ORIGINAL ARTICLE
Risk for Colorectal Neoplasia in Patients with Inflammatory Bowel Disease and Mucosa Indefinite for Dysplasia Keith K. Lai, MD,* Bela Horvath, MD, PhD,* Hao Xie, PhD,† Xianrui Wu, MD, PhD,‡ Brian L. Lewis, BA,† Rish K. Pai, MD, PhD,* Thomas Plesec, MD,* Deepa T. Patil, MD,* Ilyssa O. Gordon, MD, PhD,* Yinghong Wang, MD, PhD,§ Bo Shen, MD,§ John R. Goldblum, MD,* and Xiuli Liu, MD, PhD*
Background: The management of colonic epithelial changes indefinite for dysplasia (IND) in patients with inflammatory bowel disease (IBD) remains controversial because of a paucity of published outcome data.
Methods: We analyzed data from 93 patients with IBD who were IND and 52 IBD patients without dysplasia (controls) from the Department of Anatomic Pathology database at the Cleveland Clinic from 1989 to 2004. Pathology reports, histologic slides, clinical features, and outcomes were reviewed.
Results: Twenty-two patients (23.7%) had surgical resections within 6 months of the IND assignment; of these, 6 had dysplasia (27.3%; 1 low-grade dysplasia and 5 high-grade dysplasia [HGD]). The remaining 71 patients received regular colonoscopy examinations for a mean period of 98.6 months; 18 patients developed dysplasia or carcinoma (25.2%; 10 low-grade dysplasia, 5 HGD, and 3 colorectal cancer). There was a mean interval of 53.9 months between an IND assignment and identification of dysplasia or carcinoma. Histology review of 59 cases revealed 3.2 cases per 100 person-years for neoplasia (low-grade dysplasia, HGD, or colorectal cancer) and 1.5 cases per 100 person-years for advanced neoplasia (HGD or colorectal cancer); these values were higher than those for controls (1.9 cases per 100 person-years for neoplasia and 0.7 cases per 100 person-years for advance neoplasia; P ¼ 0.1 and P ¼ 0.2, respectively, for IND versus controls). Patients aged more than 44 years when they were found to be IND were more likely than younger patients to develop neoplasia (hazard ratio, 6.7; P ¼ 0.01). Conclusions: Patients with IBD and IND are at significant risk for colorectal dysplasia and cancer. These patients should be closely followed. (Inflamm Bowel Dis 2015;21:378–384) Key Words: Crohn’s disease, ulcerative colitis, colon cancer risk
I
t has been well established that idiopathic inflammatory bowel disease (IBD), both ulcerative colitis (UC) and Crohn’s disease, has an increased risk for developing colorectal cancer (CRC).1–4 CRC causes death in 15% of patients with IBD.5 Current management has focused on the detection of precarcinomatous change (dysplasia) in the colorectal epithelium as a marker for increased cancer risk and, thus, as a potential indicator for colectomy in patients with IBD. In the landmark article by Riddell et al,6 a standardized classification of IBD-associated dysplasia was proposed for the evaluation of biopsy specimens from patients with IBD. The pathology report should clearly state the presence or absence of dysplasia. Although most cases can be readily classified as either Received for publication September 4, 2014; Accepted October 21, 2014. From the *Department of Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio; †Cleveland Clinic Lerner College of Medicine, Case Western Reserve University, Cleveland, Ohio; and Departments of ‡Colorectal Surgery, and §Gastroenterology, Digestive Disease Institute, Cleveland Clinic, Cleveland, Ohio. The authors have no conflicts of interest to disclose. Reprints: Xiuli Liu, MD, PhD, Department of Anatomic Pathology, Cleveland Clinic, 9500 Euclid Avenue, L25, Cleveland, OH 44195 (e-mail: [email protected]). Copyright © 2015 Crohn’s & Colitis Foundation of America, Inc. DOI 10.1097/MIB.0000000000000286 Published online 7 January 2015.
378
| www.ibdjournal.org
negative for dysplasia or positive for dysplasia, the authors acknowledge a category of “mucosa indefinite for dysplasia (IND)” defined as “epithelial changes impossible to be classified as either unequivocally negative or positive for dysplasia.” However, specific diagnostic criteria for IND have not been clearly defined. Few studies have investigated the progression rate of low-grade dysplasia (LGD) in UC.7,8 These studies reported that the majority (81%) of UC patients with LGD did not progress7 and that a low rate of progression to high-grade dysplasia (HGD) or CRC is identified in patients undergoing surveillance for LGD and IND.8 Furthermore, the progression rate of LGD and IND has been reported to be dependent upon diagnostic accuracy as misclassification of dysplasia in patients with IBD significantly affects the progression rates to advanced neoplasia.9 Because of the lack of clearly defined criteria for IND, poor interobserver agreement for this diagnosis, and a paucity of published outcome data, the management of IND in patients with IBD is controversial. The aim of this study was to examine the clinical outcome of patients with IBD after a diagnosis of IND and to identify factors associated with prevalent and incident neoplasia. Inflamm Bowel Dis Volume 21, Number 2, February 2015
Inflamm Bowel Dis Volume 21, Number 2, February 2015
MATERIAL AND METHODS Study Population The pathology database in the Department of Anatomic Pathology at Cleveland Clinic was searched for IBD, UC, Crohn’s disease, colitis, and IND for the years 1989–2004. Patients who had dysplasia diagnosed previously or synchronously at the time of IND diagnosis, had incomplete or missing medical records, or were lost to follow-up (i.e., never returned to have clinical visits or hospital admission after the diagnosis of IND) were excluded. In addition, the pathology database was also searched for IBD, UC, Crohn’s disease, colitis, and negative for dysplasia for the same time period to obtain a cohort of patients with IBD who were negative for dysplasia (IBD NEG). From this cohort, 52 patients were randomly selected and included as a control group.
Demographic and Clinical Variables Data were collected on demographic variables, including age, gender, IBD subtype, duration of disease, extent of disease, presence or absence of primary sclerosing cholangitis (PSC), and clinical indication for colectomy by chart review. Data on clinical indication for colectomy were also collected for patients who underwent resection within 6 months (resection group).
Outcome Measurement The primary endpoints of this study were prevalent neoplasia (defined as the presence of LGD, HGD, or CRC in the colectomy specimen resected within 6 mo after IND diagnosis), incident neoplasia (defined as the presence of LGD, HGD, or CRC in the surveillance colonoscopy biopsy or at colectomy .6 mo), and incidence of advanced neoplasia (defined as the presence of HGD or CRC in a surveillance colonoscopy biopsy or at colectomy .6 mo).
Histologic Interpretation and Pathology Review The initial interpretation of IND was used in this study for the neoplasia progression analysis, mainly to reflect the real-time status of patients as understood by their gastroenterologists as part of routine practice. The experience of the original interpreting pathologist in the field of gastrointestinal (GI) pathology (i.e., GI pathologist versus non–GI pathologist) was recorded. A total of 125 IBD patients with a diagnosis of IND on biopsy were identified. Thirty-two patients were excluded because of a previous diagnosis of neoplasia or being lost to follow-up. A total of 93 patients with an index diagnosis of IND were included in this study, 77 of which had slides available for review. All biopsy slides on which the diagnosis of IND was made in these 77 patients were rereviewed by 5 GI pathologists according to the criteria described previously,6 but histologic slides containing prevalent and incident neoplasia in the study group and all slides from the control group were rereviewed by 1 GI pathologist (X.L.).
Outcome of IND in IBD
Statistical Analysis Patients with an index diagnosis of IND were included in this study. The study patients were divided into 2 subgroups: 1 group that underwent colectomy within 6 months and the other was followed by surveillance colonoscopy with biopsy. The prevalent neoplasia rate was obtained in the first subgroup, whereas the progression rate to any neoplasia and advanced neoplasia was determined from the subgroup of patients with IND who were followed by surveillance. Time to progression was measured in months. Patients who did not develop LGD, HGD, or CRC during follow-up were censored at the time of the last colonoscopy or colectomy. Progression probability was estimated using Kaplan-Meier method and compared using log-rank test. Univariate analysis was performed using Cox proportional hazards regression. Categorical variables were summarized as count and proportion and compared using Fisher’s exact test. Continuous variables were summarized as median and interquartile range and compared using Mann–Whitney U test. A P value of ,0.05 was considered statistically significant. Statistical analysis was performed using R 2.15.2 (R Development Core Team (2012), Vienna, Austria).
RESULTS The flowchart of study patients with IBD with a diagnosis of IND on colonic biopsy is shown in Figure 1. A total of 125 IBD patients with a diagnosis of IND on biopsy were identified. Thirty-two patients were excluded because of a previous diagnosis of neoplasia or being lost to follow-up. A total of 93 patients with an index diagnosis of IND were included in this study, 77 of which had slides available for review and thus were used for the calculation of prevalent and incident neoplasia. The demographic and clinical features of these 77 patients are shown in Table 1.
Prevalent Dysplasia in IBD Patients with IND Twenty-two of the 93 patients with IND (23.7%) had a surgical resection within 6 months of the IND diagnosis; 6 (27.3%) had dysplasia (1 LGD and 5 HGD) in their colectomy specimens. Among these 22 patients with IND who had a surgical resection within 6 months, 18 cases (of 22, 81.8%) had slides and clinical information for review. Refractory colitis and clinical impression of dysplasia were listed as colectomy indication in 11 and 4 cases, respectively. The colectomy indication in the remaining 3 cases was unknown. Review of the resection specimens confirmed the presence of HGD in 3 cases and LGD in 1 case; the prevalence of neoplasia in patients with an index diagnosis of IND was 22.2%. Univariate analysis revealed that a clinical impression of dysplasia by the surgeon as based on transcription from the surgical pathology report was associated with the presence of neoplasia in the colectomy specimen (Table 2). Other features such as age, gender, IBD type, duration of disease, extent of www.ibdjournal.org |
379
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Lai et al
FIGURE 1. Flowchart of study patients with IBD and colonic epithelial changes IND.
colitis, and experience of interpreting pathologist (GI versus nonGI) were not associated with prevalent neoplasia in the colectomy specimen.
Incident Neoplasia in IBD Patients with IND The remaining 71 patients with IBD were followed at this hospital with surveillance colonoscopy for a mean duration of 98.6 months (SD, 65.0; range, 1.7–250.5), and 90% of them had follow-up longer than 1 year. The average surveillance interval was 2.5 years, and 58% of patients had their repeat colonoscopy within 1 year of the diagnosis of IND. Of these, 18 patients (25.4%) developed dysplasia or carcinoma (10 LGD, 5 HGD, and 3 CRC) with a mean interval period of 53.9 months (SD, 36.9) between IND and dysplasia or carcinoma diagnoses. Those who progressed were significantly older than nonprogressors (P ¼ 0.006) (Table 3). A diagnosis of IND made by a GI pathologist was more than twice as likely to progress than the one made by a non–GI pathologist (34.3% versus 13.9%; P ¼ 0.04) (Table 3). A total of 59 cases from patients with IBD and IND (of 71, 83.1%) followed by colonoscopy had slides for review. None of the IND cases were collectively reclassified as positive for dysplasia after histology review by 5 GI pathologists.
380
| www.ibdjournal.org
Reexamination of the follow-up biopsy and/or resection specimens carrying a dysplasia diagnosis confirmed that 7 had LGD (time since IND diagnosis: 21–123 mo), 3 had HGD (time since IND diagnosis: 12–25 mo), and 3 had CRC (time since IND diagnosis: 34–145 mo). The Kaplan-Meier curve of progression to any neoplasia and advanced neoplasia in IBD patients with IND is shown in Figure 2. A total of 52 cases with an index diagnosis of negative for dysplasia were included as a control group. The clinical and pathological characteristics in these 2 groups were similar (Table 4). In the control group, 5 cases progressed to LGD (time since negative for dysplasia diagnosis: 115–182 mo), 1 to HGD (time since negative for dysplasia diagnosis: 15 mo), and 2 to CRC (time since negative for dysplasia diagnosis: 44 and 168 mo). The calculated progression rate (to LGD, HGD, and/or CRC) is 3.2 cases per 100 person-years at risk, and the progression rate (to advanced neoplasia [HGD and/ or CRC]) is 1.5 cases per 100 person-years at risk in the IND group. These progression rates were higher than those in the control group (1.9 and 0.7 cases per 100 person-years at risk; P ¼ 0.1 and 0.2, respectively). Univariate analysis identified that age .44 years (median age) at the time of diagnosis of index IND was associated with a greater probability of progression to neoplasia (hazard ratio, 6.7;
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Outcome of IND in IBD
TABLE 1. Demographics and Clinical Features of Patients with Slides Reviewed in This Study (N ¼ 77)
Mean age (range), yr From IND diagnosis to resection, d Median duration of colitis (range), yr No. colonoscopies No. biopsies Time of follow-up, y Gender, male (%) Follow-up results, n (%) Negative LGD HGD Cancer Original pathologist, n (%) Non-GI GI IBD subtype, n (%) CD UC Othersa Extent of colitis, n (%) Pancolitis Left-sided Rectum only Others PSC, n (%) Absent Present Unknown Indication for colectomy if resection within 6 mo, n (%) Medical refractive IBD IBD with dysplasia Unknown
Resection Group (n ¼ 18)
Colonoscopy Group (n ¼ 59)
35.5 (31.0–52.5) 33.5 (6.5–60.2)
44 (34.5–56.0) N/A
6.0 (2.0–17.8)
7.0 (1.0–11.8)
N/A N/A N/A 12 (64)
3.5 11 6.8 34
(1.2–7.0) (7.0–28.0) (3.1–9.6) (58)
14 0 4 0
46 7 3 3
(78) (12) (5) (5)
(78) (0) (22) (0)
8 (44) 10 (56)
30 (51) 29 (49)
3 (17) 15 (83) 0 (0)
6 (10) 51 (86) 2 (3)
6 5 6 1
(33) (28) (33) (6)
15 (83) 0 (0) 3 (17)
11 (61) 4 (22) 3 (17)
29 15 13 2
(49) (25) (22) (3.4)
51 (86) 8 (14) 0 (0)
N/A
a
Cases with a presumable diagnosis of UC found to have evolved into typical Crohn’s disease after colectomy. CD, Crohn’s disease; N/A, not applicable.
95% confidence interval [CI], 1.5–33.3; P ¼ 0.01) (Table 5, Fig. 2C). In addition, the frequency of surveillance colonoscopy (defined as number of colonoscopies per year) was associated with progression (hazard ratio, 2.5 for 1 more per yr; 95% CI, 1.6–3.9; P , 0.0001) (Table 5). IBD duration, the presence of PSC, the extent of disease, the number of colonoscopy procedures, or the number of biopsies was not associated with neoplasia progression (Table 5).
TABLE 2. Prevalent Neoplasia in IBD Patients with an Index Diagnosis of IND
Mean age (range), yr Gender, male (%) IBD type, n (%) UC CD Median duration of colitis (range), yr Extent of colitis, n (%) Pancolitis Left-sided Rectum only Others Interpreting pathologist, n (%) GI Non–GI pathologist Dysplasia as indication for colectomy,a n (%)
Prevalent Neoplasia (n ¼ 4)
No Prevalent Neoplasia (n ¼ 14)
P
41.5 (15–56) 3 (75)
35.5 (20–73) 9 (64.3)
0.7 1
4 (100) 0 (0) 16 (5–22)
11 (78.5) 3 (21.5) 2 (0–24)
1 (25) 1 (25) 2 (50) 0 (0)
5 4 4 1
(35.7) (28.6) (28.6) (7.1)
3 (75) 1 (25) 3 (75)
7 (50) 7 (50) 1 (9.0)
1 0.1
1
0.6 0.03
a Only 15 cases had this information available. CD, Crohn’s disease.
DISCUSSION It is well established that IBD has an increased risk for developing CRC.1–4 Detection of dysplasia by surveillance colonoscopy and biopsy is the standard of care to minimize mortality due to CRC in patients with IBD, but diagnostic difficulty and uncertainty remain and may lead to a diagnosis of IND in some cases. Little is known of the natural history of IBD–IND, and as such the management of IND in patients with IBD is controversial. We studied 2 subgroups of patients with IND; 1 group underwent colectomy within 6 months and the other was followed by surveillance colonoscopy. The first subgroup allowed us to estimate the prevalence of neoplasia associated with an index diagnosis of IND to be at least 22.2%. A clinical impression of dysplasia by the surgeon was associated with prevalent neoplasia. No other features (age, gender, IBD subtype, duration of disease, extent of colitis, or the presence of PSC) predicted prevalent dysplasia. Progression in a dysplasia–carcinoma sequence from LGD to HGD and, ultimately, to invasive adenocarcinoma in IBD is well known. However, recent data also suggest that invasive adenocarcinoma may arise from LGD.10 Therefore, in this study, neoplastic progression is defined as LGD, HGD, and CRC, in contrast to other studies that defined progression as the development of HGD or CRC. In this study, the neoplasia incidence rate www.ibdjournal.org |
381
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Lai et al
TABLE 3. Clinical Outcome of IBD patients with a Diagnosis of IND Followed with Colonoscopy
Worst neoplasiaa LGD HGD CRC Age, mean (SD), yr Gender (M/F) Extent of colitis Pancolitis Left-sided Rectum only Others Interpreting pathologist GI Non-GI
IND Progressor (n ¼ 17)
IND Nonprogressor (n ¼ 54)
10 5 2 54.4 (12.6) 13/4
0 0 0 44.1 (13.3) 30/24
9 5 2 1
24 15 13 2
0.1
12 5
23 31
0.04
P
N/A
0.006 0.2
a Worst neoplasia diagnosed in a follow-up specimen. N/A, not applicable.
in patients with index IND is 3.2 cases per 100 person-years, higher than the progression rate of 1.9 cases per 100 personyears in the control group, although the difference is not statistically significant. The progression rate to advanced neoplasia in patients with index IND is 1.5 cases per 100 person-years at risk, approaching the reported progression rate to advanced neoplasia in LGD established in 1 study,8 but is lower than the reported 3 cases per 100 person-years in a meta-analysis.11 However, this progression rate to advanced neoplasia in patients with IND is higher than the progression rate of 0.7 cases per 100 person-years seen in the control group, although the difference again does not reach statistical significance. These results suggest that some cases of IND may represent a transition stage of neoplasia between negative and LGD with an intermediate progression risk between negative for dysplasia and LGD. However, large and prospective studies are needed to further verify or refute this concept. Our finding of a lack of association of patient gender, IBD subtype, duration of disease, extent of colitis, presence of PSC, number of colonoscopy procedures, and number of biopsies is consistent with that reported in a previous study of LGD and IND.8 However, our study identified that age .44 years at diagnosis of index IND was associated with neoplasia progression. This finding differs from a previous report in which age was not associated with neoplasia progression8; however, their cohort was a mix of patients with LGD and IND (predominantly LGD) and small in size (N ¼ 35), with progression defined as HGD and/or CRC. A potential explanation for the observed association of age
382
| www.ibdjournal.org
FIGURE 2. A, Kaplan-Meier curve of progression to neoplasia (LGD, HGD, or cancer) and (B) advanced neoplasia (HGD or cancer) in patients with IND, compared with control group. C, Age at the time of IND diagnosis as a risk factor for neoplasia progression. Univariate analysis revealed the cutoff age of 44 years as having significant probability of progression (hazard ratio, 6.7; 95% CI, 1.5–33.3; P ¼ 0.01).
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Outcome of IND in IBD
TABLE 4. Clinical and Pathological Characteristics of Negative and IND Groups Negative (n ¼ 52) Mean age (SD), yr Median duration of follow-up (range), yr Median duration of colitis (range), yr Gender, male (%) Disease type, n (%) UC CD Othersa Disease extent, n (%) Pancolitis Left-sided Rectum only Other PSC, present (%)
IND (n ¼ 59)
P
42.5 (31–53.2)a 44 (34.5–56) 0.3 5.8 (2.7–13.3) 6.8 (3.1–9.7) 0.3 2.2 (0.3–9.2)
7 (1.0–11.8) 0.1
25 (48)
34 (58)
0.3
39 (75) 11 (21) 2 (4)
51 (86) 6 (10) 2 (3)
0.3
25 14 8 5 6
29 15 13 2 8
0.5
(48) (27) (15) (10) (12)
(49) (25) (22) (3) (14)
TABLE 5. Univariate Analysis of Features in Relation to Progression to Definite Dysplasia in Patients with an Index IND Diagnosis
1
Age (.44 versus #44), yr IBD duration (#7 versus .7), yr PSC (present versus absent) Interpreting pathologist (GI versus non-GI) No. colonoscopies per 1 more Frequency of surveillance per 1 more per year No. biopsies per 1 more Extent of disease (pancolitis as reference) Left-sided Rectum only Others
Hazard Ratio
95% CI
P
6.7 1.7 2.2 1.69
1.5–33.3 0.56–5.3 0.59–7.93 0.55–5.21
0.01 0.3 0.2 0.4
1.02 2.5
0.89–1.17 1.6–3.9
0.8 ,0.0001
1.3
0.96–1.8
0.09
0.66 0.21 1.54
0.17–2.51 0.03–1.72 0.19–12.51
0.5 0.1 0.7
CI, confidence interval.
a
Cases with a presumable diagnosis of UC found to have evolved into typical Crohn’s disease after colectomy. CD, Crohn’s disease.
with neoplasia progression in our study is that IND may have existed in the large bowel for a longer duration in the older patients than in the younger patients. Another possibility is that the development of neoplasia is accelerated in older patients. In addition, we identified that frequency of surveillance colonoscopy was associated with progression; however, it is unclear whether this reflects increased disease activity in these patients or possibly detection bias. There are limitations to this study. Our data are limited by the quality and availability of the clinical information found within the medical records. For example, many cases did not have detailed endoscopic reports, preventing us from determining whether the IND biopsy was from a lesion or from flat mucosa and from correlating the location of index IND with the location of prevalent or incident neoplasia. In addition, many cases did not have detailed history, such as a family history of CRC and medication history. The rate of prevalent neoplasia in our study may represent an overestimation because of the referral bias, and we could not completely rule out an outside diagnosis of dysplasia before the IND diagnosis at our institute. Similarly, the study population was based on a highly specialized tertiary care center where GI pathology subspecialty practice started in 2002, 2 years before the end of the study period, and conclusions from this study may not be applicable to other practice settings. In summary, our study reaffirms the value of IND as a diagnostic category for IBD colonic biopsies. The diagnosis
of IND carries significant prevalent and incident rates for neoplasia in patients with IBD, and clinical impression should be taken into consideration in managing these patients because of the significant rate of prevalent neoplasia. IBD patients with IND should be followed by repeat colonoscopy after a short interval (within months) and then managed based on the finding of the repeat biopsy. Our study also indicates that the adherence of IBD patients with mucosal changes IND to surveillance needs to be improved. Large studies are needed to confirm our findings.
REFERENCES 1. Rosenqvist H, Lagercrantz R, Edling N. Ulcerative colitis and carcinoma coli. Lancet. 1959;1:906–908. 2. Ekbom A, Helmick C, Zack M, et al. Increased risk of large-bowel cancer in Crohn’s disease with colonic involvement. Lancet. 1990; 336:357–359. 3. Jess T, Winther KV, Munkholm P, et al. 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. 4. Jess T, Gamborg M, Matzen P, et al. Increased risk of intestinal cancer in Crohn’s disease: a meta-analysis of population-based cohort studies. Am J Gastroenterol. 2005;100:2724–2729. 5. Breynaert C, Vermeire S, Rutgeerts P, et al. Dysplasia and colorectal cancer in inflammatory bowel disease: a result of inflammation or an intrinsic risk? Acta Gastroenterol Belg. 2008;71:367–372. 6. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol. 1983;14:931–968. 7. Zisman TL, Bronner MP, Rulyak S, et al. Prospective study of the progression of low-grade dysplasia in ulcerative colitis during current cancer surveillance guidelines. Inflamm Bowel Dis. 2012;18:2240– 2246. www.ibdjournal.org |
383
Lai et al
8. Pekow JR, Hetzel JT, Rothe JA, et al. Outcome after surveillance of lowgrade and indefinite dysplasia in patients with ulcerative colitis. Inflamm Bowel Dis. 2010;16:1352–1356. 9. van Schaik FD, ten Kate FJ, Offerhaus JA, et al. Misclassification of dysplasia in patients with inflammatory bowel disease: consequences for progression rates to advanced neoplasia. Inflamm Bowel Dis. 2011;17:1108–1116.
384
| www.ibdjournal.org
Inflamm Bowel Dis Volume 21, Number 2, February 2015
10. Harpaz N, Polydorides AD. Colorectal dysplasia in chronic inflammatory bowel disease: pathology, clinical implication, and pathogenesis. Arch Pathol Lab Med. 2010;134:876–895. 11. Thomas T, Abrams KA, Robinson RJ, et al. Meta-analysis: cancer risk of low-grade dysplasia in chronic ulcerative colitis. Aliment Pharmacol Ther. 2006;25:657–668.
Risk for Colorectal Neoplasia in Patients with Inflammatory Bowel Disease and Mucosa Indefinite for Dysplasia Keith K. Lai, MD,* Bela Horvath, MD, PhD,* Hao Xie, PhD,† Xianrui Wu, MD, PhD,‡ Brian L. Lewis, BA,† Rish K. Pai, MD, PhD,* Thomas Plesec, MD,* Deepa T. Patil, MD,* Ilyssa O. Gordon, MD, PhD,* Yinghong Wang, MD, PhD,§ Bo Shen, MD,§ John R. Goldblum, MD,* and Xiuli Liu, MD, PhD*
Background: The management of colonic epithelial changes indefinite for dysplasia (IND) in patients with inflammatory bowel disease (IBD) remains controversial because of a paucity of published outcome data.
Methods: We analyzed data from 93 patients with IBD who were IND and 52 IBD patients without dysplasia (controls) from the Department of Anatomic Pathology database at the Cleveland Clinic from 1989 to 2004. Pathology reports, histologic slides, clinical features, and outcomes were reviewed.
Results: Twenty-two patients (23.7%) had surgical resections within 6 months of the IND assignment; of these, 6 had dysplasia (27.3%; 1 low-grade dysplasia and 5 high-grade dysplasia [HGD]). The remaining 71 patients received regular colonoscopy examinations for a mean period of 98.6 months; 18 patients developed dysplasia or carcinoma (25.2%; 10 low-grade dysplasia, 5 HGD, and 3 colorectal cancer). There was a mean interval of 53.9 months between an IND assignment and identification of dysplasia or carcinoma. Histology review of 59 cases revealed 3.2 cases per 100 person-years for neoplasia (low-grade dysplasia, HGD, or colorectal cancer) and 1.5 cases per 100 person-years for advanced neoplasia (HGD or colorectal cancer); these values were higher than those for controls (1.9 cases per 100 person-years for neoplasia and 0.7 cases per 100 person-years for advance neoplasia; P ¼ 0.1 and P ¼ 0.2, respectively, for IND versus controls). Patients aged more than 44 years when they were found to be IND were more likely than younger patients to develop neoplasia (hazard ratio, 6.7; P ¼ 0.01). Conclusions: Patients with IBD and IND are at significant risk for colorectal dysplasia and cancer. These patients should be closely followed. (Inflamm Bowel Dis 2015;21:378–384) Key Words: Crohn’s disease, ulcerative colitis, colon cancer risk
I
t has been well established that idiopathic inflammatory bowel disease (IBD), both ulcerative colitis (UC) and Crohn’s disease, has an increased risk for developing colorectal cancer (CRC).1–4 CRC causes death in 15% of patients with IBD.5 Current management has focused on the detection of precarcinomatous change (dysplasia) in the colorectal epithelium as a marker for increased cancer risk and, thus, as a potential indicator for colectomy in patients with IBD. In the landmark article by Riddell et al,6 a standardized classification of IBD-associated dysplasia was proposed for the evaluation of biopsy specimens from patients with IBD. The pathology report should clearly state the presence or absence of dysplasia. Although most cases can be readily classified as either Received for publication September 4, 2014; Accepted October 21, 2014. From the *Department of Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio; †Cleveland Clinic Lerner College of Medicine, Case Western Reserve University, Cleveland, Ohio; and Departments of ‡Colorectal Surgery, and §Gastroenterology, Digestive Disease Institute, Cleveland Clinic, Cleveland, Ohio. The authors have no conflicts of interest to disclose. Reprints: Xiuli Liu, MD, PhD, Department of Anatomic Pathology, Cleveland Clinic, 9500 Euclid Avenue, L25, Cleveland, OH 44195 (e-mail: [email protected]). Copyright © 2015 Crohn’s & Colitis Foundation of America, Inc. DOI 10.1097/MIB.0000000000000286 Published online 7 January 2015.
378
| www.ibdjournal.org
negative for dysplasia or positive for dysplasia, the authors acknowledge a category of “mucosa indefinite for dysplasia (IND)” defined as “epithelial changes impossible to be classified as either unequivocally negative or positive for dysplasia.” However, specific diagnostic criteria for IND have not been clearly defined. Few studies have investigated the progression rate of low-grade dysplasia (LGD) in UC.7,8 These studies reported that the majority (81%) of UC patients with LGD did not progress7 and that a low rate of progression to high-grade dysplasia (HGD) or CRC is identified in patients undergoing surveillance for LGD and IND.8 Furthermore, the progression rate of LGD and IND has been reported to be dependent upon diagnostic accuracy as misclassification of dysplasia in patients with IBD significantly affects the progression rates to advanced neoplasia.9 Because of the lack of clearly defined criteria for IND, poor interobserver agreement for this diagnosis, and a paucity of published outcome data, the management of IND in patients with IBD is controversial. The aim of this study was to examine the clinical outcome of patients with IBD after a diagnosis of IND and to identify factors associated with prevalent and incident neoplasia. Inflamm Bowel Dis Volume 21, Number 2, February 2015
Inflamm Bowel Dis Volume 21, Number 2, February 2015
MATERIAL AND METHODS Study Population The pathology database in the Department of Anatomic Pathology at Cleveland Clinic was searched for IBD, UC, Crohn’s disease, colitis, and IND for the years 1989–2004. Patients who had dysplasia diagnosed previously or synchronously at the time of IND diagnosis, had incomplete or missing medical records, or were lost to follow-up (i.e., never returned to have clinical visits or hospital admission after the diagnosis of IND) were excluded. In addition, the pathology database was also searched for IBD, UC, Crohn’s disease, colitis, and negative for dysplasia for the same time period to obtain a cohort of patients with IBD who were negative for dysplasia (IBD NEG). From this cohort, 52 patients were randomly selected and included as a control group.
Demographic and Clinical Variables Data were collected on demographic variables, including age, gender, IBD subtype, duration of disease, extent of disease, presence or absence of primary sclerosing cholangitis (PSC), and clinical indication for colectomy by chart review. Data on clinical indication for colectomy were also collected for patients who underwent resection within 6 months (resection group).
Outcome Measurement The primary endpoints of this study were prevalent neoplasia (defined as the presence of LGD, HGD, or CRC in the colectomy specimen resected within 6 mo after IND diagnosis), incident neoplasia (defined as the presence of LGD, HGD, or CRC in the surveillance colonoscopy biopsy or at colectomy .6 mo), and incidence of advanced neoplasia (defined as the presence of HGD or CRC in a surveillance colonoscopy biopsy or at colectomy .6 mo).
Histologic Interpretation and Pathology Review The initial interpretation of IND was used in this study for the neoplasia progression analysis, mainly to reflect the real-time status of patients as understood by their gastroenterologists as part of routine practice. The experience of the original interpreting pathologist in the field of gastrointestinal (GI) pathology (i.e., GI pathologist versus non–GI pathologist) was recorded. A total of 125 IBD patients with a diagnosis of IND on biopsy were identified. Thirty-two patients were excluded because of a previous diagnosis of neoplasia or being lost to follow-up. A total of 93 patients with an index diagnosis of IND were included in this study, 77 of which had slides available for review. All biopsy slides on which the diagnosis of IND was made in these 77 patients were rereviewed by 5 GI pathologists according to the criteria described previously,6 but histologic slides containing prevalent and incident neoplasia in the study group and all slides from the control group were rereviewed by 1 GI pathologist (X.L.).
Outcome of IND in IBD
Statistical Analysis Patients with an index diagnosis of IND were included in this study. The study patients were divided into 2 subgroups: 1 group that underwent colectomy within 6 months and the other was followed by surveillance colonoscopy with biopsy. The prevalent neoplasia rate was obtained in the first subgroup, whereas the progression rate to any neoplasia and advanced neoplasia was determined from the subgroup of patients with IND who were followed by surveillance. Time to progression was measured in months. Patients who did not develop LGD, HGD, or CRC during follow-up were censored at the time of the last colonoscopy or colectomy. Progression probability was estimated using Kaplan-Meier method and compared using log-rank test. Univariate analysis was performed using Cox proportional hazards regression. Categorical variables were summarized as count and proportion and compared using Fisher’s exact test. Continuous variables were summarized as median and interquartile range and compared using Mann–Whitney U test. A P value of ,0.05 was considered statistically significant. Statistical analysis was performed using R 2.15.2 (R Development Core Team (2012), Vienna, Austria).
RESULTS The flowchart of study patients with IBD with a diagnosis of IND on colonic biopsy is shown in Figure 1. A total of 125 IBD patients with a diagnosis of IND on biopsy were identified. Thirty-two patients were excluded because of a previous diagnosis of neoplasia or being lost to follow-up. A total of 93 patients with an index diagnosis of IND were included in this study, 77 of which had slides available for review and thus were used for the calculation of prevalent and incident neoplasia. The demographic and clinical features of these 77 patients are shown in Table 1.
Prevalent Dysplasia in IBD Patients with IND Twenty-two of the 93 patients with IND (23.7%) had a surgical resection within 6 months of the IND diagnosis; 6 (27.3%) had dysplasia (1 LGD and 5 HGD) in their colectomy specimens. Among these 22 patients with IND who had a surgical resection within 6 months, 18 cases (of 22, 81.8%) had slides and clinical information for review. Refractory colitis and clinical impression of dysplasia were listed as colectomy indication in 11 and 4 cases, respectively. The colectomy indication in the remaining 3 cases was unknown. Review of the resection specimens confirmed the presence of HGD in 3 cases and LGD in 1 case; the prevalence of neoplasia in patients with an index diagnosis of IND was 22.2%. Univariate analysis revealed that a clinical impression of dysplasia by the surgeon as based on transcription from the surgical pathology report was associated with the presence of neoplasia in the colectomy specimen (Table 2). Other features such as age, gender, IBD type, duration of disease, extent of www.ibdjournal.org |
379
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Lai et al
FIGURE 1. Flowchart of study patients with IBD and colonic epithelial changes IND.
colitis, and experience of interpreting pathologist (GI versus nonGI) were not associated with prevalent neoplasia in the colectomy specimen.
Incident Neoplasia in IBD Patients with IND The remaining 71 patients with IBD were followed at this hospital with surveillance colonoscopy for a mean duration of 98.6 months (SD, 65.0; range, 1.7–250.5), and 90% of them had follow-up longer than 1 year. The average surveillance interval was 2.5 years, and 58% of patients had their repeat colonoscopy within 1 year of the diagnosis of IND. Of these, 18 patients (25.4%) developed dysplasia or carcinoma (10 LGD, 5 HGD, and 3 CRC) with a mean interval period of 53.9 months (SD, 36.9) between IND and dysplasia or carcinoma diagnoses. Those who progressed were significantly older than nonprogressors (P ¼ 0.006) (Table 3). A diagnosis of IND made by a GI pathologist was more than twice as likely to progress than the one made by a non–GI pathologist (34.3% versus 13.9%; P ¼ 0.04) (Table 3). A total of 59 cases from patients with IBD and IND (of 71, 83.1%) followed by colonoscopy had slides for review. None of the IND cases were collectively reclassified as positive for dysplasia after histology review by 5 GI pathologists.
380
| www.ibdjournal.org
Reexamination of the follow-up biopsy and/or resection specimens carrying a dysplasia diagnosis confirmed that 7 had LGD (time since IND diagnosis: 21–123 mo), 3 had HGD (time since IND diagnosis: 12–25 mo), and 3 had CRC (time since IND diagnosis: 34–145 mo). The Kaplan-Meier curve of progression to any neoplasia and advanced neoplasia in IBD patients with IND is shown in Figure 2. A total of 52 cases with an index diagnosis of negative for dysplasia were included as a control group. The clinical and pathological characteristics in these 2 groups were similar (Table 4). In the control group, 5 cases progressed to LGD (time since negative for dysplasia diagnosis: 115–182 mo), 1 to HGD (time since negative for dysplasia diagnosis: 15 mo), and 2 to CRC (time since negative for dysplasia diagnosis: 44 and 168 mo). The calculated progression rate (to LGD, HGD, and/or CRC) is 3.2 cases per 100 person-years at risk, and the progression rate (to advanced neoplasia [HGD and/ or CRC]) is 1.5 cases per 100 person-years at risk in the IND group. These progression rates were higher than those in the control group (1.9 and 0.7 cases per 100 person-years at risk; P ¼ 0.1 and 0.2, respectively). Univariate analysis identified that age .44 years (median age) at the time of diagnosis of index IND was associated with a greater probability of progression to neoplasia (hazard ratio, 6.7;
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Outcome of IND in IBD
TABLE 1. Demographics and Clinical Features of Patients with Slides Reviewed in This Study (N ¼ 77)
Mean age (range), yr From IND diagnosis to resection, d Median duration of colitis (range), yr No. colonoscopies No. biopsies Time of follow-up, y Gender, male (%) Follow-up results, n (%) Negative LGD HGD Cancer Original pathologist, n (%) Non-GI GI IBD subtype, n (%) CD UC Othersa Extent of colitis, n (%) Pancolitis Left-sided Rectum only Others PSC, n (%) Absent Present Unknown Indication for colectomy if resection within 6 mo, n (%) Medical refractive IBD IBD with dysplasia Unknown
Resection Group (n ¼ 18)
Colonoscopy Group (n ¼ 59)
35.5 (31.0–52.5) 33.5 (6.5–60.2)
44 (34.5–56.0) N/A
6.0 (2.0–17.8)
7.0 (1.0–11.8)
N/A N/A N/A 12 (64)
3.5 11 6.8 34
(1.2–7.0) (7.0–28.0) (3.1–9.6) (58)
14 0 4 0
46 7 3 3
(78) (12) (5) (5)
(78) (0) (22) (0)
8 (44) 10 (56)
30 (51) 29 (49)
3 (17) 15 (83) 0 (0)
6 (10) 51 (86) 2 (3)
6 5 6 1
(33) (28) (33) (6)
15 (83) 0 (0) 3 (17)
11 (61) 4 (22) 3 (17)
29 15 13 2
(49) (25) (22) (3.4)
51 (86) 8 (14) 0 (0)
N/A
a
Cases with a presumable diagnosis of UC found to have evolved into typical Crohn’s disease after colectomy. CD, Crohn’s disease; N/A, not applicable.
95% confidence interval [CI], 1.5–33.3; P ¼ 0.01) (Table 5, Fig. 2C). In addition, the frequency of surveillance colonoscopy (defined as number of colonoscopies per year) was associated with progression (hazard ratio, 2.5 for 1 more per yr; 95% CI, 1.6–3.9; P , 0.0001) (Table 5). IBD duration, the presence of PSC, the extent of disease, the number of colonoscopy procedures, or the number of biopsies was not associated with neoplasia progression (Table 5).
TABLE 2. Prevalent Neoplasia in IBD Patients with an Index Diagnosis of IND
Mean age (range), yr Gender, male (%) IBD type, n (%) UC CD Median duration of colitis (range), yr Extent of colitis, n (%) Pancolitis Left-sided Rectum only Others Interpreting pathologist, n (%) GI Non–GI pathologist Dysplasia as indication for colectomy,a n (%)
Prevalent Neoplasia (n ¼ 4)
No Prevalent Neoplasia (n ¼ 14)
P
41.5 (15–56) 3 (75)
35.5 (20–73) 9 (64.3)
0.7 1
4 (100) 0 (0) 16 (5–22)
11 (78.5) 3 (21.5) 2 (0–24)
1 (25) 1 (25) 2 (50) 0 (0)
5 4 4 1
(35.7) (28.6) (28.6) (7.1)
3 (75) 1 (25) 3 (75)
7 (50) 7 (50) 1 (9.0)
1 0.1
1
0.6 0.03
a Only 15 cases had this information available. CD, Crohn’s disease.
DISCUSSION It is well established that IBD has an increased risk for developing CRC.1–4 Detection of dysplasia by surveillance colonoscopy and biopsy is the standard of care to minimize mortality due to CRC in patients with IBD, but diagnostic difficulty and uncertainty remain and may lead to a diagnosis of IND in some cases. Little is known of the natural history of IBD–IND, and as such the management of IND in patients with IBD is controversial. We studied 2 subgroups of patients with IND; 1 group underwent colectomy within 6 months and the other was followed by surveillance colonoscopy. The first subgroup allowed us to estimate the prevalence of neoplasia associated with an index diagnosis of IND to be at least 22.2%. A clinical impression of dysplasia by the surgeon was associated with prevalent neoplasia. No other features (age, gender, IBD subtype, duration of disease, extent of colitis, or the presence of PSC) predicted prevalent dysplasia. Progression in a dysplasia–carcinoma sequence from LGD to HGD and, ultimately, to invasive adenocarcinoma in IBD is well known. However, recent data also suggest that invasive adenocarcinoma may arise from LGD.10 Therefore, in this study, neoplastic progression is defined as LGD, HGD, and CRC, in contrast to other studies that defined progression as the development of HGD or CRC. In this study, the neoplasia incidence rate www.ibdjournal.org |
381
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Lai et al
TABLE 3. Clinical Outcome of IBD patients with a Diagnosis of IND Followed with Colonoscopy
Worst neoplasiaa LGD HGD CRC Age, mean (SD), yr Gender (M/F) Extent of colitis Pancolitis Left-sided Rectum only Others Interpreting pathologist GI Non-GI
IND Progressor (n ¼ 17)
IND Nonprogressor (n ¼ 54)
10 5 2 54.4 (12.6) 13/4
0 0 0 44.1 (13.3) 30/24
9 5 2 1
24 15 13 2
0.1
12 5
23 31
0.04
P
N/A
0.006 0.2
a Worst neoplasia diagnosed in a follow-up specimen. N/A, not applicable.
in patients with index IND is 3.2 cases per 100 person-years, higher than the progression rate of 1.9 cases per 100 personyears in the control group, although the difference is not statistically significant. The progression rate to advanced neoplasia in patients with index IND is 1.5 cases per 100 person-years at risk, approaching the reported progression rate to advanced neoplasia in LGD established in 1 study,8 but is lower than the reported 3 cases per 100 person-years in a meta-analysis.11 However, this progression rate to advanced neoplasia in patients with IND is higher than the progression rate of 0.7 cases per 100 person-years seen in the control group, although the difference again does not reach statistical significance. These results suggest that some cases of IND may represent a transition stage of neoplasia between negative and LGD with an intermediate progression risk between negative for dysplasia and LGD. However, large and prospective studies are needed to further verify or refute this concept. Our finding of a lack of association of patient gender, IBD subtype, duration of disease, extent of colitis, presence of PSC, number of colonoscopy procedures, and number of biopsies is consistent with that reported in a previous study of LGD and IND.8 However, our study identified that age .44 years at diagnosis of index IND was associated with neoplasia progression. This finding differs from a previous report in which age was not associated with neoplasia progression8; however, their cohort was a mix of patients with LGD and IND (predominantly LGD) and small in size (N ¼ 35), with progression defined as HGD and/or CRC. A potential explanation for the observed association of age
382
| www.ibdjournal.org
FIGURE 2. A, Kaplan-Meier curve of progression to neoplasia (LGD, HGD, or cancer) and (B) advanced neoplasia (HGD or cancer) in patients with IND, compared with control group. C, Age at the time of IND diagnosis as a risk factor for neoplasia progression. Univariate analysis revealed the cutoff age of 44 years as having significant probability of progression (hazard ratio, 6.7; 95% CI, 1.5–33.3; P ¼ 0.01).
Inflamm Bowel Dis Volume 21, Number 2, February 2015
Outcome of IND in IBD
TABLE 4. Clinical and Pathological Characteristics of Negative and IND Groups Negative (n ¼ 52) Mean age (SD), yr Median duration of follow-up (range), yr Median duration of colitis (range), yr Gender, male (%) Disease type, n (%) UC CD Othersa Disease extent, n (%) Pancolitis Left-sided Rectum only Other PSC, present (%)
IND (n ¼ 59)
P
42.5 (31–53.2)a 44 (34.5–56) 0.3 5.8 (2.7–13.3) 6.8 (3.1–9.7) 0.3 2.2 (0.3–9.2)
7 (1.0–11.8) 0.1
25 (48)
34 (58)
0.3
39 (75) 11 (21) 2 (4)
51 (86) 6 (10) 2 (3)
0.3
25 14 8 5 6
29 15 13 2 8
0.5
(48) (27) (15) (10) (12)
(49) (25) (22) (3) (14)
TABLE 5. Univariate Analysis of Features in Relation to Progression to Definite Dysplasia in Patients with an Index IND Diagnosis
1
Age (.44 versus #44), yr IBD duration (#7 versus .7), yr PSC (present versus absent) Interpreting pathologist (GI versus non-GI) No. colonoscopies per 1 more Frequency of surveillance per 1 more per year No. biopsies per 1 more Extent of disease (pancolitis as reference) Left-sided Rectum only Others
Hazard Ratio
95% CI
P
6.7 1.7 2.2 1.69
1.5–33.3 0.56–5.3 0.59–7.93 0.55–5.21
0.01 0.3 0.2 0.4
1.02 2.5
0.89–1.17 1.6–3.9
0.8 ,0.0001
1.3
0.96–1.8
0.09
0.66 0.21 1.54
0.17–2.51 0.03–1.72 0.19–12.51
0.5 0.1 0.7
CI, confidence interval.
a
Cases with a presumable diagnosis of UC found to have evolved into typical Crohn’s disease after colectomy. CD, Crohn’s disease.
with neoplasia progression in our study is that IND may have existed in the large bowel for a longer duration in the older patients than in the younger patients. Another possibility is that the development of neoplasia is accelerated in older patients. In addition, we identified that frequency of surveillance colonoscopy was associated with progression; however, it is unclear whether this reflects increased disease activity in these patients or possibly detection bias. There are limitations to this study. Our data are limited by the quality and availability of the clinical information found within the medical records. For example, many cases did not have detailed endoscopic reports, preventing us from determining whether the IND biopsy was from a lesion or from flat mucosa and from correlating the location of index IND with the location of prevalent or incident neoplasia. In addition, many cases did not have detailed history, such as a family history of CRC and medication history. The rate of prevalent neoplasia in our study may represent an overestimation because of the referral bias, and we could not completely rule out an outside diagnosis of dysplasia before the IND diagnosis at our institute. Similarly, the study population was based on a highly specialized tertiary care center where GI pathology subspecialty practice started in 2002, 2 years before the end of the study period, and conclusions from this study may not be applicable to other practice settings. In summary, our study reaffirms the value of IND as a diagnostic category for IBD colonic biopsies. The diagnosis
of IND carries significant prevalent and incident rates for neoplasia in patients with IBD, and clinical impression should be taken into consideration in managing these patients because of the significant rate of prevalent neoplasia. IBD patients with IND should be followed by repeat colonoscopy after a short interval (within months) and then managed based on the finding of the repeat biopsy. Our study also indicates that the adherence of IBD patients with mucosal changes IND to surveillance needs to be improved. Large studies are needed to confirm our findings.
REFERENCES 1. Rosenqvist H, Lagercrantz R, Edling N. Ulcerative colitis and carcinoma coli. Lancet. 1959;1:906–908. 2. Ekbom A, Helmick C, Zack M, et al. Increased risk of large-bowel cancer in Crohn’s disease with colonic involvement. Lancet. 1990; 336:357–359. 3. Jess T, Winther KV, Munkholm P, et al. 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. 4. Jess T, Gamborg M, Matzen P, et al. Increased risk of intestinal cancer in Crohn’s disease: a meta-analysis of population-based cohort studies. Am J Gastroenterol. 2005;100:2724–2729. 5. Breynaert C, Vermeire S, Rutgeerts P, et al. Dysplasia and colorectal cancer in inflammatory bowel disease: a result of inflammation or an intrinsic risk? Acta Gastroenterol Belg. 2008;71:367–372. 6. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol. 1983;14:931–968. 7. Zisman TL, Bronner MP, Rulyak S, et al. Prospective study of the progression of low-grade dysplasia in ulcerative colitis during current cancer surveillance guidelines. Inflamm Bowel Dis. 2012;18:2240– 2246. www.ibdjournal.org |
383
Lai et al
8. Pekow JR, Hetzel JT, Rothe JA, et al. Outcome after surveillance of lowgrade and indefinite dysplasia in patients with ulcerative colitis. Inflamm Bowel Dis. 2010;16:1352–1356. 9. van Schaik FD, ten Kate FJ, Offerhaus JA, et al. Misclassification of dysplasia in patients with inflammatory bowel disease: consequences for progression rates to advanced neoplasia. Inflamm Bowel Dis. 2011;17:1108–1116.
384
| www.ibdjournal.org
Inflamm Bowel Dis Volume 21, Number 2, February 2015
10. Harpaz N, Polydorides AD. Colorectal dysplasia in chronic inflammatory bowel disease: pathology, clinical implication, and pathogenesis. Arch Pathol Lab Med. 2010;134:876–895. 11. Thomas T, Abrams KA, Robinson RJ, et al. Meta-analysis: cancer risk of low-grade dysplasia in chronic ulcerative colitis. Aliment Pharmacol Ther. 2006;25:657–668.
