If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Patients with previous colorectal adenomas are at increased risk of colorectal cancer. Current guidelines for postpolypectomy surveillance intervals treat all tubular adenomas 1 to 9 mm in size with low-grade dysplasia as carrying the same level of risk. We evaluated whether 6 to 9 mm adenomas detected at colonoscopy are associated with greater risk of advanced neoplasia at follow-up compared with baseline 1 to 5 mm adenomas.
We retrospectively evaluated a colonoscopy database at a single U.S. academic center. Patients with baseline examinations demonstrating tubular adenomas 1 to 9 mm in size with low-grade dysplasia and no advanced adenomas were included. Follow-up colonoscopies were performed at least 200 days later and were assessed for incident advanced neoplasia (cancer, high-grade dysplasia, adenoma ≥10 mm in size, or villous elements).
There were 2477 qualifying baseline colonoscopies. The absolute risk of metachronous advanced neoplasia increased from 3.6% in patients with 1 to 5 mm adenomas to 6.9% in patients with at least 1 adenoma of 6 to 9 mm (P = .001). Patients with 5 or more adenomas 1 of which was at least 6 to 9 mm had the highest risk of advanced neoplasia at follow-up (10.4%, P = .006). When only screening colonoscopies were considered, all baseline groups (1-2 adenomas, 3-4 adenomas, ≥5 adenomas) with adenomas 6 to 9 mm in size had an increased risk for metachronous advanced neoplasia (odds ratio [OR], 4.07; 95% confidence interval [CI], 1.50-11.04; OR, 4.91; 95% CI, 1.44-16.75; OR, 4.71; 95% CI, 1.30-17.05, respectively).
Patients with baseline small (6-9 mm) adenomas have an increased risk of advanced lesions on follow-up compared with patients with only diminutive (1-5 mm) adenomas. Postpolypectomy guidelines should consider risk stratification based on small versus diminutive adenomas.
Patients with previous colorectal adenomas are at increased risk of colorectal cancer, although compared with the general population, the increased risk may be largely confined to those with advanced or multiple (3 or more) adenomas.
patients are considered at higher risk if they have had an advanced conventional adenoma (adenoma ≥10 mm, or with high-grade dysplasia or villous elements), or 3 or more adenomas. High-risk patients are recommended to undergo their next surveillance colonoscopy at a shorter interval compared with those who have only 1 or 2 low-risk adenomas (tubular adenomas ≤9 mm in size with only low-grade dysplasia).
Adenoma detection rates (ADRs) in clinical practice are increasing
As a result, tiny conventional adenomas are being detected in an increasingly large fraction of screening and surveillance populations. Patients with even tiny adenomas are assigned to shorter surveillance intervals, which results in increased cost, risks, and inconvenience to patients, with perhaps limited benefit in cancer protection. Further, the effectiveness of postpolypectomy surveillance is limited by marked variability in the quality of performance of the baseline colonoscopy.
Currently, postpolypectomy surveillance guidelines do not account for baseline performance or the ADR of examiners, and generally assume that baseline colonoscopy has uniform performance. Future postpolypectomy surveillance guidelines might be specified to only apply for doctors with ADRs above a certain threshold, or could vary recommended intervals according to the ADR. Longer surveillance intervals could be particularly appropriate for doctors with high ADRs and patients with only tiny adenomas.
For these reasons, we expect increasing pressure to expand surveillance intervals for patients with tiny adenomas. In this regard, in many postpolypectomy surveillance observational studies, low-risk adenomas include all tubular adenomas 1 to 9 mm in size with only low-grade dysplasia. However, adenoma size is a known predictor of subsequent risk, as has been repeatedly shown for adenomas ≥10 mm in size compared with 1 to 9 mm in size.
We hypothesized that adenomas 6 to 9 mm in size, often called “small adenomas,” could be associated with a greater risk of subsequent development of advanced neoplasia compared with diminutive adenomas (defined as 1-5 mm in size). Indeed, although most previous studies have lumped 1 to 5 mm and 6 to 9 mm adenomas together, some studies have suggested that small adenomas are associated with a higher risk of subsequent advanced neoplasia than are diminutive adenomas.
In this report, we describe our experience with the risk of subsequent advanced neoplasia in patients with small versus only diminutive conventional adenomas.
A database of colonoscopies conducted at a single center (Indiana University Hospital and an associated outpatient endoscopy center) from 1999 to 2016 was used. The database is updated periodically with procedure information, and we have reported previously on surveillance findings collected from this database.
Briefly, the database contains patient demographics, polyp findings, including histology and size, and procedure characteristics, including indication. Completion of the procedure and bowel preparation quality were added to the database in 2012. Permission to review the database for the current study was obtained from the Institutional Review Board at Indiana University on November 21, 2018. The database at this time consisted of procedures from January 1999 to June 2016.
We identified patients with only diminutive or small adenomas (<10 mm) and categorized them into either having diminutive adenomas or small adenomas and then into 6 groups: 1 to 2 adenomas both ≤5 mm, 1 to 2 adenomas with 1 at least 6 to 9 mm, 3 to 4 adenomas all ≤5 mm, 3 to 4 adenomas with 1 at least 6 to 9 mm, 5 or more adenomas all ≤5 mm, 5 or more adenomas with 1 at least 6 to 9 mm. Polyp size was estimated by the endoscopist at the time of the procedure. Patients with concomitant serrated lesions at baseline were not excluded.
We excluded patients with inflammatory bowel disease, colon cancer syndromes (familial adenomatous polyposis, hereditary nonpolyposis colorectal cancer, serrated polyposis syndrome), advanced adenomas (≥10 mm tubular adenoma, villous component, high-grade dysplasia), or cancer at or before the baseline colonoscopy. We only included patients who had at least 1 follow-up examination at our institution at least 200 days later than the baseline colonoscopy.
Advanced neoplasia at follow-up was defined as finding a conventional adenoma that was 10 mm or larger, with high-grade dysplasia or villous elements or cancer. We also included sensitivity analyses with an expanded definition of advanced neoplasia that considered a sessile serrated polyp (SSP) with cytological dysplasia or SSP ≥10 mm along with the above-mentioned criteria as the definition for advanced neoplasia. We also performed subgroup analyses first limiting the sample to screening patients and then limiting the sample to patients who had 1 to 2 adenomas with a follow-up interval between 4.5 and 5.5 years.
We report demographics, interval to follow-up colonoscopy, and advanced neoplasia at follow-up for each group. Analysis of variance was used for continuous variables and chi-squared tests were used for categorical variables. We then assessed the effects of age, gender, bowel preparation quality, indication for baseline colonoscopy, time to follow-up, and our baseline adenoma categories for risk of advanced neoplasia at follow-up using a logistic regression model. We report odds ratios (ORs) with 95% confidence intervals (CIs). All analyses were performed with IBM SPSS Statistics, Version 25 (IBM Corp, Armonk, NY, USA).
Among 34,467 patients who underwent procedures from January 1999 to June 2016, we identified 2477 patients who had diminutive or small adenomas at baseline colonoscopy, who never had advanced neoplasia before this examination, and underwent a follow-up examination at least 200 days later at 1 of our institutions (Fig. 1). The largest fraction of the baseline cohort had only 1 or 2 diminutive adenomas (66%). The absolute risk of advanced neoplasia at follow-up was 3.6% among patients with only diminutive adenomas compared with 6.9% among patients with at least 1 small adenoma (71 of 1997 vs 33 of 480, P = .001) (Table 1). We performed 2 separate logistic regression analyses using different adenoma subgroups (Table 2). The first logistic regression indicated that age and the presence of a small adenoma were risk factors for metachronous advanced neoplasia, whereas gender and time to follow-up colonoscopy were not (Table 2). The fraction of baseline bowel preparations that were excellent or good was 89.2%. In the multivariable analysis, excellent or good preparation at baseline was not associated with advanced neoplasia at follow-up colonoscopy (OR, 0.65; 95% CI, 0.32-1.31; P = .229; Table 2).
Table 1Demographic features and advanced neoplasia at follow-up in patients with diminutive adenomas (1-5 mm) only and patients with at least 1 small adenoma (6-9 mm) at baseline colonoscopy
Patients with diminutive adenomas only at baseline colonoscopy
Patients with at least 1 small adenoma at baseline colonoscopy
Results from the second logistic regression. Results for age, gender, time to follow-up, bowel preparation quality, and indication were nearly identical to the first analysis. Therefore, only the results for adenoma subgroups are shown for the second analysis.
1-2 adenomas both ≤5 mm
1-2 adenomas one 6-9 mm
3-4 adenomas all ≤5 mm
3-4 adenomas one 6-9 mm
≥5 adenomas all ≤ 5 mm
≥5 adenomas one 6-9 mm
∗ Logistic regression using the enter method.
† Results from the first logistic regression.
‡ Results from the second logistic regression. Results for age, gender, time to follow-up, bowel preparation quality, and indication were nearly identical to the first analysis. Therefore, only the results for adenoma subgroups are shown for the second analysis.
We performed a second logistic regression using more adenoma subgroups (Table 2). In the second analysis, there was only minimal difference in the results for age, gender, years to follow-up, bowel preparation quality, or indication from the first analysis, and the results are only shown for these factors for the first logistic regression in Table 2. The second logistic regression indicated that age (results not shown) and the subcategories of patients with at least 3 adenomas 1 of which was ≥6 mm in size predicted advanced neoplasia at follow-up colonoscopy (Table 2).
The absolute risk of metachronous advanced neoplasia was 10.4% (11 of 106) among patients with ≥5 adenomas at least 1 of which was a small adenoma (Table 3). Patients who only had 1 to 2 diminutive adenomas had an absolute risk of 3.3% (54 of 1625) for advanced neoplasia at follow-up.
Table 3Advanced lesions at follow-up according to the baseline colonoscopy findings
Tables 4 and 5 show findings at follow-up colonoscopy according to baseline screening colonoscopies only. When the analysis was confined to baseline screening examinations only, multivariable analyses showed that all subgroups with adenomas 6 to 9 mm had a higher incidence of advanced neoplasia at follow-up compared with 1 to 2 adenomas <6 mm in size, regardless of the total number of adenomas (Table 5).
Table 4Advanced lesions among 6 baseline groups limited to screening indication only
Tables 6 and 7 show an analysis limited to patients with 1 to 2 baseline adenomas and who had a follow-up examination between 4.5 and 5.5 years. Patients with a small adenoma had a significantly higher risk of advanced neoplasia at follow-up compared with patients with diminutive adenomas only (OR, 5.23; 95% CI, 1.55-17.68).
Table 6Advanced lesions at follow-up according to baseline colonoscopy findings limited to patients with 1-2 adenomas at baseline whose follow-up colonoscopy was at 5 ± 0.5 years
In this study, we describe experience from a single U.S. center of the impact of small versus diminutive adenomas at colonoscopy on the risk of advanced neoplasia at follow-up. Among patients with ≥5 adenomas, the risk for advanced neoplasia at follow-up was greater for patients who had at least 1 small adenoma compared with patients with all diminutive adenomas Thus, we found that adenomas in the 6 to 9 mm size range increase the risk of subsequent advanced neoplasia compared with diminutive adenomas. Our results suggest that caution is appropriate in lumping all adenomas of 1 to 9 mm together in postpolypectomy surveillance guidelines.
In addition to 6 to 9 mm adenomas predicting subsequent advanced neoplasia compared with having only diminutive adenomas in our entire study population, we also found that the discriminating effect of 6 to 9 mm polyps on the subsequent risk of advanced neoplasia was present when only baseline colonoscopies performed for screening were considered. Thus, when only baseline screening colonoscopies were considered, each of the 3 subgroups of adenomas organized by number of lesions (1-2 adenomas, 3-4 adenomas, and ≥5 adenomas) and containing at least 1 adenoma 6 to 9 mm in size had a significantly higher risk of advanced neoplasia at follow-up compared with the subgroup with the same number of adenomas composed of only diminutive adenomas. The predictive value of 6 to 9 mm adenomas in screening patients was more striking than the same effect for the entire study population. The explanation for this finding is uncertain, but one possibility is that patients undergoing baseline surveillance examinations had previous examinations at other centers that removed high-risk adenoma findings, but of which we were unaware and could not identify from our database. Such patients would remain at increased risk of advanced neoplasia despite an intervening surveillance examination at our center showing only low-risk adenomas.
However, 2 small studies that did stratify risk by small versus diminutive size of adenomas at baseline suggested an increased risk of subsequent advanced lesions in those with small baseline adenomas.
we suspect that our results have good generalizability. Our results may not apply in settings where ADRs are low because the absolute risk of advanced neoplasia could be substantially higher even in individuals with very low-risk findings (because more high-risk lesions are missed when ADRs are low). Many of the studies that underlie current postpolypectomy surveillance recommendations
were performed before the current emphasis on ADRs and before high-definition colonoscopes were available. Thus, the prevalence of patients with only tiny adenomas may be substantially different in the current study from earlier studies. Indeed, the difference in risk of subsequent advanced neoplasia observed in the current study between the baseline small versus diminutive adenoma groups may largely reflect the current study identifying a large and distinct cohort of very low-risk tiny adenomas that were less-frequently observed in earlier studies. This suggestion warrants additional evaluation.
Limitations of our study include that it is a retrospective analysis of a prospectively developed and maintained database. The number of patients in certain of the risk groups was small. Nevertheless, the results supporting the predictive value of adenomas in the 6 to 9 mm size range compared with the 1 to 5 mm size range reached statistical significance. ADRs in our unit are likely to be relatively high, which suggests that our findings will be increasingly relevant as detection improves in community-based colonoscopy. Another potential limitation might be that we did not choose to systematically exclude patients with serrated lesions at baseline. Further, we did not present data on the impact of concomitant SSPs at baseline, primarily because our pathology department has not consistently interpreted SSPs versus hyperplastic polyps over the study interval, as we have twice demonstrated.
but these effects are common to postpolypectomy surveillance studies generally.
In conclusion, these data indicate that lumping small and diminutive adenomas together in postpolypectomy surveillance guidelines may create a risk for some patients with 6 to 9 mm adenomas and/or increase procedure-related costs and risks for patients with 1 to 5 mm adenomas. We recommend that additional groups evaluate the impact of small versus diminutive adenomas on the subsequent risk of advanced neoplasia in postpolypectomy surveillance cohorts.
This work was supported by a gift from Scott Schurz of Bloomington, Indiana, and his children to the Indiana University Foundation in the name of Douglas K. Rex, MD.
Supplementary Table 1Advanced neoplasia according to baseline findings and including sessile serrated polyps (SSPs) with cytological dysplasia and SSPs ≥10 mm as advanced lesions
Supplementary Table 2Risk factors for advanced neoplasia according to the baseline findings and including sessile serrated polyps (SSPs) with cytological dysplasia and SSPs ≥10 mm as advanced lesions in the multivariable analysis
Supplementary Table 4Risk factors for advanced neoplasia according to the baseline findings and including sessile serrated adenomas/polyps (SSPs) with cytological dysplasia and SSPs ≥10 mm as advanced lesions in the multivariable analysis