Post-polypectomy recurrence: Low detector or high-risk polyp?

The most logical explanation for why high-detecting colonoscopists (high detectors) achieve much greater reductions in post-colonoscopy colorectal cancer (CRC) risk is the removal of all those neoplastic lesions that would be missed by low-detecting colonoscopists (low detectors).¹ However, a secondary and often neglected result of high detection is that many more patients are instructed to return at shorter intervals for surveillance colonoscopy. Since the patients of high-adenoma detection rate (ADR) colonoscopists receive both superior baseline-clearing colonoscopies and then return on average at shorter intervals, high detection provides two forms of protection against post-colonoscopy cancer. Indeed, since post-polypectomy surveillance has been shown clearly to contribute to reducing long-term CRC incidence and mortality in high-risk groups,² the shortened surveillance intervals of patients colonoscoped by high-ADR examiners may represent a significant aspect of the long-term CRC prevention achieved by high detectors.

Conversely, the high rate of missed lesions comprises the main CRC risk for the unfortunate patients of low-ADR colonoscopists. However, the application of post-polypectomy surveillance recommendations to the patients of low-ADR doctors subjects these patients to a second form of CRC risk. This second risk applies because surveillance intervals understandably, but incorrectly, assume uniform and reliable performance of baseline examinations. When low-ADR colonoscopists are working, unwitting patients are being told to return at longer intervals than the true status of their colorectal neoplasia would justify. Low detectors leave more neoplasia in the colon where it can progress to cancer or advance to later stages of cancer, and prescribe more time before the next colonoscopy should be performed so that progression can proceed during that period. If the above-described logic applies, another expected outcome of low-level baseline detection is excessively high detection of apparently metachronous lesions at surveillance, and possibly even when the surveillance examination is performed by the same low-ADR colonoscopist who performed the baseline colonoscopy. Thus, low-baseline detection results in more neoplasia remaining in the colon and available to progress. Also, the low-ADR colonoscopist leaves this neoplasia in the colon longer and, finally, during this longer surveillance interval it grows to a size at which even a low detector can find it at follow-up colonoscopy. For opposite reasons, high detectors might achieve such a minimization of the detection rate at surveillance to question the actual need for intensive surveillance or perhaps for any surveillance in some or all of their patients. All of these assumptions have been tested by Mangas et al. in this issue of the Journal.³ In a post-hoc analysis of the COLNPREV study, a three-fold difference in the rate of metachronous advanced neoplasia (AN) at first-surveillance colonoscopy between low and high detectors – stratified at baseline colonoscopy with an ADR cut-off of 33% – was shown in a cohort of 270 patients with high-risk adenomas. If the 13% rate of recurrent AN by low detectors was in line with similar estimates from pre-quality colonoscopy studies, the very low 4% achieved by high detectors is another powerful demonstration of the value that the competent individual endoscopist brings in long-term cancer protection to screened patients. This was ultimately confirmed by the finding that the ADR of the baseline endoscopist was the only independent predictor of metachronous AN, marginalising the role of polyp features, including the differentiation between intermediate- and high-risk groups according to European guidelines.

Based on these data, the detection rate at surveillance is a mere proxy of the initial quality of colonoscopy, supporting the hypothesis that most of the apparent recurrent AN is actually the result of lesions missed at baseline by low detectors. The dangerous idea that more intensive surveillance may be used to compensate for the initially suboptimal performance of the low detectors must be ruled out immediately. Such a high rate of AN is a passive epiphenomenon of an excessively high missing rate, and at least part of the missed lesions will progress to interval CRC before, or irrespective of, any surveillance. On the other hand, the very low rate of AN recurrence when the baseline colonoscopy is performed by high detectors confers an additional attribute to ADR as a universal quality indicator of colonoscopy.^4,5 ADR is not only associated with the risk of post-colonoscopy CRC, detection of serrated lesions, level of cleansing or withdrawal time, but it also predicts the recurrence rate of AN at surveillance.

It may be argued that it is too preliminary, and that additional data are needed, but the simple suggestion would appear as an appealing Copernican revolution. According to the data by Mangas et al.,³ the endoscopist and not the polyp is the main predictor of AN recurrence. Since the beginning of colonoscopy, size and unfavourable histology were considered to be intrinsically associated with a more severe cancerogenesis and a higher risk of AN recurrence.^6–8 Now, we realise that a high detector may dominate these features, downgrading their role in predicting an unfavourable outcome. This observation might open a completely new era where post-polypectomy guidelines admit a hierarchy between the competence of the endoscopist and polyp-related features. High-level detection has already been found to be more cost-effective than low-level detection despite the larger number of surveillance examinations performed when high-ADR doctors are at work, because the cost of cancer care is so high.⁹ If post-polypectomy surveillance recommendations were adjusted to expand the definitions of low-risk surveillance groups when baseline colonoscopy is performed by high-ADR doctors, then high-ADR doctors would become even more cost-effective relative to low detectors. By reducing the burden of surveillance with less intensive strategies, high detectors will save economic and financial resources that may be critical when coupling limited endoscopy capacity with mass population screening. In a post-polypectomy surveillance system that considers ADR, low-level detection would be likely to become increasingly unacceptable.

In conclusion, we learned that the quality of screening colonoscopy has a beneficial effect on post-polypectomy surveillance, further strengthening the need for proactive monitoring of the ADR of individual endoscopists and dedicated retraining of low detectors.

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.