Background and Aims
Inflammatory bowel disease (IBD) is a well-known risk factor for colorectal cancer (CRC). Current guidelines propose complete endoscopic resection of dysplasia in IBD patients with close endoscopic follow-up. Current data on the risk of neoplasia after endoscopic resection of dysplasia in IBD patients are limited.
Methods
Multiple databases were searched from inception through August 2019 to identify studies that reported on incidence and/or recurrence of neoplasia after resection of dysplasia in patients with IBD. Outcomes from the included studies were pooled to estimate the risk of neoplasia after dysplasia resection in IBD patients.
Results
From 18 studies, 1037 IBD patients underwent endoscopic resection for a total of 1428 colonic lesions. After lesion resection, the pooled risk (rate per 1000 person-years of follow-up) of CRC was 2 (95% confidence interval [CI], 0-3), the pooled risk of high-grade dysplasia was 2 (95% CI, 1-3), and the pooled risk of any lesion was 43 (95% CI, 30-57). Meta-regression analysis based on lesion location (right, left), lesion size (mean and/or median size in mm), lesion type (Paris type I, Paris type II), endoscopic resection technique (EMR, endoscopic submucosal dissection, or polypectomy), and lesion histology (low-grade dysplasia, high-grade dysplasia) did not influence the reported outcomes.
Conclusions
Risk of CRC after dysplasia resection in IBD patients appears to be low, supporting the current strategy of resection and surveillance.
Abbreviations:
CD (Crohn’s disease), CI (confidence interval), CRC (colorectal cancer), ESD (endoscopic submucosal dissection), HGD (high-grade dysplasia), IBD (inflammatory bowel disease), LGD (low-grade dysplasia), PI (prediction interval), UC (ulcerative colitis)Inflammatory bowel disease (IBD) is a well-established risk factor for colorectal cancer (CRC). CRC accounts for approximately 10% to 15% of all deaths in patients with IBD, particularly in patients with ulcerative colitis (UC) as compared with Crohn’s disease (CD). Extent and duration of the disease pose a greater risk, and frequent surveillance colonoscopy is recommended to detect dysplasia. Endoscopically visible colon dysplasia in IBD can occur as nonpolypoid or polypoid lesions.
1
,2
Polypoid lesions are typically well circumscribed, sessile, or pedunculated and resemble adenomas in patients without IBD. Nonpolypoid lesions are flat, broad, sessile, irregular, ulcerating, or constricting lesions that have a high association with metachronous or synchronous adenocarcinoma.
3
Nonpolypoid dysplasias are increasingly being detected with the help of high-definition white-light endoscopy along with chromoendoscopy. In the past, a finding of dysplasia within the affected segment of bowel was an indication for colectomy because of the risk of cancer in 32% to 42% of patients.3
International guidelines now recommend that dysplasia with clearly visible and distinct margins be considered as “endoscopically resectable.”
4
The surveillance for colorectal endoscopic neoplasia detection and management in IBD patient per international consensus recommendations is close surveillance colonoscopy rather than colectomy after complete removal of endoscopically resectable dysplastic lesions. The risk of CRC in polypoid and/or nonpolypoid lesions in IBD patients is not well studied, and the guideline is based on low quality of evidence.1
In patients with IBD, underlying disease inflammation can make the borders unclear and submucosal fibrosis can render the resection technically challenging. EMR is based on traditional polypectomy principles and is relatively easier to complete with low adverse event rates. However, completeness of excision can be difficult to ascertain. On the other hand, endoscopic submucosal dissection (ESD) facilitates en bloc resection, is technically challenging, is associated with higher rate of adverse events, and generally is expected to have lower rates of recurrence.
It is clear that the diagnosis, prognosis, and management guidelines of polypoid and/or nonpolypoid colonic lesions in IBD are an area of ongoing debate, mainly because of paucity of good published data. We therefore conducted this meta-analysis to add quality evidence to the current literature with respect to clinical outcomes after resection of colon dysplasia in patients with IBD. We focused on neoplasia risk in terms of incidence and/or recurrence after resection of dysplastic lesions in patients with IBD and also looked into the safety of the endoscopic resection procedure in this population.
Methods
Search strategy
A comprehensive search of several databases from inception to August 30, 2019, limited to English language only, and excluding animal studies was conducted. The databases included Ovid Medline and Epub ahead of print, In-Process and Other Non-Indexed Citations and Daily, Ovid Embase, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus. The search strategy was designed and conducted by an experienced librarian with input from the study’s principal investigator. Controlled vocabulary supplemented with key words was used to search for studies describing endoscopic resection in patients with IBD. The full search strategy is available in Appendix 1 (available online at www.giejournal.org). In addition, reference lists of the original articles and relevant review articles were scrutinized to identify studies missed on the initial search.
Study selection
Inclusion criteria were studies that reported on the risk of neoplasia incidence and/or recurrence after endoscopic resection of dysplasia in IBD patients. Studies were included irrespective of the type of IBD (UC and/or CD), inpatient/outpatient setting, follow-up time, abstract/article status, and geography as long as they provided the appropriate data needed for the analysis.
Exclusion criteria were studies with study sample size <10, studies done in pediatric population (age <18 years), and studies not published in English. In cases of multiple publications from a single research group reporting on the same patient same cohort and/or overlapping cohorts, data from the most recent and/or most appropriate comprehensive report were retained. When needed, authors were contacted via e-mail for clarification of data and/or study cohort overlap.
Data abstraction
Data on study-related outcomes from individual studies were abstracted independently onto a standardized form by at least 2 authors (L.L.K., S.R.K.). In case of discrepancy, a third author (S.C.) assessed the data to achieve consensus. The primary study authors were contacted via e-mail as and when needed for further information and/or clarification on data. We assessed the following outcomes: pooled risk of CRC after dysplasia resection, pooled risk of high-grade dysplasia (HGD) after dysplasia resection, pooled risk of any lesion after dysplasia resection, pooled rate of patients requiring surgery, and pooled risk of adverse events (perforation and bleeding) after resection of dysplasia.
Subgroup analyses and meta-regression analyses
Predetermined subgroup analyses were carried out based on the type of IBD (UC alone, UC + CD) and study sample size (<50, >50). Meta-regression was planned based on lesion location (right, left), lesion size (mean and/or median size in mm), lesion type (Paris type I, Paris type II), endoscopic resection technique (EMR, ESD, polypectomy), and lesion histology (low-grade dysplasia [LGD], HGD) to assess potential causes of heterogeneity and predictors of clinical outcomes.
Assessment methodology and definitions
This study was performed according to the Meta-analysis of Observational Studies in Epidemiology recommendations and following its checklist (Appendix 2, available online at www.giejournal.org).
5
The Newcastle-Ottawa scale for cohort studies was used to assess the quality of studies.6
Two authors (B.P.M., S.C.) did the quality scoring independently. This quality score consisted of 8 questions, the details of which are provided in Supplementary Table 1 (available online at www.giejournal.org).Data on the reported incidence and/or recurrence of CRC, HGD, or any lesion in IBD patients undergoing endoscopic dysplasia removal were collected as reported in the original articles. Any recurrent lesion was defined tubular adenoma, tubulovillous adenoma, villous adenoma, flat dysplasia, dysplasia-like lesion and/or mass, HGD, and CRC. Bleeding events were considered as adverse events if they warranted blood transfusion and/or inpatient stay. Perforation events were considered as adverse events as reported in the original articles. Data on patients referred for surgery were collected based on the following: secondary to adverse event of ESD, EMR, polypectomy including perforation, and/or bleeding; diagnosis of invasive cancer; presence of noncurative synchronous lesions; progression of disease form LGD to HGD; and recurrence of disease not amenable to mucosal resection.
Statistical analysis
We used meta-analysis techniques to calculate the pooled estimates in each case following the methods suggested by DerSimonian and Laird using the random-effects model.
7
When the incidence of an outcome was 0 in a study, a continuity correction of .5 was added to the number of incident cases before statistical analysis.8
We assessed heterogeneity between study-specific estimates by using Cochran Q statistical test for heterogeneity; 95% prediction interval (PI), which deals with the dispersion of the effects10
, 11
, 9
; and the I2 statistics.12
,13
For this, values of <30%, 30% to 60%, 61% to 75%, and >75% were suggestive of low, moderate, substantial, and considerable heterogeneity, respectively.14
Publication bias was ascertained qualitatively by visual inspection of funnel plots and quantitatively by the Egger test.15
When publication bias was present, further statistics using the fail-Safe N test and Duval and Tweedie’s Trim and Fill test was used to ascertain the impact of the bias.16
Three levels of impact were reported based on the concordance between the reported results and the actual estimate if there were no bias. The impact was reported as minimal if both versions were estimated to be same, modest if effect size changed substantially but the final finding would still remain the same, and severe if basic final conclusion of the analysis is threatened by the bias.17
All analyses were performed using Comprehensive Meta-Analysis software, version 3 (BioStat, Englewood, NJ, USA).Results
Search results and population characteristics
From an initial 799 studies, 419 records were screened and 33 full-length articles were assessed. Eighteen studies were included in the final analysis.
18
, 19
, 20
, - Buchner A.M.
- Mahmud N.
- Elhanafi S.
- et al.
Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].
Gastrointest Endosc. 2019; 89: AB69
21
, 22
, 23
, 24
, 25
, 26
, 27
, 28
, 29
, 30
, 31
, 32
, 33
, 34
, 35
The schematic diagram demonstrating our study selection is illustrated in Supplementary Figure 1 (available online at www.giejournal.org).A total of 1037 patients with IBD had resections for 1428 colonic lesions and were followed up for disease recurrence. Of these, 933 patients had UC and 126 patients had CD. Sixty-three percent were men and 37% women. The mean and/or median age ranged from 53 to 69 years, and patients were followed for a mean duration of 6 to 82 months. Six studies
22
,25
,28
,29
,32
,33
reported on the resection of 359 nonpolypoid lesions in 224 patients, 2 studies19
,27
on the resection of 135 polypoid lesions in 122 patients, and the remaining 10 studies18
,20
,- Buchner A.M.
- Mahmud N.
- Elhanafi S.
- et al.
Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].
Gastrointest Endosc. 2019; 89: AB69
21
,23
,24
,26
,30
,31
,34
,35
reported on a combination, which were 934 lesions in 691 patients. Further baseline study and population characteristics are summarized in Supplementary Table 1.Characteristics and quality of included studies
Six included studies were prospectively done,
20
,- Buchner A.M.
- Mahmud N.
- Elhanafi S.
- et al.
Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].
Gastrointest Endosc. 2019; 89: AB69
22
,23
,30
, 31
, 32
and 6 studies were based on multicenter data.18
,25
,26
,28
,29
,33
The detailed assessment of study quality can be found in Supplementary Table 2. Overall, 14 studies were considered to be of high quality18
,19
,21
,23
,25
, 26
, 27
, 28
, 29
, 30
, 31
,33
, 34
, 35
and 4 of medium quality.20
,- Buchner A.M.
- Mahmud N.
- Elhanafi S.
- et al.
Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].
Gastrointest Endosc. 2019; 89: AB69
22
,24
,31
There were no low-quality studies.Meta-analysis outcomes
From 18 studies, 1037 patients were treated for 1428 colon lesions.
18
, 19
, 20
, - Buchner A.M.
- Mahmud N.
- Elhanafi S.
- et al.
Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].
Gastrointest Endosc. 2019; 89: AB69
21
, 22
, 23
, 24
, 25
, 26
, 27
, 28
, 29
, 30
, 31
, 32
, 33
, 34
, 35
The pooled risk per 1000 years of patient follow-up of CRC was 2 (95% confidence interval [CI], 0-3) (Fig. 1), of HGD was 2 (95% CI, 1-3) (Fig. 2), and of any lesion was 43 (95% CI, 30-57) (Fig. 3). In the subcategory of any lesions seen on follow-up, 67 of 184 recurrent lesions (36.4%) were treated endoscopically. More than half of these lesions (55%) were treated by EMR. The exact endoscopic technique of treatment was not mentioned in 24 patients, and the rest were treated by ablation (n = 1), polypectomy (n = 1), and EMR-assisted ESD (n = 4).
Figure 1Forest plot of CRC incidence. CRC, Colorectal cancer; CI, confidence interval.
Figure 2Forest plot of HGD incidence. HGD, High-grade dysplasia; CI, confidence interval.
Figure 3Forest plot of any dysplasia incidence. CI, Confidence interval.
The pooled rate of patients requiring surgical management was 9.9% (95% CI, 6.5-14.7) (Supplementary Fig. 2, available online at www.giejournal.org). One hundred seventeen patients required surgery. Thirty-six patients (30.7%) were referred to surgery for a confirmed diagnosis of adenocarcinoma, 28 patients (23.9%) underwent surgery for unresectable and/or metachronous lesions, 23 patients (19.7%) had a diagnosis of LGD, 12 patients (10.3%) had HGD, 15 patients (12.8%) were referred because of strictures/stenosis/fistula, and 3 patients (2.6%) because of IBD disease progression.
The pooled rate of bleeding reported after dysplasia resection was 2.2% (95% CI, 1.1-4.4) (Supplementary Fig. 3, available online at www.giejournal.org), and the pooled rate of perforation was 2% (95% CI, .9-4.4), (Supplementary Fig. 4, available online at www.giejournal.org). The pooled results are summarized in Table 1, along with the PIs and I2 values.
Table 1Pooled results
| Pooled rate (95% confidence interval, I2) Pooled rate per 1000 person-years of follow-up | Heterogeneity (95% prediction interval, I2) | |
|---|---|---|
| Pooled risk of colorectal cancer | 2 (0-3)/1000 person-years; 17 studies | 1.2-4.2; 0 |
| Pooled risk of high-grade dysplasia | 2 (1-3)/1000 person-years; 17 studies | 1-7; 0 |
| Pooled risk of any lesion | 43 (30-57)/1000 person-years; 18 studies | 1-63; 88 |
| Pooled rate of patients requiring surgery | 9.9% (6.5-14.7); 18 studies | 2-40; 80 |
| Pooled rate of bleeding | 2.2% (1.1-4.4); 13 studies | 1-14; 47 |
| Pooled rate of perforation | 2% (.9-4.4); 13 studies | 1-17; 50 |
| Publication bias (Egger’s 2-tailed P value) | .01 |
Subgroup and meta-regression analysis
Subgroup analysis based on IBD type (UC only and UC + CD)
Eleven studies reported on neoplasia incidence in patients with UC alone, and 5 studies had sample sizes that included both UC and CD patients. The pooled CRC and/or HGD risk was 2 per 1000 person-years of follow-up in both groups. Risk of any lesion was 35 (95% CI, 20-51) in the UC-only sample and 50 (95% CI, 25-75) in the UC + CD sample (Table 2).
Table 2Pooled risk per 1000 person-years evaluated by stratified analyses
| CRC (incidence per 1000 person-years of follow-up) | HGD (incidence per 1000 person-years of follow-up) | Any lesion (incidence per 1000 person-years of follow-up) | |
|---|---|---|---|
| UC only vs UC + CD | 2 (0-3); 11 studies | 2 (0-3); 11 studies | 35 (20-51); 12 studies |
| 2 (0-5); 5 studies; P = .72 | 2 (0-4); 5 studies; P = .95 | 50 (25-75); 5 studies; P = .08 | |
| Sample size: <50, >50 | 10 (1-20); 7 studies | 12 (2-22); 7 studies | 50 (23-77); 7 studies |
| 2 (0-3); 10 studies; P = .07 | 2 (0-3); 10 studies; P = .04 | 41 (26-56); P = .57 |
CRC, Colorectal cancer; HGD, high-grade dysplasia; UC, ulcerative colitis; CD, Crohn’s disease.
Subgroup analysis based on sample size (<50 and >50)
Seven studies had patient sample size <50 and 10 studies >50. The pooled CRC risk per 1000 person-years of follow-up in studies with <50 patients was 10 (95% CI, 1-20) and in studies with >50 patients was 2 (95% CI, 0-3). The pooled HGD risk per 1000 person-years follow-up in studies with <50 patients was 12 (95% CI, 2-22) and in studies with >50 patients was 2 (95% CI, 0-3). The pooled risk of any lesion was 50 (95% CI, 23-77) in studies with <50 patients and 41 (95% CI, 26-56) in studies with >50 patients (Table 2).
Meta-regression analysis
Meta-regression analysis based on lesion location (right, left), lesion size (mean and/or median size in mm), lesion type (Paris type I, Paris type II), endoscopic resection technique (EMR, ESD, polypectomy), en bloc/piecemeal, and lesion histology (LGD, HGD) did not reveal any influence on the reported outcomes (Table 3).
Table 3Results of meta-regression analysis
| Variable | Random effects, Knapp-Hartung, 2-sided P value (CRC, HGD, any lesion) |
|---|---|
| IBD type: UC only | .77, .89 |
| IBD type: UC + CD | .96, .74 |
| Paris type I | .35, .45, .89 |
| Paris type II | .79, .6, .18 |
| EMR | .41, .37, .54 |
| ESD | .95, .58, .45 |
| Polypectomy | .34, .65, .89 |
| En bloc removal | .2, .4, .3 |
| Piecemeal removal | .5, .1, .6 |
| Right-sided lesion | .28, .18, .44 |
| Left-sided lesion | .07, .24, .18 |
| Lesion size | .06, .46, .97 |
| LGD | .06, .1, .16 |
| HGD | .09, .6, .05 |
CRC, Colorectal cancer; HGD, high-grade dysplasia; IBD, inflammatory bowel disease; UC, ulcerative colitis; CD, Crohn’s disease; ESD, endoscopic submucosal dissection; LGD, low-grade dysplasia.
Validation of meta-analysis results
Sensitivity analysis
To assess whether any 1 study had a dominant effect on the meta-analysis, we excluded 1 study at a time and analyzed its effect on the main summary estimate. On this analysis, no single study significantly affected the outcome or heterogeneity. The cumulative effect of pooled data based on the year of publication was done for the pooled rates of CRC, HGD, recurrence of any dysplasia, and patients referred to surgery (Supplementary Figure 5, Supplementary Figure 6, Supplementary Figure 7, Supplementary Figure 8, available online at www.giejournal.org). Based on this analysis, the pooled rate of CRC and/or HGD seemed to stabilize at 2 per 1000 person-years of follow-up from year 2007 onward, and no changes were observed in the referral pattern for surgical resection over the study period from 1999 to 2017 with a slight decrease from 2017 to 2019.
Heterogeneity
We assessed dispersion of the calculated rates using the PI and I2 percentage values. The PI gives an idea of the range of the dispersion, and I2 tell us what proportion of the dispersion is true versus chance.
11
The calculated PIs are reported with pooled rates in Table 1. The calculated PI was narrow, with minimal heterogeneity for the calculated pooled rates of risk of CRC, HGD, bleeding, and perforation. Based on subgroup (Table 2) and meta-regression (Table 3) analysis, the study sample size seemed to have contributed to the observed heterogeneity.Publication bias
Based on visual inspection of the funnel plot and quantitative measurement that used the Egger regression test, there was evidence of publication bias (Supplementary Fig. 9 [available online at www.giejournal.org], Egger’s 2-tailed P = .01). Further statistics, using the fail-Safe N test and Duval and Tweedie’s Trim and Fill test, did not reveal any major changes to the calculated pooled rates.
Quality of evidence
The quality of evidence was rated for results from the meta-analysis according to the grading of recommendations assessment, development and evaluation working group approach.
36
Observational studies begin with a low-quality rating, and based on the risk of bias, heterogeneity, and publication bias, the quality of this meta-analysis would be considered as low-quality evidence.Discussion
In this meta-analysis, we report the pooled risk of incidence and/or recurrence of neoplastic lesion after resection of colonic dysplasia in patients with IBD. We report the pooled risk of CRC and HGD to be 2 per 1000 person-years of follow-up. The incidence and/or recurrence rate of any lesion was 43 per 1000 person-years of follow-up. Although CRC and HGD were unusual after endoscopic treatment, the recurrence of any dysplasia was not unusual. However, many recurrent lesions could be managed endoscopically, highlighting the effectiveness of endoscopic resection and follow-up for patients with dysplastic lesions.
We hypothesized that our low numbers could be related to recent technological advances in detecting and treating dysplastic lesions in patients with IBD. High-definition white-light endoscopy and chromoendoscopy has greatly influenced the detection of these lesions by clearly demarcating the borders. However, we did not observe any changes in the pooled rates from 2007 onward based on our cumulative analysis stratified by year of study publication. The other possibility is that the advent of newer biologic medications introduced in the treatment of IBD may have a role.
37
With 1037 patients treated for 1428 colon lesions from 18 studies, this is the largest and most updated meta-analysis in this topic.The pooled rate of IBD patients diagnosed with any dysplasia requiring surgery, in terms of total colectomy or total proctocolectomy, was 9.9%. This result needs to be interpreted with caution, because nearly half of the included studies were published in a time when the current techniques of endoscopic resections (like EMR and ESD) were still in their infancy and the possibility of endoscopically removing a lesion in its entirety was low. Our cumulative analysis demonstrates that referral to surgery seemed to slightly decrease between 2017 and 2019.
The pooled rate of bleeding and of perforation were 2%. Recent advancements in endoscopic resection techniques (EMR, ESD) and the steeply improving learning curve in these techniques hold promise and can have significant impact on the need for surgery. This needs to be evaluated in future studies.
Our subgroup analysis, based on the type of IBD (UC alone and UC + CD) and sample size (<50 and >50) did not reveal major changes and/or differences to the reported rates of CRC. However, the risk of HGD based on the study sample size (<50 vs >50) was significantly lower in studies with sample size >50 patients (12 vs 2, P = .04), highlighting the importance of adequate sample size.
We assessed the effects of lesion characteristics based on the Paris classification; EMR, ESD, or polypectomy; en bloc or piecemeal removal; location of the lesion (right/left); and lesion histology (LGD or HGD) by mean of meta-regression analysis. A subgroup analysis was not possible, because the original studies did not categorize the outcomes based on these parameters. Our meta-regression analysis based on lesion location (right, left), lesion size (mean and/or median size in mm), lesion type (Paris type I, Paris type II), endoscopic resection method (EMR, ESD, polypectomy), and lesion histology (LGD, HGD) did not reveal any major changes to the reported pooled outcomes or the heterogeneity.
There was, however, a trend toward significance with left-sided lesions (P = .06) and lesion size (P = .06) that support the well-known notion that left-sided lesions as well as large lesions are more prone to malignant conversion than right-sided or smaller lesions. We also noted a trend toward significance with LGD (P = .06) and HGD (P = .09). This is a novel finding of this study and seems to support the hypothesis that tumor progression in the inflammatory bowel segment does not always follow the LGD to HGD to cancer pathway. The failure to achieve statistical significance could be related to insufficient sample size and variability in the definition of outcomes in the individual studies. However, it is important to note that meta-regression is a weak statistic when it comes to the assessment of potential variables that may affect the calculated outcomes.
We compared our study findings with previously published data. In the 2014 meta-analysis by Wanders et al,
2
the pooled CRC rate was 5.3 per 1000 person-years of follow-up (from 10 studies that included 376 patients) and the pooled incidence rate for any dysplasia was 65 occurrences per 1000 patient-years. The updated pooled rates in this meta-analysis are much lower. The 2007 meta-analysis by Thomas et al38
reported a CRC rate of 14 per 1000 patient-years of follow-up from 7 studies that evaluated 239 chronic UC patients with LGD. The aforementioned differences could be related to the higher total number of studies included in this analysis and to the recent improvements in chromoendoscopy and resection techniques (EMR, ESD).The strengths of this review are as follows: systematic literature search with well-defined inclusion criteria, careful exclusion of redundant studies, inclusion of good quality studies with detailed extraction of data, and rigorous evaluation of study quality. We calculated and presented the PI, which helps to understand the range of the neoplasia incidence in the larger universe of patients with IBD who undergo resection of dysplasia. The PIs were narrow and heterogeneity percentage was minimal with the calculated CRC and/or HGD rates.
There are limitations to this study, most of which are inherent to any meta-analysis. The included studies were not entirely representative of the general population and community practice, with most studies performed in tertiary care referral centers with a large range in follow-up times. Our analysis included studies that were retrospective in nature, contributing to selection bias. We were not able to assess other important determinants of neoplasia recurrence in IBD patients such as coexistent primary sclerosing cholangitis, unifocality versus multifocality of neoplasia at baseline, coexistent “invisible” lesions in random biopsy samples, and the accompanying disease activity.
Studies included hot biopsy sampling as a procedure for endoscopic removal, which might seem inappropriate, because the resected margin cannot be evaluated in the specimen. However, based on our sensitivity analysis and removing the studies that performed hot biopsy sampling (Choi et al
21
and Kisiel et al27
), our overall CRC and HGD rates did not change. The rate of any lesion recurrence was 37 per 1000 person-year of follow-up instead of 43, highlighting that complete resection of lesions with EMR/ESD or polypectomy demonstrates better results than using a hot biopsy forceps.A subgroup analysis of the pooled rates by ESD and EMR is one of the most vital clinical questions we were unable to answer. This would have provided the ability to see the impact of individual studies and a better sense of effect. Unfortunately, the studies did not report on the number of patients who underwent ESD or EMR with outcomes of CRC and/or HGD. As a result, a subgroup analysis of pooled rates pertaining to ESD and EMR was not feasible. Additionally, whether the lesion was removed from within or outside the area of colitis is difficult to know with absolute certainty. Nevertheless, our study is the best currently available estimate in literature thus far with respect to the cancer risk in patients with IBD undergoing dysplasia resection.
In conclusion, based on our meta-analysis, the pooled risk to CRC and/or HGD in IBD patients undergoing dysplasia resection is 2 per 1000 person-years of follow-up and the risk of any lesion is 43 per 1000 person-years follow-up. Our results are much lower than previously reported, and therefore our findings support the current strategy of resection and surveillance after dysplasia resection in IBD patients. Our analysis is limited by low-quality evidence because of the observational nature of the studies. We observed that the follow-up time was highly variable in the included studies. There is a need for a prospective, controlled clinical trial that incorporates high-definition white-light endoscopy and chromoendoscopy for dysplasia detection and targeted removal of premalignant lesions with EMR/ESD over a prolonged follow-up time. Until then, risk assessment of individual IBD patients should not only be based on the present meta-analysis results, but also on other pertinent individual patient parameters.
Acknowledgment
We thank Leslie Hassett, MLS (Outreach librarian, Mayo Clinic Libraries, Rochester, MN) for help with the systematic literature search.
Appendix
Supplementary Figure 1Study selection flow chart.
Supplementary Figure 2Forest plot of pooled rate of patients requiring surgery. CI, Confidence interval.
Supplementary Figure 3Forest plot of pooled rate of bleeding. CI, Confidence interval.
Supplementary Figure 4Forest plot of pooled rate of perforation. CI, Confidence interval.
Supplementary Figure 5Cumulative analysis of CRC by year of study publication. CRC, Colorectal cancer; CI, confidence interval.
Supplementary Figure 6Cumulative analysis of HGD by year of study publication. HGD, High-grade dysplasia; CI, confidence interval.
Supplementary Figure 7Cumulative analysis of recurrence (any dysplasia) by year of study publication. CI, Confidence interval.
Supplementary Figure 8Publication bias, funnel plot. CI, Confidence interval.
Supplementary Figure 9Funnel plot of standard error by logit event rate.
Supplementary Table 2Study quality assessment
| Study | Selection | Comparability | Outcome | Score | Quality | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Representativeness of the average adult in community Population based: 1; multicenter: .5; single center: 0 | Cohort size >40 patients: 1; 39 to 20: 0.5; <20: 0 | Information on clinical outcomes Information with clarity: 1; information derived from percentage value: .5; unclear: 0 | Outcome not present at start Not present: 1; present: 0 | Factors comparable between the groups Yes: 1; no: 0 | Adequate clinical assessment Yes: 1; no: 0 | Follow-up time Yes: 1; not mentioned: 0 | Adequacy of follow-up All patients followed up: 1; >50% followed up: 0.5; <50% followed up or not mentioned: 0 | Maximum = 8 | High >6, medium 4-6, low <4 | |
| Al-Kandari, 2019 | .5 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 7 | High |
| Branquinho, 2016 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Buchner, 2019 | 0 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6 | Medium |
| Choi, 2014 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Gulati, 2018 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | .5 | 5.5 | Medium |
| Hurlstone, 2007 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Ignjatovic, 2011 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | .5 | 5.5 | Medium |
| Kaltenbach, 2019 | .5 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Kinoshita, 2018 | .5 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Kisiel, 2011 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Odze, 2004 | .5 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Quinn, 2013 | .5 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 7 | High |
| Rubin, 1999 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Saoula, 2017 | 0 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6 | Medium |
| Smith, 2008 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Suzuki, 2017 | .5 | .5 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Vieth, 2006 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
| Yadav, 2019 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | .5 | 6.5 | High |
- Supplementary Table 1
Appendix 1
Literature search strategy: data sources and search strategies
Database(s): Ovid MEDLINE(R) 1946 to Present and Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Ovid MEDLINE(R) Daily, EBM Reviews - Cochrane Central Register of Controlled Trials July 2019, EBM Reviews - Cochrane Database of Systematic Reviews 2005 to August 29, 2019, Embase 1974 to 2019 August 29
Tabled
1
| # | Searches | Results |
|---|---|---|
| 1 | Endoscopic Mucosal Resection/ | 7237 |
| 2 | (endoscop∗ adj2 resect∗).ti. | 8878 |
| 3 | 1 or 2 | 13,850 |
| 4 | (colon∗ or colorect∗ or colitis or colectom∗).ti. | 588,426 |
| 5 | exp Colon/ | 144,733 |
| 6 | 4 or 5 | 655,035 |
| 7 | 3 and 6 | 2902 |
| 8 | Colonic Polyps/ and (surg∗ or resect∗).ti,ab,hw,kw,kf. | 3915 |
| 9 | exp ∗Colon/ and (surg∗ or resect∗).ti,ab,hw,kw,kf. | 15,115 |
| 10 | exp ∗Colon/su | 8321 |
| 11 | Colonic Polyps/su or colon polyp/su | 4888 |
| 12 | or/8-11 | 23,199 |
| 13 | endoscop∗.ti. | 214,186 |
| 14 | exp Endoscopy, Digestive System/ or exp digestive tract endoscopy/ | 312,140 |
| 15 | 13 or 14 | 450,911 |
| 16 | 12 and 15 | 6257 |
| 17 | 7 or 16 | 8594 |
| 18 | exp Inflammatory Bowel Diseases/ | 220,418 |
| 19 | ("inflammatory bowel disease∗" or "ulcerative colitis" or crohn∗).ti,ab,hw,kw,kf. | 282,028 |
| 20 | 18 or 19 | 286,468 |
| 21 | 17 and 20 | 579 |
| 22 | 3 and 20 | 237 |
| 23 | 21 or 22 | 679 |
| 24 | (pediatric∗ or children).ti. | 1,404,030 |
| 25 | 23 not 24 | 659 |
| 26 | limit 25 to english language [Limit not valid in CDSR; records were retained] | 611 |
| 27 | 26 not ((exp animals/ or exp nonhuman/) not exp humans/) | 610 |
| 28 | 27 not ((case∗ adj3 (report∗ or stud∗)).ti,ab,hw,kw. or case report/ or exp case study/) | 470 |
| 29 | remove duplicates from 28 | 415 |
Tabled
1
| 1 | TITLE-ABS-KEY ( ( endoscop∗ W/2 resect∗ ) AND ( "inflammatory bowel disease∗" OR "ulcerative colitis" OR crohn∗ ) ) |
| 2 | INDEX(embase) OR INDEX(medline) OR PMID(0∗ OR 1∗ OR 2∗ OR 3∗ OR 4∗ OR 5∗ OR 6∗ OR 7∗ OR 8∗ OR 9∗) |
| 3 | #1 and not #2 |
Appendix 2Meta-analysis of observational studies in epidemiology checklist
| Item no. | Recommendation | Reported on page no. |
|---|---|---|
| Reporting of background should include | ||
| 1 | Problem definition | 4 |
| 2 | Hypothesis statement | -na- |
| 3 | Description of study outcome(s) | 6-7 |
| 4 | Type of exposure or intervention used | 6-7 |
| 5 | Type of study designs used | 6-7 |
| 6 | Study population | 6 |
| Reporting of search strategy should include | ||
| 7 | Qualifications of searchers (eg, librarians and investigators) | 6, Appendix 1 |
| 8 | Search strategy, including time period included in the synthesis and key words | 6, Appendix 1 |
| 9 | Effort to include all available studies, including contact with authors | 7 |
| 10 | Databases and registries searched | Appendix 1 |
| 11 | Search software used, name and version, including special features used (eg, explosion) | NA |
| 12 | Use of hand searching (eg, reference lists of obtained articles) | 6 |
| 13 | List of citations located and those excluded, including justification | Supplementary Figure 1 |
| 14 | Method of addressing articles published in languages other than English | 6 |
| 15 | Method of handling abstracts and unpublished studies | 6 |
| 16 | Description of any contact with authors | 7 |
| Reporting of methods should include | ||
| 17 | Description of relevance or appropriateness of studies assembled for assessing the hypothesis to be tested | 6-8 |
| 18 | Rationale for the selection and coding of data (eg, sound clinical principles or convenience) | 6-8 |
| 19 | Documentation of how data were classified and coded (eg, multiple raters, blinding and interrater reliability) | 6-8 |
| 20 | Assessment of confounding (eg, comparability of cases and control subjects in studies where appropriate) | 7 |
| 21 | Assessment of study quality, including blinding of quality assessors, stratification or regression on possible predictors of study results | 7 |
| 22 | Assessment of heterogeneity | 8 |
| 23 | Description of statistical methods (eg, complete description of fixed or random-effects models, justification of whether the chosen models account for predictors of study results, dose–response models, or cumulative meta-analysis) in sufficient detail to be replicated | 8 |
| 24 | Provision of appropriate tables and graphics | Provided |
| Reporting of results should include | ||
| 25 | Graphic summarizing individual study estimates and overall estimate | Figure 1, Figure 2, Figure 3 Supplementary Figure 2, Supplementary Figure 3, Supplementary Figure 4 |
| 26 | Table giving descriptive information for each study included | Supplementary Table 1 |
| 27 | Results of sensitivity testing (eg, subgroup analysis) | NA |
| 28 | Indication of statistical uncertainty of findings | 10-11 |
| Reporting of discussion should include | ||
| 29 | Quantitative assessment of bias (eg, publication bias) | 11 |
| 30 | Justification for exclusion (eg, exclusion of non-English language citations) | 6 |
| 31 | Assessment of quality of included studies | Supplementary Table 2 |
| Reporting of conclusions should include | ||
| 32 | Consideration of alternative explanations for observed results | 12-14 |
| 33 | Generalization of the conclusions (ie, appropriate for the data presented and within the domain of the literature review) | 12-14 |
| 34 | Guidelines for future research | 14 |
| 35 | Disclosure of funding source | 2 |
From Stroup DF, Berlin JA, Morton SC, et al, for the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) Group. Meta-analysis of observational studies in epidemiology. A proposal for reporting. JAMA 2000;283:2008-12.
NA, Not applicable.
References
- SCENIC international consensus statement on surveillance and management of dysplasia in inflammatory bowel disease.Gastroenterology. 2015; 148: 639-651
- Cancer risk after resection of polypoid dysplasia in patients with longstanding ulcerative colitis: a meta-analysis.Clin Gastroenterol Hepatol. 2014; 12: 756-764
- Are we telling patients the truth about surveillance colonoscopy in ulcerative colitis?.Lancet. 1994; 343: 71-74
- ECCO guidelines on therapeutics in Crohn's disease: surgical treatment.J Crohns Colitis. 2020; 14: 155-168
- Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group.JAMA. 2000; 283: 2008-2012
- Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses.Eur J Epidemiol. 2010; 25: 603-605
- Meta-analysis in clinical trials.Control Clin Trials. 1986; 7: 177-188
- Methods for meta-analysis in medical research.John Wiley & Sons, New York, NY2000: 205-228
- A re-evaluation of random-effects meta-analysis.J R Stat Soc Ser A. 2009; 172: 137-159
- Interpretation of random effects meta-analyses.BMJ. 2011; 342: d549
- Heterogeneity in systematic review and meta-analysis: how to read between the numbers.Gastrointest Endosc. 2019; 89: 902-903
Kanwal F, White D. Systematic reviews and meta-analyses in clinical gastroenterology and hepatology. Clin Gastroenterol Hepatol 2012;10:1184-6.
- Measuring inconsistency in meta-analyses.BMJ. 2003; 327: 557
- GRADE guidelines: 7. Rating the quality of evidence—inconsistency.J Clin Epidemiol. 2011; 64: 1294-1302
Easterbrook PJ, Gopalan R, Berlin JA, et al. Publication bias in clinical research. Lancet 337:867-872.
- Trim and fill: a simple funnel-plot–based method of testing and adjusting for publication bias in meta-analysis.Biometrics. 2000; 56: 455-463
- Publication bias in meta-analysis: prevention, assessment and adjustments.John Wiley & Sons, New York, NY2006
- Endoscopic resections in inflammatory bowel disease: a multicentre European outcomes study.J Crohns Colitis. 2019; 4: 17
- Dysplasia in inflammatory bowel disease: Is endoscopic mucosal resection ready for primetime?.J Crohns Colitis. 2016; 10: S390
- Surveillance chromocolonoscopy with endoscopic mucosal resection of colitis-associated “defiant” dysplastic lesions in patients with long term inflammatory bowel disease (IBD): results from a large tertiary referral medical center [abstract].Gastrointest Endosc. 2019; 89: AB69
- Endoscopic resection of raised dysplastic lesions in ulcerative colitis: long-term outcome [abstract].Gastrointest Endosc. 2014; 1: AB466
- Outcomes of endoscopic resections of large laterally spreading colorectal lesions in inflammatory bowel disease: a single United Kingdom center experience.Inflamm Bowel Dis. 2018; 24: 1196-1203
- Endoscopic mucosal resection for flat neoplasia in chronic ulcerative colitis: Can we change the endoscopic management paradigm?.Gut. 2007; 56: 838-846
- Endoscopic treatment of dysplasia in ulcerative colitis-medium term outcome.Gut. 2011; 1: A207-A208
- Longitudinal outcomes of the endoscopic resection of nonpolypoid dysplastic lesions in patients with inflammatory bowel disease [abstract].Gastrointest Endosc. 2019; 89: AB69-AB70
- The role of colorectal endoscopic submucosal dissection in patients with ulcerative colitis.Gastrointest Endosc. 2018; 87: 1079-1084
- Outcome of sporadic adenomas and adenoma-like dysplasia in patients with ulcerative colitis undergoing polypectomy.Inflamm Bowel Dis. 2011; 18: 226-235
- Long-term follow-up after polypectomy treatment for adenoma-like dysplastic lesions in ulcerative colitis.Clin Gastroenterol Hepatol. 2004; 2: 534-541
- Polypectomy is adequate treatment for adenoma-like dysplastic lesions (DALMs) in Crohn's disease.Inflamm Bowel Dis. 2013; 19: 1186-1193
- Colonoscopic polypectomy in chronic colitis: conservative management after endoscopic resection of dysplastic polyps.Gastroenterology. 1999; 117: 1295-1300
- Outcome of endoscopic ally resected dysplasic lesions in ulcerative colitis.United Eur Gastroenterol J. 2017; 5: A507
- Endoscopic resection of adenoma-like mass in chronic ulcerative colitis using a combined endoscopic mucosal resection and cap assisted submucosal dissection technique.Inflamm Bowel Dis. 2008; 14: 1380-1386
- Endoscopic submucosal dissection of colitis-related dysplasia.Endoscopy. 2017; 49: 1237-1242
- Sporadic adenoma in ulcerative colitis: endoscopic resection is an adequate treatment.Gut. 2006; 55: 1151-1155
- Outcome of endoscopic resection of colonic polyps larger than 10 mm in patients with inflammatory bowel disease.Endosc Int Open. 2019; 7: E994-E1001
- A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analysis.BMJ. 2014; 349: g5630
- ECCO guidelines on therapeutics in Crohn's disease: medical treatment.J Crohns Colitis. 2020; 14: 4-22
- Meta-analysis: cancer risk of low-grade dysplasia in chronic ulcerative colitis.Aliment Pharmacol Ther. 2007; 25: 657-668
Article Info
Publication History
Published online: June 24, 2020
Accepted:
June 8,
2020
Received:
February 4,
2020
Footnotes
DISCLOSURE: The following authors disclosed financial relationships: B. Shen: Consultant for Abbvie and Takeda. M. Iacucci: Research support from NIHR Biomedical Research Centre, Birmingham, UK. U. Navaneethan: Consultant for Takeda, Abbvie, Janssen, and Pfizer. All other authors disclosed no financial relationships.
If you would like to chat with an author of this article, you may contact Dr Mohan at [email protected] .
Identification
Copyright
© 2021 by the American Society for Gastrointestinal Endoscopy
34512-0&id=gr1.jpg){kind=link}
34512-0&id=gr2.jpg){kind=link}
34512-0&id=gr3.jpg){kind=link}
34512-0&id=fx1.jpg){kind=link}
34512-0&id=fx2.jpg){kind=link}
34512-0&id=fx3.jpg){kind=link}
34512-0&id=fx4.jpg){kind=link}
34512-0&id=fx5.jpg){kind=link}
34512-0&id=fx6.jpg){kind=link}
34512-0&id=fx7.jpg){kind=link}
34512-0&id=fx8.jpg){kind=link}
34512-0&id=fx9.jpg){kind=link}