Abstract
Background
Most post-colonoscopy colorectal cancers (PC-CRCs) are thought to develop from missed or incompletely resected adenomas.
Aims
We aimed to assess the incidence rate of PC-CRC overall and per colorectal segment, as a proxy for PC-CRC due to incomplete adenoma resection, and to identify adenoma characteristics associated with these PC-CRCs.
Methods
We performed a nationwide, population-based cohort study, including all patients with a first colorectal adenoma between 2000–2010 in the Dutch Pathology Registry (PALGA). Outcomes were the incidence rate of PC-CRC overall and of PC-CRC in the same colorectal segment, occurring between six months and five years after adenoma resection. A multivariable Cox proportional hazard analysis was performed to identify factors associated with PC-CRCs in the same segment.
Results
We included 107,744 patients (mean age 63.4 years; 53.6% male). PC-CRC was detected in 1031 patients (0.96%) with an incidence rate of 1.88 per 1000 person years. PC-CRC in the same segment was found in 323 of 133,519 adenomas (0.24%) with an incidence rate of 0.56 per 1000 years of follow-up. High-grade dysplasia (hazard ratio (HR) 2.54, 95% confidence interval (CI) 1.99–3.25) and both villous (HR 2.63, 95% CI 1.79–3.87) and tubulovillous histology (HR 1.80, 95% CI 1.43–2.27) were risk factors for PC-CRC in the same segment.
Conclusions
Approximately one-third of PC-CRCs are found in the same colorectal segment after adenoma resection and could therefore be a consequence of incomplete adenoma resection, occurring in one in 400 adenomas. The risk of PC-CRC in the same segment is increased in adenomas with high-grade dysplasia or (tubulo)villous histology.
Keywords: Post-colonoscopy colorectal cancer, advanced adenoma, endoscopic resection, high grade dysplasia, villous histology
Introduction
Resection of adenomas during colonoscopy effectively reduces the incidence of subsequent colorectal cancer (CRC),1 but does not preclude its occurrence within the first three to five years after the procedure.2,3 Rather than developing from rapidly growing new lesions, most post-colonoscopy CRCs (PC-CRCs) are thought to be a consequence of missed lesions, but also incompletely resected adenomas.4–6
There is a considerable risk of colorectal adenomas being incompletely resected, especially when larger lesions are removed in piecemeal fashion.7,8 The proportion of PC-CRCs attributed to incompletely resected adenomas has been reported to range from 9–50% in previous studies,5,6,9 suggesting that a substantial number of PC-CRCs could have been prevented by improving the quality of endoscopic resection. Data in these studies were either derived from prevalent CRC cases, retrospectively identifying PC-CRC, or from cohorts of adenoma patients, prospectively identifying small numbers of PC-CRCs. In order to assess the absolute risk of developing PC-CRC after adenoma resection and to identify high-risk adenomas, a large cohort of adenoma patients with a substantial number of PC-CRCs is needed.
We performed a nationwide, population-based cohort study to determine the rate of PC-CRC overall and in the same colorectal segment, to estimate the incidence of PC-CRC due to incomplete adenoma resection and to identify adenoma characteristics associated with PC-CRC due to incomplete adenoma resection.
Methods
Data source and study population
We performed a population-based cohort study using data from the nationwide network and registry of histopathology and cytopathology diagnoses in the Netherlands using the Dutch Pathology Registry (PALGA). In this registry, summaries of pathology reports generated by all pathology departments in the Netherlands are centrally archived.10 The registry has accomplished nationwide coverage since 1991. Each pathology report in PALGA is linked to a diagnostic code, in line with the Systemized Nomenclature of Medicine (SNOMED).11
We conducted a search in PALGA and identified all patients with a first colorectal adenoma between 1 January 2000 and 31 December 2009. We did not include patients without adenomas, as our aim was to determine the rate of PC-CRC due to incomplete resection, not to compare the rate of PC-CRC between patients with and without adenomas at baseline. Inclusion of patients was based on the combination of diagnostic codes for ‘benign tumour’ and ‘colon’ or ‘rectum’. For each patient, we retrieved all histopathology reports of specimens of colorectal origin, starting at the first report and ending on 1 September 2013. Every report comprised the date of resection, the summary text and diagnostic codes. Age and gender of the patient were available for each PALGA report.
Data extraction and definitions
We excluded patients in whom no adenoma was found in the summary text or diagnostic codes. These included patients with a benign tumour that was not reported to be an adenoma or to contain dysplasia, for example hyperplastic polyps. In addition, patients with prevalent CRC, defined as patients in whom a CRC was found before, at the same time, or within six months after the first adenoma, were excluded. It was assumed that CRCs diagnosed within six months of the index procedure had already been detected or suspected during the index colonoscopy.3,12
Pathological features consisted of tubular adenoma (TA), tubulovillous adenoma (TVA), villous adenoma (VA), sessile serrated adenoma/polyp (SSA/P), low-grade dysplasia (LGD), high-grade dysplasia (HGD; including carcinoma in situ) and CRC. Adenomas with villous histology and/or high-grade dysplasia were considered advanced adenomas.
Locations were categorized into six segments of the colon: caecum; ascending colon (including hepatic flexure); transverse colon (including splenic flexure); descending colon; sigmoid colon and rectum. We extracted locations of adenomas and carcinomas through either the reported specific colon segment or the type of surgery (for CRC). Only if the colon segment or type of surgery were not mentioned, the reported distance from the anal verge was used, according to the following definitions: rectum as 0–15 cm from the anal verge; sigmoid colon as 16–35 cm; descending colon as 36–50 cm and transverse colon as 51–70 cm. When the distance was >70 cm from the anal verge, the location was considered to be the ascending colon, unless the caecum was mentioned. Using this approach we assumed that the readily identifiable caecum would have been mentioned as such in the pathology report.
The number of adenomas and the locations of the adenomas in each pathology report were extracted from summary texts and diagnostic codes using automated searches, counting the number of terms and synonyms for adenomas and locations separately. As a consequence, cumulative numbers of adenomas did not always correspond to cumulative numbers of locations. For example, if a report mentioned ‘adenoma’ or ‘adenomas’ once and synonyms for location (i.e. ‘ascending’ and ‘transverse’) twice, a discrepancy existed between the number of adenomas and the number of locations. The discrepancy was 5% for the total number of adenomas in the entire study population. In case of metachronous CRC, all pathology reports of the patient were hand-checked to prevent false positive outcomes.
Outcome measures
Outcome measures were the incidence rate of PC-CRC, defined as a CRC occurring between six months and five years after colonoscopy with adenoma removal, and the risk of PC-CRC after removal of an adenoma from the same colorectal segment, used as a proxy for PC-CRC due to incomplete resection.
Statistical analysis
We used SAS version 9.3 (SAS Institute, Cary, North Carolina, USA) for data management and performed data analysis using Statistical Packages for Social Sciences version 22 (IBM Corp., Armonk, New York, USA).
We performed per patient and per adenoma analyses. In the per patient analysis, patients were followed from the date of their first adenoma until CRC occurred anywhere in the colon. If no CRC occurred, they were followed until five years after the last adenoma was detected, or until the end of study (1 September 2013), whichever came first.
In the per adenoma analysis, we included only adenomas with a known location, both those occurring at baseline and those occurring during follow up. Each adenoma in a unique colon segment was included separately. Adenomas occurring in a segment where an adenoma was found previously were also included separately. Follow up of the previous adenoma in that segment was stopped at the date of detection of a new adenoma. Adenomas occurring simultaneously in one segment were not included separately. The per adenoma analysis is therefore a per segment analysis, including all segments in which at least one adenoma was detected.
Segments with at least one adenoma were followed until CRC was detected anywhere in the colon or an adenoma was detected in the same segment, whichever came first. A new adenoma in the same segment was not an outcome, but rather a new start of follow up. If neither an adenoma in the same segment nor CRC were found, adenomas were followed until five years after detection or until the end of study (1 September 2013), whichever came first. Adenomas with a follow-up of less than six months were excluded.
We performed a multivariable Cox proportional hazard regression analysis to identify adenoma related risk factors for PC-CRC in the same colorectal segment. Multiple adenomas within one patient were included as separate cases. Therefore, we performed the analysis with and without adjusting for patient related factors and included the timing of detecting the adenoma, i.e. at baseline or during follow up, as a separate variable. Results were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). We considered a p-value < 0.05 to be significant.
To overcome the bias of failed biopsy detection, meaning that CRC was suspected already at baseline, but biopsy failed to confirm the histologic diagnosis, we performed the same multivariable analysis, excluding all CRCs occurring within one year after the previous adenoma.
Ethical considerations
This study was performed with the approval of, and in accordance with, the privacy and ethical guidelines of the privacy committee of PALGA. The Medical Ethical Committee of the University Medical Center Utrecht, the Netherlands, exempted the study from informed consent according to Dutch regulations.
Results
Per patient
In total, 121,378 patients with a benign colorectal tumour were identified in PALGA, of whom 119,233 with one or more adenomas were included. After exclusion of patients with CRC diagnosed before or at the same time as the first adenoma (n = 7222) and patients with CRC within six months of the first adenoma (n = 4267), a total of 107,744 patients were included in the final analysis.
Mean age of the included patients was 63.4 years (standard deviation (SD) 12.8) and 53.6% were male (Table 1). Mean number of resected adenomas at the time of first adenoma detection was 1.23 (SD 0.57), with a total number of 132.974 adenomas at baseline. The sigmoid colon (39.7%) and rectum (27.6%) were the most frequently reported locations. Villous features, i.e. tubulovillous or villous histology, were reported at baseline in 31.4% of patients, while high-grade dysplasia was present in 13.8% of patients. The incidence of SSA/Ps was 2.4%. We observed an increase in the incidence of first adenomas in the Netherlands from 5798 in the year 2000 to 19,039 adenomas in 2009.
Table 1.
Baseline characteristics
| Total number of patients | 107,744 (%) |
|---|---|
| Male gender | 57,784 (53.6) |
| Mean age at first adenoma, years (SD) | 63.4 (12.8) |
| Mean number of adenomas per patient (SD) | 1.23 (0.57) |
| Number of adenomas reported at baseline | 132,974 |
| 1 | 87,742 (81.4) |
| 2 | 16,327 (15.2) |
| ≥3 | 3675 (3.4) |
| Locations of adenoma(s) reported at baselinea | 139,871 |
| Caecum | 9539 (8.9) |
| Ascending colon | 13,049 (12.1) |
| Transverse colon | 10,530 (9.8) |
| Descending colon | 10,424 (9.7) |
| Sigmoid colon | 42,812 (39.7) |
| Rectum | 29,710 (27.6) |
| Colon not otherwise specified | 23,807 (22.1) |
| Non tubular histology of first adenoma(s) | |
| Tubulovillous | 30,295 (28.1) |
| Villous | 3527 (3.3) |
| Sessile serrated adenoma/polyp | 2619 (2.4) |
| High-grade dysplasia in first adenoma(s) | 14,900 (13.8) |
| Year of first adenoma diagnosis | |
| 2000–2003 | 25,621 (23.8) |
| 2004–2006 | 32,251 (29.9) |
| 2007–2009 | 49,872 (46.3) |
SD: standard deviation.
Sum can be > 100%, as more than one adenoma can be found at index colonoscopy.
A cumulative number of 1.44 (SD 0.94) adenomas per patient was found at baseline and during follow up, adding up to a total number of 155,239 adenomas. Locations were known for 133,519 of these adenomas (86.0%), of which 89.2% were found at baseline and 10.8% during follow up. The majority of the adenomas with known locations were in the sigmoid colon or rectum (59.7%) (Table 2). Villous features were found in 32.0% and high-grade dysplasia in 13.6% of adenomas.
Table 2.
Characteristics of adenomas with known location, including first adenomas and adenomas during follow up
| Number of adenomas | 133,519 (%) |
|---|---|
| Number of adenomas in the same segment | |
| 1 | 122,802 (92.0) |
| 2 | 9668 (7.2) |
| ≥3 | 1049 (0.8) |
| Adenoma location | |
| Caecum | 11,601 (8.7) |
| Ascending colon | 16,614 (12.4) |
| Transverse colon | 13,308 (10.0) |
| Descending colon | 12,294 (9.2) |
| Sigmoid colon | 46,747 (35.0) |
| Rectum | 32,955 (24.7) |
| Non-tubular adenoma histology | |
| Tubulovillous | 38,194 (28.6) |
| Villous | 4530 (3.4) |
| Sessile serrated adenoma/polyp | 4665 (3.5) |
| High-grade dysplasia | 18,138 (13.6) |
| Year of adenoma diagnosis | |
| 2000–2003 | 24,546 (18.4) |
| 2004–2007 | 34,395 (25.8) |
| 2007–2009 | 58,157 (43.6) |
| 2010–2013 | 16,421 (12.3) |
| Timing of adenoma within one patient | |
| First adenoma | 119,129 (89.8) |
| Adenoma during follow-up | 14,390 (10.2) |
PC-CRC, per patient analysis
In per patient analysis, the risk of developing PC-CRC anywhere in the colon was 0.96% (1031 of 107,744 patients). Mean follow-up time per patient was 5.1 years (SD 1.2) and the incidence rate of PC-CRC was 1.88 per 1000 person years (Table 3). PC-CRC was found in the same colorectal segment after adenoma resection in 323 (0.30%) patients, corresponding to an incidence rate of 0.59 per 1000 person years.
Table 3.
Rates of post-colonoscopy colorectal cancer (PC-CRC)
| Per patient analysis | |
| Total number of patients | 107,744 |
| PC-CRC (%) | 1031 (0.96%) |
| PC-CRC in the same colorectal segment (%) | 323 (0.30%) |
| Mean follow-up per patient, years (SD) | 5.08 (1.19) |
| Incidence rate of PC-CRC | 1.88 per 1000 person years |
| Incidence rate of PC-CRC in the same segment | 0.59 per 1000 person years |
| Per adenoma analysis | |
| Total number of adenomas | 133,519 |
| PC-CRC in the same segment (%) | 323 (0.24%) |
| Mean follow-up per adenoma, years (±SD) | 4.35 (1.11) |
| Incidence rate of PC-CRC in the same segment | 0.56 per 1000 years of follow up |
| Distribution of occurrence of PC-CRCs in the same segment, according to time after adenoma resection | |
| Between 6 months and one year | 84 (26.0%) |
| Second year | 65 (20.1%) |
| Third year | 61 (18.9%) |
| Fourth year | 58 (18.0%) |
| Fifth year | 55 (17.0%) |
SD: standard deviation.
PC-CRC in the same colorectal segment, per adenoma analysis
In per adenoma analysis, PC-CRC was detected in the same segment within five years in 323 (0.24%) of 133,519 adenomas. Mean follow-up per adenoma was 4.4 years (SD 1.1), and the incidence rate of PC-CRC due in the same colorectal segment was 0.56 per 1000 years of follow-up (Table 3). The number of PC-CRCs in the same segment decreased from 84 in the first year after resection to 55 in the fifth year after resection. PC-CRC in the same segment occurred in 123 (0.58%) of 21,257 resected advanced adenomas.
Comparison of locations of PC-CRCs overall with PC-CRCs in the same segment shows that PC-CRCs overall were more often located in the proximal colon (47.5% versus 32.5%) (Table 4). PC-CRCs in the same segment were more frequently found in the rectum and sigmoid (62.5%) compared to PC-CRCs overall (48.3%). The distribution of PC-CRCs in the same segment was highly comparable to the distribution of all adenomas found at baseline.
Table 4.
Comparison of locations of index adenomas, post-colonoscopy colorectal cancer (PC-CRC) overall and PC-CRC in the same colorectal segment
| Index adenoma (%)a | PC-CRC overallb (%) | PC-CRC in the same segment (%) | |
|---|---|---|---|
| Location | n = 139,871 | n = 826 | n = 323 |
| Caecum | 9539 (8.9) | 144 (17.4) | 31 (9.6) |
| Ascending colon | 13,049 (12.1) | 179 (21.7) | 46 (14.2) |
| Transverse colon | 10,530 (9.8) | 70 (8.5) | 28 (8.7) |
| Descending colon | 10,424 (9.7) | 34 (4.1) | 16 (5.0) |
| Sigmoid | 42,812 (39.7) | 168 (20.3) | 108 (33.4) |
| Rectum | 29,710 (27.6) | 231 (28.0) | 94 (29.1) |
Only adenomas with known location.
Only PC-CRCs with known location.
In multivariable analysis, high-grade dysplasia (HR 2.57, 95% CI 2.02–3.26) and a villous adenoma (HR 2.69, 95% CI 1.85–3.91) or tubulovillous adenoma (HR 1.74, 95% CI 1.93–2.19) were independently associated with PC-CRC in the same segment (Table 5). Compared to adenomas in the rectum, adenomas in the descending colon were less likely to lead to PC-CRC in the same segment. We found no association with PC-CRC in the same segment for sessile serrated histology, year of adenoma resection or number of synchronous adenomas in the one colon segment. In addition, no difference was found in the risk of CRC in the same segment after resection of adenomas detected at baseline versus adenomas detected during follow up. The addition of patients’ age (HR per year increase 1.03, 95% CI 1.02–1.04) and gender (HR 0.99, 95% CI 0.79–1.23) to the multivariable analysis did neither affect significance nor effect size of adenoma-related risk factors.
Table 5.
Hazard ratios (HRs) and 95% confidence intervals (CIs) for post-colonoscopy colorectal cancer (PC-CRC) due to incomplete adenoma resection
| Adenomas leading to PC-CRC due to incomplete resection (%) | Other adenomas (%) | Adjusted |
||
|---|---|---|---|---|
| n = 323 | n = 133,196 | HR | 95% CI | |
| Number of adenomas in the same segment | ||||
| 1 | 290 (89.8) | 122,512 (92.0) | Reference | Reference |
| 2 | 28 (8.7) | 9640 (7.2) | 1.20 | 0.81–1.77 |
| 3 or more | 5 (1.5) | 1044 (0.8) | 1.84 | 0.76–4.45 |
| Adenoma location | ||||
| Rectum | 94 (29.1) | 32,861 (24.7) | Reference | Reference |
| Sigmoid | 108 (33.4) | 46,639 (35.0) | 0.82 | 0.62–1.09 |
| Descending colon | 16 (5.0) | 12,278 (9.2) | 0.52 | 0.31–0.88 |
| Transverse colon | 28 (8.7) | 13,280 (10.0) | 0.92 | 0.61–1.41 |
| Ascending colon | 46 (14.2) | 16,568 (12.4) | 1.18 | 0.83–1.70 |
| Caecum | 31 (9.6) | 11,570 (8.7) | 1.09 | 0.72–1.63 |
| Histology | ||||
| Tubulovillous adenoma | 141 (43.7) | 38,053 (28.6) | 1.74 | 1.38–2.19 |
| Villous adenoma | 32 (9.9) | 4498 (3.4) | 2.69 | 1.85–3.91 |
| SSA/P | 10 (3.1) | 4655 (3.5) | 1.15 | 0.61–2.18 |
| High-grade dysplasia | 106 (32.8) | 18,032 (13.5) | 2.57 | 2.02–3.26 |
| Year of adenoma diagnosis | ||||
| 2000–2003 | 59 (18.3) | 24,487 (18.4.0) | 1.21 | 0.45–3.24 |
| 2004–2006 | 105 (32.5) | 34,290 (25.7) | 1.57 | 0.60–4.14 |
| 2007–2009 | 147 (45.5) | 58,010 (43.6) | 1.40 | 0.53–3.65 |
| 2010–2013 | 12 (3.7) | 16,409 (12.3) | Reference | Reference |
| Adenoma during follow up (vs first adenoma) | 10 (3.1) | 13,671 (10.3) | 0.90 | 0.32–2.54 |
SSA/P: sessile serrated adenoma/polyp.
We also performed a multivariable analysis excluding the PC-CRC cases within one year, to overcome the bias of failed biopsy detection. This resulted in a reduction in effect size of histologic parameters. The HRs for tubulovillous adenomas, villous adenomas and adenomas with HGD were 1.40 (95% CI 1.07–1.84), 1.77 (95% CI 1.05–2.96) and 2.25 (95% CI 1.69–3.01), respectively. The simultaneous occurrence of 3 or more adenomas significantly increased the risk of PC-CRC in the same colorectal segment (HR 2.57, 95% CI 1.06–6.27), while the reduced risk for adenomas in the descending colon was no longer significant (HR 0.57, 95% CI 0.31–1.05). No other differences in significance or effect size were observed.
Discussion
In this nationwide, population-based cohort study, we found that the overall incidence rate of PC-CRC within five years after adenoma resection was 1.88 per 1000 person years. PC-CRC in the same colorectal segment, probably due to to incomplete adenoma resection, occurred in 0.24% of adenomas with an incidence rate of 0.56 per 1000 years of follow-up. Villous features and high-grade dysplasia were associated with an increased risk of PC-CRC in the same colorectal segment.
In recent years, accumulating evidence has shown that most PC-CRCs are not due to new, rapidly growing and biologically more aggressive tumours,12–14 but rather to missed or incompletely resected lesions.6 Incidence rates of PC-CRC overall and PC-CRC attributed to incomplete adenoma resection in three previous prospective studies varied from 1.7–2.9 and 0.7–1.5 per 1000 person years, respectively.4,5,9 The percentage of PC-CRCs due to incomplete resection was estimated to be 31–50%, which is comparable to our study. The small number of PC-CRCs in the abovementioned studies, in which the same proxy for PC-CRC due to incomplete adenoma resection was used as in our study, precludes identification of potential risk factors. The overall incidence rate of PC-CRC in the studies of Huang et al. and Pabby et al. was higher than in our study (1.88 cases per 1000 person years).4,5 Both studies included patients with a higher risk (i.e. more previous and simultaneous adenomas with more advanced features) compared to our study. Another explanation may be that the data in our study are derived from daily clinical practice, in which follow-up has been less strict than in controlled studies. Robertson et al. also included studies with higher risk patients under close surveillance, but interventions in two of these three studies reduced the risk of PC-CRC.9
Le Clercq et al. performed a population-based study including all CRCs occurring in the South-Limburg region of the Netherlands from 2001–2010.6 The authors concluded that 2.9% of all CRCs were PC-CRCs, defined as CRCs diagnosed within five years after index colonoscopy. Nine percent of these PC-CRCs could be attributed to incomplete adenoma resection. However, identification of PC-CRCs from prevalent CRC cases precludes calculating the absolute risk of developing PC-CRC after (incomplete) adenoma resection.
It is not surprising that adenomas with villous features and high-grade dysplasia were associated with an increased risk of PC-CRC due to incomplete adenoma resection. These advanced pathological features have already previously been associated with an increased risk of adenoma recurrence and subsequent development of advanced adenomas or CRC.15,16 Advanced adenomas are often large and resected in piecemeal fashion, which has been associated with an increased risk of local recurrence.7 Information regarding adenoma size was unfortunately not available in our study, precluding adjustment of the results for size. It is probable that such an adjustment would have led to a reduction in effect size of pathological risk factors.
Proximal location is repeatedly reported to be a risk factor for PC-CRC,2,3,12 presumably because lesions in the proximal colon are more prone to be missed during colonoscopy. In our study, proximal location of the resected adenoma was not associated with PC-CRC in the same colorectal segment. We do not have data on patients with negative colonoscopies or neoplastic lesions that may have been missed during colonoscopy. Neither Pabby et al. nor Le Clercq et al. found an association between proximal adenoma location and PC-CRC due to incomplete adenoma resection.5,6 We found a protective effect of location in the descending colon as compared to the rectum. This could be explained by the presence of more large and non-pedunculated adenomas in the rectum.
Although SSA/Ps have been reported to be more frequently incompletely resected than conventional adenomas,8 no association was found between SSA/Ps and PC-CRC due to incomplete resection. Knowledge about SSA/Ps and their malignant potential has largely increased in the last decade. It may therefore well be that a substantial proportion of SSA/Ps may have been missed during the early years of this study, rather than having been (incompletely) resected.
Our study is the first to estimate the absolute risk of PC-CRC due to incomplete adenoma resection in a large and unselected cohort of adenoma patients. The major strength of this study is its population-based approach, which reduces possible selection bias. Furthermore, we used a data source with national coverage and excellent accuracy, which allows a pseudo-prospective follow up of patients over time making it unlikely that incident CRC cases were missed.
Attention to quality of colonoscopy tends to focus on adenoma detection,17 but CRC prevention programs will only be fully effective in case of adequate polypectomy. Our study underlines the significance of PC-CRC due to incomplete adenoma resection, thereby increasing awareness and supporting the ongoing evolution of resection techniques to decrease recurrence risk by improving quality of polypectomy.18
The most important limitation of our study is the uncertainty that PC-CRC occurring in the same colon segment is indeed a direct consequence of incomplete adenoma resection and not the result of a missed lesion. As it is practically impossible to identify the exact polypectomy site, unless marking techniques are routinely applied, we used PC-CRC in the same colorectal segment as a proxy for PC-CRC due to incomplete adenoma resection, as previous studies did.4–6 Localization based on depth of insertion could be inaccurate, but this method was only used if the segment itself was not mentioned in the pathology report, which was the case in less than 25% of adenomas.
In an attempt to estimate whether PC-CRC in the same segment was not mainly a consequence of missed lesions, we compared the locations of all PC-CRCs to the locations of PC-CRCs in the same colon segment. We found that 39% of all PC-CRCs were in the caecum and ascending colon vs 24% of PC-CRCs in the same segment, probably reflecting the association of proximal location with PC-CRC due to missed lesions. PC-CRCs in the same segment were more often located in the sigmoid or rectum (63%) than PC-CRCs anywhere (48%), probably reflecting the higher risk of incomplete resection for advanced lesions found more often in the sigmoid or rectum. In addition, even after exclusion of PC-CRC occurring within one year, advanced features were independently associated with PC-CRC in the same segment, which is another argument for incomplete resection rather than missed lesions causing PC-CRC in the same segment.
We had no information on endoscopic data that might have affected the risk of PC-CRC. The endoscopic morphology of the removed adenomas, i.e. size and sessile versus pedunculated morphology, resection technique, i.e. cold snaring, snare coagulation or (piecemeal) endoscopic mucosal resection, and quality indicators of the colonoscopy, such as caecal intubation and bowel preparation quality, were not available.
As approximately one-third of all PC-CRCs occurs in the same segment after adenoma resection, it is likely that incomplete adenoma resection contributes significantly to the risk of PC-CRC. PC-CRCs in the same segment occur after one in 400 adenomas (0.24%) and one in 170 advanced adenomas (0.58%). The results of our nationwide cohort study support current guidelines, recommending intensified follow-up after resection of adenomas with villous features and/or high-grade dysplasia.19,20
Acknowledgements
The authors would like to thank Esther van den Broek from PALGA for her help in successfully acquiring the relevant data.
Declaration of conflicting interests
None declared.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References
- 1.Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993; 329: 1977–1981. [DOI] [PubMed] [Google Scholar]
- 2.Bressler B, Paszat LF, Chen Z, et al. Rates of new or missed colorectal cancers after colonoscopy and their risk factors: A population-based analysis. Gastroenterology 2007; 132: 96–102. [DOI] [PubMed] [Google Scholar]
- 3.Singh H, Nugent Z, Demers AA, et al. Rate and predictors of early/missed colorectal cancers after colonoscopy in Manitoba: A population-based study. Am J Gastroenterol 2010; 105: 2588–2596. [DOI] [PubMed] [Google Scholar]
- 4.Huang Y, Gong W, Su B, et al. Risk and cause of interval colorectal cancer after colonoscopic polypectomy. Digestion 2012; 86: 148–154. [DOI] [PubMed] [Google Scholar]
- 5.Pabby A, Schoen RE, Weissfeld JL, et al. Analysis of colorectal cancer occurrence during surveillance colonoscopy in the dietary Polyp Prevention Trial. Gastrointest Endosc 2005; 61: 385–391. [DOI] [PubMed] [Google Scholar]
- 6.Le Clercq CM, Bouwens MW, Rondagh EJ, et al. Postcolonoscopy colorectal cancers are preventable: A population-based study. Gut 2014; 63: 957–963. [DOI] [PubMed] [Google Scholar]
- 7.Belderbos TD, Leenders M, Moons LM, et al. Local recurrence after endoscopic mucosal resection of nonpedunculated colorectal lesions: Systematic review and meta-analysis. Endoscopy 2014; 46: 388–402. [DOI] [PubMed] [Google Scholar]
- 8.Pohl H, Srivastava A, Bensen SP, et al. Incomplete polyp resection during colonoscopy-results of the complete adenoma resection (CARE) study. Gastroenterology 2013; 144: 74–80. [DOI] [PubMed] [Google Scholar]
- 9.Robertson DJ, Greenberg ER, Beach M, et al. Colorectal cancer in patients under close colonoscopic surveillance. Gastroenterology 2005; 129: 34–41. [DOI] [PubMed] [Google Scholar]
- 10.Casparie M, Tiebosch AT, Burger G, et al. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol 2007; 29: 19–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Côté RA, Robboy S. Progress in medical information management. Systematized nomenclature of medicine (SNOMED). JAMA 1980; 243: 756–762. [DOI] [PubMed] [Google Scholar]
- 12.Pullens HJ, Leenders M, Schipper ME, et al. No decrease in the rate of early or missed colorectal cancers after colonoscopy with polypectomy over a 10-year period: A population-based analysis. Clin Gastroenterol Hepatol 2015; 13: 140–147. [DOI] [PubMed] [Google Scholar]
- 13.Erichsen R, Baron JA, Stoffel EM, et al. Characteristics and survival of interval and sporadic colorectal cancer patients: A nationwide population-based cohort study. Am J Gastroenterol 2013; 108: 1332–1340. [DOI] [PubMed] [Google Scholar]
- 14.Samadder NJ, Curtin K, Tuohy TM, et al. Characteristics of missed or interval colorectal cancer and patient survival: A population-based study. Gastroenterology 2014; 146: 950–960. [DOI] [PubMed] [Google Scholar]
- 15.Martínez ME, Baron JA, Lieberman DA, et al. A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy. Gastroenterology 2009; 136: 832–841. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.De Jonge V, Sint Nicolaas J, van Leerdam ME, et al. Systematic literature review and pooled analyses of risk factors for finding adenomas at surveillance colonoscopy. Endoscopy 2011; 43: 560–572. [DOI] [PubMed] [Google Scholar]
- 17.Dominitz JA, Spiegel B. Editorial: On the quality of quality metrics: Rethinking what defines a good colonoscopy. Am J Gastroenterol 2016; 111: 730–732. [DOI] [PubMed] [Google Scholar]
- 18.Moss A, Williams SJ, Hourigan LF, et al. Long-term adenoma recurrence following wide-field endoscopic mucosal resection (WF-EMR) for advanced colonic mucosal neoplasia is infrequent: Results and risk factors in 1000 cases from the Australian Colonic EMR (ACE) study. Gut 2015; 64: 57–65. [DOI] [PubMed] [Google Scholar]
- 19.Winawer SJ, Zauber AG, Fletcher RH, et al. American Cancer Society. Guidelines for colonoscopy surveillance after polypectomy: A consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. Gastroenterology 2006; 130: 1872–1885. [DOI] [PubMed] [Google Scholar]
- 20.Centraal Begeleidingsorgaan (CBO). Dutch guideline colonoscopy surveillance. [Nederlandse richtlijn colonoscopie surveillance.] 2013.