Synergistic or independent? Revisiting colonoscopic surveillance for conventional and serrated lesions
Journal
Digestive Endoscopy
Journal Volume
34
Journal Issue
4
Pages
858-861
Date Issued
2022
Author(s)
Abstract
Sessile serrated lesions (SSLs) are unique colorectal neoplasms with distinct molecular alterations and histological features. The prevalence of SSLs has varied widely from 2% to 20% in previous observational studies1 and were considered premalignant lesions, as they could progress from benign adenoma to colorectal carcinoma (CRC) through unique carcinogenesis steps, known as the serrated pathway.2 Although the prevalence of SSLs is much lower than that of conventional adenoma, CRCs derived from SSLs still account for approximately one-third of all sporadic CRCs2 and a considerable number of post-colonoscopy CRCs (PCCRCs) owing to their subtle endoscopic morphology, previously unclear clinical entity and management strategy, and unique natural course. Considering the malignant potential of such lesions, SSLs should be treated as a precursor lesion of CRC similar to conventional adenoma, that is, a complete resection followed by colonoscopy in an optimal surveillance interval. The major concerns regarding tailoring one individual’s surveillance interval after polypectomy were basically his or her risk of developing metachronous advanced neoplasm (AN), as it is a more clinical proxy for CRC, during the follow-up period. To stratify this risk, the most important reference is the index colonoscopy finding, such as the number, size and histology component (having high-grade dysplasia or villous component) of the detected neoplasms. However, when considering the surveillance interval recommendation for SSLs, the evidence level is much lower than that in conventional adenoma. Current guidelines treat the existence of SSLs at index colonoscopy in different ways (Table 1). Some of them, such as the British guidelines,3 mix SSLs and conventional adenomas together and only define an umbrella diagnosis of “premalignant” or “advanced colorectal polyps”. In other words, subjects with two low-risk tubular adenomas (TAs) or two low-risk SSLs would be classified into the same risk group, and their recommended surveillance interval would be the same. On the other hand, the rest of the guidelines, such as Japanese,4 European5 and US guidelines,6 put SSLs in a separate category apart from TA with a surveillance recommendation specific to SSLs. For example, in the US guidelines, subjects with three low-risk TAs or three low-risk SSLs would be classified into different risk groups; therefore, a different surveillance interval would be recommended. Additionally, SSLs may coexist with conventional adenomas, which adds further confusion to surveillance planning in clinical practice. High-risk criteria: High-risk criteria: In real-world practice, subjects with synchronous SSLs and TAs are not uncommon; thus, the risk of metachronous AN after polypectomy in this particular population should also be addressed. Since the majority of colorectal neoplasms are conventional adenomas, it is intriguing to know whether the coexistence of SSLs would add extra or alter the risk of metachronous AN formation in comparison with subjects without synchronous SSL. A better understanding of the risk alteration brought by this synchronicity can help us better tailor the surveillance interval after colonoscopic polypectomy. However, none of the current guidelines provide a particular explanation or recommendation for subjects with both types of lesions. To make a proper post-polypectomy surveillance recommendation for this specific population, we first have to understand whether a synchronous SSL may change the subsequent risk of metachronous AN. A previous study by Melson et al.7 discovered that, in comparison with having only TA, the presence of SSL was associated with an approximately two-fold increased risk of metachronous AN. A more recent work by Anderson et al.8 discovered that subjects with high-risk adenoma (HRA) and synchronous SSL carried a much higher risk of metachronous AN (OR 16.04; 95% CI 6.95–37.00) than subjects with negative index colonoscopy findings, and this risk was much higher than in subjects having only HRA (OR 3.86; 95% CI 2.77–5.39). However, as Melson’s study enrolled many subjects with only SSLs in the TA-SSL group and Anderson’s study used subjects with negative index colonoscopy findings as the control group, the lack of a direct comparison between subjects with similar TA findings (with and without synchronous SSLs) made the impact of the coexistence of SSLs and TA elusive. In the recent study by Hong and colleagues,9 the authors classified their study cohort into HRA and low-risk adenoma (LRA) groups according to the TA findings (number, size and histology) from screening colonoscopy. Each risk group was further divided into subjects based on the coexistence of SSL. This classification method enabled a direct comparison between subjects with similar adenoma findings; therefore, it could emphasize the extra impact on the incidence of metachronous AN, if any, carried out by synchronous SSL, which was the most important value of this study. The first finding from this study was that synchronous SSLs would not change the risk of metachronous AN among subjects with similar conventional adenoma findings. Nevertheless, although having no clinical impact on metachronous AN, synchronous SSLs could significantly increase the risk of metachronous SSL, and this association was much stronger than with the coexistence of high-risk SSL. In addition to current knowledge of the carcinogenesis of CRC, the results of this study further imply that the two major colorectal carcinogenesis pathways, conventional and serrated pathways, could exist simultaneously but work independently and have no significant impact on each other, as the coexistence of serrated pathways would not alter the existing risk derived from individuals’ conventional adenomatous burden. Moreover, this result can also help us to reconsider whether it is justified to recommend a surveillance interval to a patient regardless of his or her synchronous SSLs because this would not change the risk of metachronous AN. However, this study is not without limitation. First, the histological diagnosis criteria of SSLs were revised in 2019; therefore, the current study could not completely apply such a definition. To re-evaluate the clinical impact of these lesions, we need SSL studies that use the most updated diagnostic definition. Second, the scarcity of subjects with synchronous high-risk SSL (only 36 subjects in the LRA group and 21 subjects in the HRA group), either in the baseline or surveillance colonoscopies, will underestimate the clinical impact of these high-risk lesions. To the best of our knowledge, most SSLs are indolent in nature and have a long indwelling time from a precancerous status to CRC. However, carcinogenesis via the serrated pathway may speed up when dysplasia develops. Accordingly, when “high-risk” SSLs are detected during colonoscopy, a more intense surveillance interval (3 years) should be recommended. The current study, however, disclosed that a synchronous high-risk SSL was associated with higher metachronous high-risk SSL incidence only in subjects with LRA but not in the HRA group. This discrepancy in the clinical impact of synchronous high-risk SSL in subjects with different conventional adenomatous conditions was possibly due to the small subject number and insufficient statistical power. Third, more than 40% of the subjects in either the LRA or HRA groups were not included in this study due to missing surveillance colonoscopy data, which might have induced bias. Because subjects would have better compliance with surveillance recommendations if they had a more advanced finding at index colonoscopy, such a high percentage of subjects without documented surveillance results would lead to a significant surveillance bias. Last, but not least, this study enrolled only an Asian population, in which the prevalence of SSL was reported to be much lower than that in the Western population. Thus, the study results may not be applicable to other ethnicities. Regarding the growing concern about the epidemiology of SSL and its contribution to CRC or PCCRC in recent years, there are still several areas worth studying. First, to understand the true clinical significance of SSL, an accurate prevalence study of this neoplasm entity is necessary. The colonoscopy detection of SSL by endoscopists was, however, lower than that for conventional adenoma, and a high detection rate of conventional adenoma may not necessarily be associated with high detection of SSL.10 Effort to improve endoscopists’ ability to detect SSL is the first step to fill the knowledge gap in this regard. Second, the efficacy of post-polypectomy surveillance regarding SSLs under current surveillance guidelines remains largely unknown; therefore, randomized control trials or well-designed prospective cohort studies using advanced SSL or CRC arising via the serrated pathway as the endpoint are mandatory to clarify this issue. Third, current surveillance guidelines define high-risk SSL as SSL with any degree of dysplasia. Whether the different grades of dysplasia in SSL may affect metachronous AN or even CRC development differently still needs to be investigated. Fourth, several modifiable factors, such as obesity, metabolic derangement, smoking or exercise, have proven to be associated with the risk of metachronous AN. Whether these correlations are still present regarding the risk of metachronous SSL is also waiting to be disclosed. Finally, conciseness and explicitness may affect the practicality of surveillance guidelines in clinical practice, and the balance between precision medicine and feasibility is still an important consideration. Author H.M. Chiu is an Associate Editor of Digestive Endoscopy. The other author declares no conflict of interest for this article. None.
SDGs
Publisher
John Wiley and Sons Inc
Type
editorial