Serrated polyps (SPs) are the second most common type of colorectal polyps with a distinct histological appearance of saw-toothed colonic crypts. According to the 2019 World Health Organization’s criteria, SPs are histologically classified as hyperplastic polyps (HPs), sessile serrated lesions (SSLs), or traditional serrated adenomas (TSAs) (1). Unlike the canonical adenoma-carcinoma sequence, SPs are proven to be early precursors to about 15%–30% colorectal carcinoma (CRC) through serrated pathway, frequently associated with BRAF or KRAS oncogenic mutations, microsatellite instability (MSI), and high CpG island methylator phenotype (CIMP) (2).

There is no consensus in the literature on which SPs are clinically relevant. A wide risk variation exists among SPs, as some case reports present SP rapid progression to invasive cancer within months (3), while other findings suggest a mean interval of 15 years for malignant transformation (4). Several longitudinal studies have investigated the relationship between the future risks of colorectal neoplasms and the clinicopathological characteristics of SPs, such as histology, size, anatomic location, or numbers (5–7). Evidence indicates that the detection of large SPs at the first endoscopy is more likely to have metachronous advanced neoplasms than those with no SPs (8), and it is an independent risk factor for subsequent CRC even with stronger association than that for advanced adenomas (9, 10). Cross-sectional reports have also substantially manifested that over-representation of malignancy hallmarks in SPs including BRAF mutation, MLH1 methylation, MUC5AC demethylation, and CIMP proposes possibly higher risks of CRC compared to conventional adenomas (11–13). Nevertheless, little is known about the genomic aberrations of SPs in relation to recurrence risks.

Therefore, we applied targeted next-generation sequencing (NGS) approach on groups of polyp-relapsed SPs (PRSPs) and polyp-free SPs (PFSPs) based on the surveillance outcomes to compare the intergroup differences of variants in 71 colorectal cancer-associated genes. In order to get insight about the value of genomic variants in predicting polyps’ recurrence and in planning colonoscopic surveillance, we further validated our results in a multicenter validation cohort established longitudinally from 2016 to 2019.

It has not been widely appreciated until 2003–2005 that distinct from conventional sequence, SPs are integrated into clinical practice as malignant precursors of CRCs through a serrated tumorigenic pathway. Clinical surveillance recommendations for SP patients on colonoscopy are difficult to set up uniformly on account of unknown underlying genetic mechanisms behind risks of subsequent CRCs (19–22). Controversy exists around the effectiveness of colonoscopy even based on the foremost guidelines, as unexpected interval cancers are sporadically developed within serial surveillance colonoscopies (23). Research regarding the remarkable resemblance of interval cancers to SPs in right-sided preference and epigenetic features such as MSI and CIMP (24, 25) has aroused extensive concern on identification of high-risk SPs in the clinic. Though various studies have documented generally accepted features of high-risk serrated lesions such as SSLs with size larger than 10 mm or SLs harboring dysplasia including TSAs that require intensive surveillance colonoscopy (5–7, 19), they are based on limited evidence. A recent study firstly provided longitudinal evidence on molecular markers (including BRAF V600E, CIMP, and MLH1 methylation) of SPs, suggesting that epigenetic defect of MLH1 methylation was in relation to subsequent advanced neoplasms (26). However, somatic mutations have not been screened yet.

In the present study, we firstly investigated the genetic alterations of SPs in relation to polyp recurrence by NGS. The mutational profile analyses demonstrated pathogenic variants in genes as previously reported, such as BRAF V600E and KRAS codon12/13 but also illustrated genes less recorded such as FBXW7, ATM, and DMD. This work provided several interesting findings. First, BRAF, KRAS, and APC were the key drivers of serrated tumorigenesis but not the useful markers to identify high-risk SPs. Second, higher prevalence of FBXW7, MSH2, and ERBB2 mutants in PRSPs rather than PFSPs might suggest their association with higher risks of developing colorectal polyps. Third, lower incidences of DMD, BRCA1, and BRCA2 mutants in PRSPs rather than PFSPs delineated the unlikelihood of their contribution to interval cancer. Finally, ERBB2 mutants demonstrated a significant relationship with colorectal advanced neoplasm recurrence, implicating its value as an important marker of high-risk SPs.

In line with previous studies, our findings recognized BRAF V600E or KRAS mutations as the triggering event of the serrated pathway (27). The overall rate of BRAF V600E (53/90, 59%) was within the range reported in the existing literatures from 50% to 83% (28, 29). The high level of mutant BRAF in SPs is becoming growingly relevant as a poor prognostic factor (30). However, our result indicated that BRAF V600E was not related to the future risks of neoplasms, consistent with the result of Hua et al. (26). The prevalence of KRAS mutations is discrete, ranging from 15% to 75% in colorectal polyps in previous studies (31, 32). The frequency of KRAS mutations in our NGS analysis was 48% (43/90). Interestingly, APC was the third most mutated gene among the SPs. APC gene typically acts as a driver gene in adenoma-to-carcinoma sequence through aberrant Wnt signaling (33). The potential role of APC mutations in the serrated pathway remains controversial. It is generally believed that Wnt pathway of SPs is mainly activated by a number of alternative mechanisms such as PTPRK-RSPO3 fusions and RNF43 mutations other than APC (28, 34). Although recently, research is suggesting that APC mutations are likely the main pathogenic reason for WNT signaling activation in serrated pathway based on their high frequency (35, 36).

ERBB2 is a member of the ErbB receptor tyrosine kinase (RTK) family that transduces downstream signaling pathway, such as Phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K)/protein kinase B (AKT) axis through the active form of homodimer or heterodimer complexes with other RTKs (37). ERBB2 overexpression occurs in many kinds of human cancers such as breast and ovarian cancers (38), and its association with poor prognosis in these cancers has been widely proven (39, 40). Moreover, the use of NGS has revealed the presence of ERBB2 sequence mutations in human tumors over recent 10 years. GENIE consortium data published in 2017 have exhibited that ERBB2 is altered in 4.69% of 2,081 CRC patients with R678Q and V842I present in 0.48% and 0.38% of all CRC patients, respectively (41). Given the nearly 5% frequency of ERBB2 alterations in CRCs based on the current literatures, ERBB2 becomes a promising target in anti-ERBB2-targeted therapies and prognoses. Whereas the presence of ERBB2 point mutations in colorectal serrated precursors has not been published before. Our findings have yielded important insights into the functional consequences of ERBB2 mutants. In our study, ERBB2 was mutated in 5.56% of 90 NGS subjects, and it was altered in 9.97% of 321 validation subjects with no statistical differences. Given the current identification of ERBB2 mutations in high-risk SPs, it could help shed some light on better understanding of polyp development and recurrence and could mark the significance that ERBB2 mutations may serve as an independent biomarker for guidance of clinical surveillance strategy. Patients with ERBB2 mutation-positive SPs may be recommended to receive more aggressive follow-up colonoscopy.

Our study has several strengths. First, SPs in this study are reclassified both historically and clinically. The historical reclassification was based on the newest 2019 WHO definitions by two independent pathologists, while the clinical subgroups were defined according to the follow-up outcomes. Second, this is the first cohort study exploring the correlation between high-risk genomic aberrations of SPs and recurrent outcome over time by using a targeted colorectal NGS panel. Prior literature only identified several endoscopic, histologic, or epigenetic features of serrated precursors that were strongly related to subsequent high-risk adenomas or CRCs identified during surveillance colonoscopy. Finally, the correlation between the ERBB2 mutational targets and colorectal polyp recurrence is first identified in the NGS cohort and was ulteriorly validated in a multicenter cohort. ERBB2 mutations are qualified as a clinically relevant biomarker to predict risks of future ANLs that have HRs >3.

This study has likewise some limitations. The retrospective study contained only a small number of SPs, and the follow-up duration was relatively short. Sanger sequencing platform had the limitation for reliable detection that required more than 10% fraction of mutated allele fraction. A sensitive digital PCR-based ERBB2 assay should be established in our future work to accurately evaluate its performance. The exploration of relevant factors of recurrence was limited in genomic features, while epigenetic characteristics of SPs were not dug out. Hence, the prognostic application of ERBB2 mutants in serrated pathway should be confirmed prospectively and extensively in the future.

In summary, for the first time, we have identified distinct genomic features of SPs in relation to subsequent polyp recurrence. This is also the first study detecting ERBB2 mutants in SPs. The clinical relevance of ERBB2 mutants with higher risks of subsequent colorectal neoplasms suggests their prospects as molecular markers for high-risk SP identification.

The original contributions presented in the study are publicly available. These data can be found here: https://ngdc.cncb.ac.cn/gsa-human, PRJCA007048 and GSA-Human: HRA001502.

The studies involving human participants were reviewed and approved by the Ethics Committee of the Renji Hospital, School of Medicine, Shanghai Jiao Tong University. The patients/participants provided their written informed consent to participate in this study.

Guarantor of the paper: X-BL and H-MC. Specific author contributions: Q-WW and X-YW performed the experiments, analyzed data, and wrote the original draft. H-YC revised the article. Q-WZ, J-NC, and Y-JZ collected clinical and pathological information. Z-RT and R-LW provided tissue specimens. X-BL and H-MC conceived and supervised the study. All the authors revised and approved the final article.

The project described was supported by a funding grant from the National Natural Science Foundation of China (NSFC) (General Program: 81772519) and by a grant from the Excellent Youth Program of Shanghai Municipal Commission of Health and Family Planning (2018YQ29). There was no influence from the NSFC or Shanghai Municipal Commission of Health and Family Planning on study design or conduct, data collection and management, analysis, interpretation, preparation and review or approval of the article.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.