 1999;86:947–950
- . p53 expression in ulcerative colitis: a longitudinal study. Gut. 1995;37:802–804
- Clinical significance of microsatellite instability in the inflamed mucosa for the prediction of colonic neoplasms in patients with ulcerative colitis. J Gastroenterol Hepatol. 2005;20:710–715
- K-ras mutations in sera of patients with colorectal neoplasias and long-standing inflammatory bowel disease. Scand J Gastroenterol. 2002;37:715–718
- Screening for p53 and K-ras mutations in whole-gut lavage in chronic inflammatory bowel disease. Eur J Gastroenterol Hepatol. 2002;14:1061–1066
- Fecal DNA versus fecal occult blood for colorectal-cancer screening in an average-risk population. N Engl J Med. 2004;351:2704–2714
- Improved fecal DNA test for colorectal cancer screening. Clin Gastroenterol Hepatol. 2007;5:111–117
- Molecular profiling of ulcerative colitis-associated neoplastic progression. Exp Mol Pathol. 2006;80:1–10
Reprint requests Address requests for reprints to: Chair, Clinical Practice and Quality Management Committee, AGA National Office, 4930 Del Ray Avenue, Bethesda, Maryland 20814. Phone: (301) 272-1189; e-mail: SAgyeman@gastro.org.
Conflicts of interest The authors disclose the following: Dr Farraye has received research support from Prometheus Laboratories; is a consultant and a member of the speaker's bureau for Abbott, Centocor, Proctor & Gamble, Prometheus Laboratories, Salix, and Shire; and is a consultant for UCB. The remaining authors disclose no conflicts.
PII: S0016-5085(09)02200-8
doi:10.1053/j.gastro.2009.12.035
© 2010 AGA Institute. Published by Elsevier Inc. All rights reserved.
Refers to article:
- IBD Guide 2009 CME Exam 3 , 23 December 2009