Surveillance of colorectal neoplasia: a review of guidelines, current evidence, and future directions

Abstract

About one third of individuals undergoing colonoscopy for colorectal cancer (CRC) screening are found to have neoplasms that require ongoing colonoscopic surveillance for metachronous neoplasia. Current guidelines for post-polypectomy surveillance interval are based on polyp features of size, histology, and multiplicity. Recent large, observational cohort studies delineate CRC risk among individuals with previous neoplasia while others investigate the role of incorporating patient and endoscopist characteristics in addition to polyp features to more precisely identify patients most likely to benefit from surveillance. Several gaps in colorectal neoplasia surveillance remain, including the role of noninvasive stool-based tests and surveillance cessation approaches for older adults. The purpose of this narrative review is to summarize current post-polypectomy surveillance guidelines from around the world, describe recent observational cohort studies quantifying post-polypectomy CRC risk, highlight novel approaches to risk-stratification for surveillance, summarize ongoing clinical trials of surveillance modalities and intervals that will impact practice, and outline an agenda for future research to further improve the efficiency and effectiveness of colorectal neoplasia surveillance.

1. Introduction

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-related deaths worldwide [1]. Up to one third of patients undergoing screening or diagnostic colonoscopy have neoplasms that require surveillance, and colonoscopic surveillance in turn contributes to an estimated 30% reduction in CRC incidence after initial index colonoscopy [2–4]. Following the identification of colorectal neoplasia, the process of periodically repeating colonoscopy to identify subsequent neoplasia is known as surveillance, the current standard of care. In many parts of the world, such as the US, colonoscopy is the preferred and only recommended, post-polypectomy surveillance modality. The demand for colonoscopy far exceeds capacity in most health systems, yet there is ongoing overuse of surveillance for low-risk and underuse for high-risk colorectal neoplasia [5,6]. While clinical trials are underway to determine the effectiveness of noninvasive surveillance approaches, such as with fecal immunochemical test (FIT) for low-risk neoplasia, risk stratification based on factors beyond polyp characteristics is gaining attention to reduce patient burden and allocate surveillance resources more efficiently [7–9].

In this narrative review, we first highlight the most up-to-date guidelines for post-polypectomy surveillance from around the world. We then review recent, large observational cohort studies describing CRC risk based on polyp characteristics, and summarize novel approaches to optimizing surveillance. We conclude by describing ongoing surveillance-related clinical trials and highlight areas for future research. Recognizing that evidence-based guidelines for surveillance differ for individuals with high-risk conditions such as inflammatory bowel disease, hereditary CRC syndromes, malignant polyps, and personal or family history of CRC, we focus this review on surveillance for individuals with colorectal neoplasia who are otherwise at average risk for CRC.

2. Methods

To identify relevant studies, we searched Pubmed for articles using the Medical Subject Headings (MeSH terms) “colorectal neoplasms” combined with “prevention and control,” “mortality,” and “epidemiology.” We limited the search to 10 years (through February 2025) and full text articles available in the English language. We also reviewed references of the retrieved articles to identify studies that were not identified electronically. To identify ongoing clinical trials, we searched clinicaltrials.gov and added additional trials that were not included in this database based on recommendations from the peer review process. We acknowledge the vast amount of literature on colorectal neoplasia and aim to provide an overview using selected studies. Variation in terminology used for polyp subgroups throughout this manuscript is due to variation in terms used in the reference articles. As some of these subgroups differ across studies (i.e., low-risk polyps vs low-risk adenomas), we used terms as they were used in each referenced article and provided definitions where needed.

3. Current guidelines

All guidelines that were reviewed emphasized that surveillance intervals assume high-quality colonoscopy as defined by adequate bowel preparation, and a complete exam to cecum. Surveillance intervals should be adjusted for piecemeal polyp resection, suboptimal bowel preparation, and incomplete colonoscopy.

3.1. United States

Nearly three decades ago, shortly after results from one of the earliest randomized trials of post-polypectomy surveillance intervals were published, the first comprehensive report for CRC screening and surveillance was published in the US [10,11]. Evidence had started to emerge to show that the risk of progression from adenoma to carcinoma was associated with polyp size, histology, and multiplicity; though in practice, surveillance intervals varied considerably. Based on results from these early studies, including the National Polyp Study [10], the first guidance for CRC screening and surveillance was published in 1997 [11] with a recommendation for surveillance colonoscopy every 3 years for individuals with adenoma >1 cm in size, or with multiple adenomas regardless of size, on index colonoscopy. Since this report, surveillance guidelines have continued to evolve in the US, further differentiating CRC risk among individuals with a few small adenomas from those with multiple, large adenomas, and serrated polyps (SPs). The overall trend over the past 3 decades has been to gradually de-intensify surveillance for low-risk neoplasia, although colonoscopy continues to be the exclusive surveillance modality.

In 2020, the US Multi-Society Task Force (USMSTF), published the most recent guidelines for post-polypectomy surveillance, emphasizing the importance of high-quality colonoscopy and standardized terminology for colorectal neoplasia (Table 1) [12]. Individuals with low-risk adenomas (LRAs), defined as 1–2 tubular adenomas, each <10 mm in size, were recommended to undergo surveillance colonoscopy every 7–10 years (vs 5–10 years previously) [12,13]. Individuals with high-risk adenomas (HRAs) were defined as any advanced adenoma (AA; ≥10 mm in size, or with villous features or high-grade dysplasia), or 3 or more adenomas of any size. These individuals were further risk-stratified based on multiplicity for updated surveillance intervals including: 3–5 years for those with 3–4 adenomas all <10 mm in size; and 3 years for individuals with 5–10 adenomas all <10 mm in size, at least one adenoma ≥10 mm in size, or an adenoma with villous features or high-grade dysplasia regardless of size. For individuals with >10 adenomas, regardless of size, a 1-year surveillance interval was recommended.

Table 1.

Comparison of post-polypectomy guidelines.

Baseline colonoscopy	First follow-up surveillance recommendationa (in years)
	USMSTF [12]	Canada			Europe		Asia-Pacificl [20]
		Ontario [16]	Alberta [15]	BC [17]	ESGE [18]	BSG/ACGBI/PHEi [19]
Adenomas
1–2 tubular adenomas, <10 mm	7–10	5 (FIT)	5 (FIT)	10	Return to screening (10)e	Return to screening (FIT)	10m
3–4 tubular adenomas, <10 mm	3–5	3	5				3
5–10 tubular adenomas, <10 mm	3		3	5–9 TAs = 3	3	3
>10 adenomas	1	≤1	1	≥10 TAs = 1	–
Adenoma ≥10 mmb, with villous/tubulovillous features, and/or high-grade dysplasia	3	3	3	3	3f,g	3g
Serrated Polyps
≤20 HP in the sigmoid/rectum, <10 mmc	10	10 (FIT)	10 (FIT)	10 (FIT)	Return to screening (10)e	≥5 HP = 3j <5 HP = Return to screening (FIT)	Return to screening
≤20 HP anywhere proximal to sigmoid colon, <10 mmc
Hyperplastic polyp ≥10 mm	3–5	–	3d	3	3	3	3n
1–2 SSLs, <10 mm	5–10	5	5	10	Return to screeninge,h (10)	Return to screening (FIT)h	Return to screening
3–4 SSLs, <10 mm	3–5		3
5–10 SSLs, <10 mm	3			5–9 SSLs = 3; ≥10 SSLs = 1	5–9 SSLs = 10; >10 = genetics referral	3k
SSL ≥10 mmb, with dysplasia, and/or a traditional serrated adenoma		3		3	3		3n

USMSTF = United States Multi-Society Task Force; BC = British Columbia; ESGE = European Society of Gastrointestinal Endoscopy; BSG = British Society Guidelines; ACGBI = Association of Coloproctology of Great Britain and Ireland; PHE = Public Health England; FIT=fecal immunochemical test; TA=tubular adenoma(s); SSL = sessile serrated lesion; HP = hyperplastic polyp(s).

^aUnless otherwise specified, recommendations are for surveillance colonoscopy.

^bPiecemeal resection of polyps ≥20 mm in size should prompt surveillance colonoscopy in 6 months.

^cWorld Health Organization criteria for serrated polyposis syndrome (SPS) include: 1) presence of ≥5 SSLs proximal to the rectum that are ≥5 mm and ≥2 of which are ≥10 mm; and 2) >20 serrated polyps of any size with ≥5 located proximal to the rectum. Individuals who meet these criteria should undergo tailored surveillance [63].

^dFor HPs ≥10 mm in size and proximal to sigmoid colon, 3-year surveillance is recommended. For HPs ≥10 mm in size and in the rectosigmoid, a 5-year surveillance colonoscopy is recommended.

^eReturn to the local organized screening program, or if such a program is not available, then colonoscopy in 10 years.

^f3-year surveillance if size ≥10 mm AND high-risk histologic features. If size <10 mm then 10-year surveillance regardless of other high-risk histologic features.

^gVillous/tubulovillous histology is not considered high-risk in the European guidelines.

^hAs long as no dysplasia, return to screening is recommended.

ⁱSurveillance colonoscopy is recommended to subside if life-expectancy is <10 years, or at age older than 75 years.

^jExcludes 1–5 mm rectal hyperplastic polyps, for which a return to screening is recommended.

^kTraditional serrated adenoma follows the same guidelines as other premalignant polyps such that this histology by itself does not indicate a 3-year surveillance. Rather, at least 2 premalignant polyps (defined by adenomas or SPs inclusive of HPs, SSLs, and TSAs) including at least 1 advanced adenoma or advanced serrated polyps (defined by size >10 mm or with dysplasia), OR 5 or more premalignant polyps are required for a 3-year surveillance.

^lA return to screening is assumed for categories of polyps for which a surveillance period is not specified.

^mSurveillance can be performed with FIT or colonoscopy.

ⁿSize criteria includes surveillance in 3 years if SSL is >10 mm in size.

^oSurveillance cessation considered for individuals 80 years of age, or earlier if life expectancy limited by comorbidities.

The USMSTF also updated guidance for SP surveillance in 2020 [12]. Serrated polyps include hyperplastic polyps (HPs), sessile serrated lesions (SSL; synonymous with sessile serrated polyps), and traditional serrated adenomas (TSA). For individuals with ≤20 HPs in the sigmoid or rectum colon, and those with ≤20 HPs proximal to the sigmoid colon, all <10 mm in size, a 10-year surveillance colonoscopy interval was recommended. The overall approach was to consider the SSL as the adenoma’s counterpart in the serrated neoplasia pathway, while recognizing that the supporting literature was lower in quantity and quality than that for adenomas, and the known interobserver variability in distinguishing SSLs from HPs. Thus, the risk stratification of SSLs also depended on size and histology, separating a group of higher-risk lesions for more intensive surveillance broadly similar to the approach for adenomas. Unchanged from the 2012 guidelines, 1–2 SSLs <10 mm in size without dysplasia were recommended for a 5- to10-year surveillance interval whereas 3–4 SSLs, each of which is <10 mm in size, were recommended for a 3- to 5-year surveillance interval. Individuals with 5–10 SSLs, all <10 mm in size, were recommended for a 3-year surveillance interval, as were SSLs ≥10 mm in size. A 3-year surveillance interval was also recommended for any SSL with dysplasia, regardless of size or number, and for a TSA. In addition to these updates, a HP that is ≥10 mm in size, regardless of location in the colon, was recommended for surveillance in 3- to 5-years. It is important to note here that surveillance recommendations for serrated lesions are affected by pathologist variability in the distinction between HP and SSL. Thus, if there are concerns that SSLs are being misdiagnosed as HP locally, then clinicians can adapt surveillance recommendations to be more closely aligned with those that apply to SSLs.

3.2. Canada

In Canada, unless an individual is identified as high-risk for CRC based on personal or family history, screening is recommended every 1–2years starting at age 50 through 74years using the FIT [14]. Beyond screening, post-polypectomy surveillance guidelines are developed under the jurisdiction of each province and territory due to variation in resources. For this review, we highlight post-polypectomy guidelines developed by Cancer Care Ontario, Alberta Health Services, and Provincial Health Authorities for British Columbia, which represent three of the four most populated provinces [15–17]. Surveillance guidelines in these regions are largely consistent with post-polypectomy surveillance recommendations in the US and similar between provinces, but with a few notable differences.

Adenomas are defined as low- and high-risk consistent with the USMSTF definitions but with variation in surveillance intervals and type of test based on size and number (Table 1) [12]. The first distinction from the US guidelines includes a return to FIT-based screening in 10 years for individuals with HPs <10 mm in size for all three provinces, and a return to FIT-based surveillance in 5 years for those with 1–2 tubular adenomas in Ontario and Alberta [15–17]. In British Columbia, for 1–4 tubular adenomas, each <10 mm in size, the recommendation is for 10-year surveillance colonoscopy, and if normal, then a return to FIT-based screening in 10 years for individuals without family history of CRC [17]. Surveillance for HRAs is consistent with the USMSTF guideline recommendation of 3 years; however, individuals with >10 adenomas are recommended to undergo a clearing colonoscopy within 1 year of index exam, with subsequent colonoscopies to be performed in <3 years with the precise interval at the endoscopist’s discretion. Also distinct from the USMSTF guidelines is a return to FIT-based screening in 5 years if the second surveillance colonoscopy for individuals with high-risk polyps on index exam is normal, rather than ongoing colonoscopic surveillance. Additional post-polypectomy surveillance recommendations in Ontario, Alberta, and British Columbia are highlighted in Table 1.

3.3. Europe

In Europe, the European Society of Gastrointestinal Endoscopy (ESGE) and the British Society of Gastroenterology in collaboration with Association of Coloproctology of Great Britain and Ireland, and Public Health England (BSG/ACGBI/PHE) have published surveillance guidance [18,19]. While this review focuses on ESGE and BSG/ACGBI/PHE guidelines, it should be noted that guidance may vary within countries in Europe. Notable differences between ESGE guidelines, relative to the US and Canada, include the recommendations that patients who are found to have 1–4 tubular adenomas <10 mm in size, irrespective of villous features, return to general screening (i.e., return to the local organized screening program or if such a program is not available, then colonoscopy in 10 years) [18]. Individuals with 5–10 tubular adenomas <10 mm in size, and adenoma ≥10 mm in size or with high-grade dysplasia, are recommended to undergo surveillance colonoscopy in 3 years. For SSLs, ESGE recommends returning to screening for individuals who are found to have up to 10 SSLs <10 mm in size. For SSLs ≥10 mm in size, SSLs with dysplasia regardless of size, and TSA regardless of size, surveillance colonoscopy is recommended in 3 years. Surveillance cessation is recommended at the age of 80 years, or earlier depending on life expectancy.

The BSG/ACGBI/PHE guidelines define high-risk findings from index colonoscopy as: ≥2 premalignant polyps (adenomas and/or SPs) including ≥1 advanced colorectal polyp (adenoma ≥10 mm in size and/or with HGD; SP ≥10 mm and/or containing dysplasia), or ≥5 premalignant polyps, all of which are recommended for surveillance colonoscopy in 3 years [19]. Premalignant polyp includes adenomas and SPs (HPs, SSLs, and TSAs, but excludes 1–5 mm rectal HPs). Similar to the ESGE guidelines, villous and tubulovillous histology are not considered part of the AA definition in the BSG/ACGBI/PHE guidelines. A return to screening with biennial FIT is recommended for individuals with up to 4 small, non-advanced adenomas, assuming a high-quality index colonoscopy. In addition, surveillance colonoscopy is recommended to subside for individuals with expected life-expectancy of less than 10 years, or at age older than 75 years (Table 1).

3.4. Asia-Pacific region

The Asia-Pacific region, comprising a panel of experts from China, Japan, Korea, Australia, and New Zealand, also published post-polypectomy surveillance guidelines in 2022 using a modified Delphi process [20]. Notable post-polypectomy surveillance recommendations that differ from other international professional societies include: 1) surveillance for LRAs can be performed by colonoscopy or FIT, at intervals used for average-risk individuals (i.e., 10 years for colonoscopy; every year or every other year with FIT); and 2) 3-year surveillance recommendation for 3 or more tubular adenomas, AA, and advanced SPs. Additional post-polypectomy surveillance guidelines for this region, based on the Asia-Pacific guidelines, with note of similarities and differences as compared to other international guidelines are highlighted in Table 1, although it is important to note that local guidance may vary within countries in this region.

4. Neoplasia and long-term colorectal cancer risk

4.1. Individual studies

Several recent cohort studies provide evidence for metachronous CRC risk based on polyp characteristics at index colonoscopy. Below we highlight selected studies based on publication within the past 10 years, duration of follow-up period >5 years, representation of large populations from around the world, and summarize recent pooled analyses that include some of these and other cohorts. Selected large observational studies are summarized in Supplementary Table 1.

A recent post hoc analysis of 15,935 participants from a large, multicenter trial in the US, the Prostate, Lung, Colorectal, and Ovarian Cancer Screening (PLCO) trial, showed that participants who had index AA were more likely to develop CRC compared to those without index adenomas (rate ratio (RR) 2.7; 95% confidence interval (CI) 1.9–3.7) during a median follow-up period of 12.9 years [21]. The presence of AA also increased risk for CRC-related mortality (RR 2.6; 95% CI 1.2–5.7). When comparing AA as defined by size ≥1 cm, compared to those that were <1 cm but with advanced histology (villous or tubulovillous features, or high-grade dysplasia), CRC incidence rates per 10,000 person-years were similar (19.2, 95% CI 13.9–24.5 vs 22.4, 95% CI 12.3–32.5; p = 0.58). Further, for adenomas <1 cm but with advanced histology, there was no significant difference in CRC incidence rates per 10,000 person-years between those with villous/tubulovillous features as compared to those with high-grade dysplasia (21.2, 95% CI 10.4–31.8 vs 28.6, 95% CI 0.6–56.6; p = 0.59). In contrast, individuals with non-advanced adenomas had no increase in CRC risk (RR 1.2; 95% CI 0.8–1.7) or CRC-related mortality (RR 1.2; 95% CI 0.5–2.7) as compared to those with no adenomas on index exam. When evaluating CRC risk for those with 3 or more non-advanced adenomas compared to those with no adenomas on index exam, there was also no significant difference (RR 1.4; 95% CI 0.6–3.0). Participants of the PLCO trial were 55–74 years of age upon enrollment, which took place between 1993 and 2001, and all had abnormal flexible sigmoidoscopies that prompted colonoscopy. Significantly increased CRC incidence and mortality among those with 3 or more adenomas, along with adenomas >10 mm in size or those with advanced histology, and lack thereof among individuals with non-advanced adenomas at index colonoscopy, are findings that were corroborated by others including an observational cohort study comprising 64,422 members of the Kaiser Permanente Northern California health system with a follow-up period up to 14 years and 3,121 patients in the Veterans Affairs Cooperative Studies Program with a follow-up period of 10 years (Supplementary Table 1) [22,23].

Another large observational cohort study in the US, comprising 122,899 participants from the Nurses’ Health Studies (NHS1 and NHS2) and Health Professionals Follow-up Study, showed that compared to individuals with no index polyps, those with non-advanced adenomas or small SPs (<10 mm) had no increase in CRC risk (HR 1.21, 95% CI 0.68–2.16; HR 1.25, 95% CI 0.76–2.08, respectively) [24]. However, individuals with AA (HR 4.07, 95% CI 2.89–5.72) and those with large SPs as defined by size ≥10 mm (HR 3.35, 95% CI 1.37–8.15) had greater risk of developing CRC as compared to those with no index polyps. Notably, CRC risk also increased among those with 3–10 adenomas (HR 3.15; 95% CI 1.29–7.67), and those with adenomas with villous (HR 8.51; 95% CI 4.50–16.1) or tubullovillous (HR 3.17; 95% CI 1.88–5.36) features, regardless of size, as compared to those without index polyps. And finally, CRC risk for adenomas with high-grade dysplasia, regardless of size, was also increased (HR 5.95; 95% CI 1.88–18.8) as compared to those with no index polyps.

In a separate study, Polychronidis and colleagues studied a cohort of 156,699 participants who underwent index colonoscopy between 2007 and 2017 as part of the Massachusetts General Brigham Colonoscopy Cohort to estimate subsequent risk of CRC and high-risk polyps after index colonoscopy with polypectomy [25]. Polyps were characterized as HRAs based on the USMSTF definitions for advanced adenoma or based on multiplicity (≥3 adenomas) [12]. All other adenomas were defined as low-risk. Serrated polyps included HPs, SSLs, and TSAs with high-risk SPs defined as those that were ≥10 mm in size or had dysplasia. All other SPs were considered low-risk SPs. Patients with HRAs and high-risk SPs had a higher risk of CRC during a median follow-up period of 5.3 years, with the greatest risk within 3 years after polypectomy (HR 5.44, 95% CI 3.56–8.29; HR 8.35, 95% CI 4.20–16.59, respectively) as compared to individuals without polyps on index colonoscopy. When surveillance colonoscopy was performed for high-risk polyps (including HRAs and high-risk SPs), a lower risk of CRC was noted (HR 0.61, 95% CI 0.39–0.98) as compared to those with high-risk polyps who did not undergo surveillance. Lower risk of CRC was also noted among those with low-risk polyps (including LRAs and low-risk SPs) who underwent surveillance (HR 0.57, 95% CI 0.35–0.92) as compared to those with low-risk polyps who did not undergo surveillance. The authors noted that among nearly 1,550 patients with high-risk polyps at index and surveillance colonoscopies, about one third had index and interval polyps in the same colon segments, suggesting that perhaps incomplete resection or a “field effect” (in this case, an area of the colon in which neoplasia-associated molecular alterations, such as chromosomal instability or epigenetic changes, are present without discernible mucosal abnormalities) led to interval neoplasia identified at time of surveillance.

Results from these large, multicenter US studies are similar to other large observational cohort studies from other countries. A multicenter study comprising 17 hospitals in the UK in which a total of 28,972 patients underwent colonoscopy between 1984 and 2010 (majority between 2000 and 2010), also showed a lack of significantly increased risk for CRC in the low-risk neoplasia cohort compared to the general population [26]. In this study, adenomas were risk-stratified based on size and multiplicity into low- (1–2 adenomas each <10 mm), intermediate- (3–4 adenomas each <10 mm, or 1–2 adenomas with at least one ≥10 mm), and high-risk (5 or more adenomas <10 mm, or 3 or more adenomas with at least one ≥10 mm), with further risk stratification within each of these categories based on additional risk factors such that there was a lower-risk subgroup and higher-risk subgroup in each of the categories. For example, for the low-risk cohort, the number of surveillance exams, age, adenoma histology, proximal location of adenomas, and complete colonoscopy were independent risk factors for CRC. After adjusting for these factors, a single surveillance exam was associated with a 44% reduction in CRC incidence compared to no surveillance. Without surveillance, the cumulative 10-year CRC indence was 1.7% (95% CI 1.2–2.1) in the low-risk cohort, which differed significantly between the lower- and higher-risk subgroups within this cohort (1.2%; 95% CI 0.8–1.7 and 2.1%; 95% CI 1.7–2.6, respectively). Overall, when considering the low-risk cohort, CRC risk was not significantly increased for those without surveillance as compared to the general population (SIR 0.86; 95% CI 0.73–1.02), which was also true for the higher-risk subgroup within this cohort (SIR 1.07; 95% CI 0.88–1.28). However, the CRC risk was lower in the lower-risk subgroup (SIR 0.51; 95% CI 0.35–0.73) as compared to the general population [26]. This finding contrasts to the increased CRC risk in the high-risk neoplasia cohort, in which the number of surveillance exams, presence of high-grade dysplasia, and complete colonoscopy were independently associated with CRC incidence. Without surveillance, the cumulative CRC incidence at 10 years was 5.7% (95% CI 4.0–8.3) in the high-risk cohort, which differed between the lower- and higher- risk subgroups within this cohort (3.8%; 95% CI 2.1–6.8 and 9.9%; 95% CI 6.2–15.7, respectively). As compared to the general population, CRC incidence was greater for the high-risk cohort (SIR 1.91; 95% CI 1.39–2.56) and higher-risk subgroup within this cohort (SIR 3.55; 95% CI 2.34–5.17) but similar for the lower-risk subgroup (SIR 1.10; 95% CI 0.64–1.76). Based on these data, the authors concluded that there is heterogeneity within the low- and high-risk categories based on adenoma size and mulitpilicity alone, and that complete colonoscopy is independently associated with CRC risk. Assuming a high-quality index exam that is complete to cecum, for distal adenomas with non-advanced histology, the results of this study support that those in the low- and intermediate-risk neoplasia categories (i.e., up to 4 small tubular adenomas) could return to screening with biennial FIT rather than undergoing surveillance colonoscopy, as reflected in their current guidelines [19]. Limitations of this study include the lack consideration of SSLs given the timeframe from which data were collected and the assumption that all colonoscopies after the index exam were performed for surveillance purposes.

Another multicenter study of 236,089 individuals enrolled in the Polish National CRC Screening program between 2000 and 2011 further supports the variable risk of CRC among individuals based on specific characteristics of polyps [27]. Compared with individuals without adenomas, those with LRAs had a higher risk of CRC (HR 1.49; 95% CI 1.13–1.98) although risk was even higher among those with adenomas with advanced features (3 or more adenomas, ≥10 mm in size, villous features, and/or high-grade dysplasia) (HR 2.94; 95% CI 2.28–3.81). When the authors further compared CRC risk between individuals with adenomas, size ≥20 mm was associated with the highest risk of CRC (HR 4.83; 95% CI 3.05–7.63) and when these patients were removed, the next highest risk was among those with high-grade dysplasia (HR 2.06; 95% CI 1.08–3.92). Neither number of adenomas (≥3 vs 1–2) nor villous features were independently associated CRC risk factors among those with adenomas. Compared to individuals without adenomas, those with adenomas that were <20 mm without high-grade dysplasia had greater CRC risk (age-adjusted HR 1.55; 95% CI 1.22–1.96). When compared to individuals without adenomas, those with adenomas ≥20 mm (age-adjusted HR 9.25, 95% 6.39–13.39) and those with adenomas <20 mm in size with high-grade dysplasia (age-adjusted HR 3.58, 95% CI 1.96–6.54) also had greater CRC risk. Using these findings, the authors concluded that a novel risk stratification system, based on 3 categories inclusive of adenoma size and grade of dysplasia, could more accurately identify those with highest risk for CRC and death after adenoma resection as compared to the current risk stratification categories of adenomas ≥10 mm in size, 3 more adenomas, or those with villous/tubulovillous features or high-grade dysplasia. The 3 novel categories, in order of highest to lowest CRC risk as compared to the general population, included: 1) adenomas ≥20 mm in size; 2) adenomas <20 mm in size with high-grade dysplasia; and 3) adenomas <20 mm without high-grade dysplasia, and recommended that intensive surveillance could be considered for those in the first two categories.

Another population-based study from Denmark compared 704 patients with screen-detected adenomas to 1,221,726 individuals in the general population [28]. Individuals with adenomas were categorized according to the European Union (EU) or ESGE. Using the EU risk-stratification, adenomas were classified as low- (1–2 small adenomas), intermediate- (3–4 small adenomas or 1 adenoma 10–19 mm in size), and high-risk (≥5 adenomas or 1 adenoma ≥20 mm). During a median of 10.7 years of follow-up, CRC incidence rates were 251 (95% CI 94–671), 218 (95% CI 104–456), and 542 (95% CI 300–978) per 100,000 person-years among individuals with low-, intermediate-, and high-risk adenomas, respectively, using the EU risk-stratification system. When adjusted for age and sex, the high-risk subgroup had a higher CRC risk (aHR 1.80, 95% CI 1.00–3.26) as compared to the general population based on index colonoscopy findings. Colorectal cancer risk was not significantly elevated for those with low- and intermediate-risk adenomas [(HR 0.88, 95% CI 0.33–2.35) and (HR 0.72, 95% CI 0.34–1.51), respectively] compared to the general population. Without surveillance, CRC risk was signficiantly higher for the EU high-risk subgroup (aHR 3.34; 95% CI 1.67–6.69) but was reduced for those who had at least one follow-up exam during the study period (aHR 0.81; 95% CI 0.26–2.51) as compared to the general population.

Finally, Zhu and colleagues studied CRC risk among 34,382 patients 40–74 years of age who underwent colonoscopy from 2007 to 2016 in community-based practices in China [29]. After a median follow-up period of 7.7 years, they also found a significant increase in CRC risk among those with advanced adenomas (HR 4.01; 95% CI 2.37–6.77) as compared to those with no adenomas. For those with LRAs, no significant increase in CRC risk was found.

While these large multicenter observational studies were published recently, their generalizability to current practice may be limited by a few factors, the first of which is the inclusion of individuals who underwent colonoscopy in the late 1990s or early 2000s, a period during which colonoscopy may have been performed with standard (vs. high definition) colonoscopes. Second is that quality indicators were not as well-defined as they are currently. These two limitations may have resulted in underdetection of polyps. A third factor potentially limiting generalizability is the variation in surveillance practices at the time. Also, these studies largely represent cohorts of individuals in the US and Europe, where regional variation in neoplasia prevalence or other lifestyle-related risk factors may further limit generalizability.

4.2. Systematic reviews

Several recently published pooled analyses comprise some of the previously discussed studies and additional cohorts. While we present these reviews, it should be noted that some studies within each review did not have long follow-up periods (i.e., 5 years or less). The first is a systematic review and meta-analysis of studies from several regions throughout the world (e.g., Asia, Israel, US Europe) comparing the risk of CRC between otherwise average-risk individuals with 1–2 small, tubular adenomas to individuals with normal index colonoscopies or the general population [30]. By pooling 8 observational studies, all of which were published after 2006 and including 10,319 individuals, the authors found an increased risk for developing AA among individuals with LRAs in comparison to those without adenomas on index colonoscopy (RR 1.55; 95% CI 1.24–1.95 with no significant statistical heterogeneity among studies). However, when compared to the general population, one population-based case-control study from Germany [31] and one retrospective cohort study from France [32], showed that individuals with LRAs had a reduced risk of CRC (OR 0.4; 95% CI 0.2–0.6 at 3–5 years; and standardized incidence ratio 0.68; 95% CI 0.44–0.99 at a median follow up of 7.7 years, in these studies, respectively). Another retrospective cohort study [33] from Norway, included in this review, showed a 25% relative risk reduction in CRC-mortality in the LRA group compared to the general population after a median follow up of 7.7 years. In another meta-analysis comprising 12 studies, 510,019 patients (mean age 59.2 years ± 2.6 years) and mean follow-up of 8.5 ± 3.3 years, CRC incidence was higher in those with LRAs compared to those without adenomas (OR 1.26; 95% CI 1.06–1.51) but with no difference in CRC-related mortality (OR 1.15; 95% CI 0.76–1.74) [34]. This finding contrasts to patients with HRAs compared to those without adenomas, where there was a significant increase in CRC risk and mortality (OR 2.92; 95% CI 2.31–3.69 and OR 2.69; 95% CI 1.87–3.87), respectively. It should be noted that not all studies applied the same definitions for low- and high-risk adenomas.

A third pooled analysis of 55 studies, comprising a total of 936,540 patients with adenomas or CRC at time of index colonoscopy aimed to assess risk for metachronous CRC based on specific characteristics of HRAs [35]. Increased risk for metachronous CRC was associated with the following adenoma features on index colonoscopy: size ≥20 mm as compared to size <20 mm (RR 2.98; 95% CI 1.43–6.19), high-grade dysplasia as compared to low-grade dysplasia (RR 2.89; 95% CI 1.88–4.44), and villous/tubulovillous as compared to tubular histology (RR 1.75; 95% CI 1.33–2.31). In contrast, multiplicity of ≥3 adenomas (vs <3) and ≥5 (vs <5) were not associated with increased risk of CRC incidence (RR 1.24; 95% CI 0.84–1.83 and RR 1.22; 95% CI 0.49–3.06, respectively) over a mean follow-up period of 5.4 years in the pooled studies. Notably, SPs were not included in the analysis and when patients had SPs and adenomas, risk stratification was based on features of the adenomas only.

To specifically evaluate the risk of SPs and metachronous advanced neoplasia (AN), Jung and colleagues pooled 11 studies comprising a total of 1,079,315 patients [36]. When comparing individuals with only SSLs to those with only adenomas, they found no significant difference between risk of metachronous AN (OR 0.91; 95% CI 0.23–3.63) and risk of CRC (OR 1.11, 95% CI 0.42–2.97) although significant statistical heterogeneity was observed among studies. Synchronous SSLs and any adenomas, particularly if HRAs had a higher risk of metachronous AN compared to HRAs alone (OR 3.10, 95% CI 1.92–4.99). Limited information on number and size of SSLs, limited studies with TSAs, and poor histopathological agreement regarding SSLs vs HPs are notable limitations of this analysis. To address the gap of CRC risk in the setting of TSAs, we highlight a retrospective cohort study [37] published more recently than the studies included in the review by Jung et al. This study by Trivedi et al, comprising 853 Veterans with TSAs on index colonoscopy between 1999 and 2018, showed cumulative 5-year CRC incidence of 1.34% (95% CI 0.67–2.68%) and CRC mortality of 0.12% (95% CI 0–0.35%) [37]. Compared to patients with normal index colonscopy, those with TSA had a higher risk of CRC (HR 3.70; 95% CI 1.63–8.41) but no significant difference in mortality (HR 1.16, 95% CI 0.10–7.26). Compared to those with adenomas on index colonoscopy, patients with TSA had similar CRC incidence (HR 0.86, 95% CI 0.45–1.64) and mortality (HR 0.78, 95% CI 0.09–6.53). Among 378 patients who had at least 1 surveillance exam, 65% of patients with TSAs had advanced neoplasia.

Another systematic review by Baile-Maxia et al, comprising 14 studies and a total of 493,949 patients with a median follow-up period of 4.9 years, compared metachronous CRC incidence between patients with advanced and nonadvanced SPs, those with nonadvanced SPs and normal colonoscopies, and assessed the association of specific characteristics of SPs on metachronous CRC risk [38]. They defined SPs as including SSLs, TSAs, and proximal HPs. Among patients with advanced SPs (SSL, TSA, or proximal HP ≥10 mm or with dysplasia), the overall CRC incidence was 2.09 per 1000 person-years (95% CI 1.29–2.90), compared to 0.50 per 1000 person-years (95% CI 0.35–0.66) among patients with nonadvanced SPs, and 0.44 (95% CI 0.41–0.46) among those with normal colonoscopies. When comparing specific characteristics of SPs, metachronous CRC incidence rates were highest in SPs with dysplasia (5.86 per 1000 person-years; 95% CI 2.16–9.56), followed by SPs ≥10 mm (1.52; 95% CI 0.78–2.25), proximal SPs (1.18; 95% CI 0.77–1.60), and multiplicity of >3 SPs (0.52; 95% CI 0.08–1.12).

Based on these observational data, several trends are apparent. The first of these is the overall low risk of metachronous CRC in the setting of nonadvanced (vs advanced) adenomas and small SPs on index exam. The second is that risk of metachronous CRC among individuals with advanced adenomas and large SPs is likely more associated with size and advanced histology rather than multiplicity (i.e., multiple, small polyps). Third, prospectively collected data, contemporaneous to currently established standards for high-quality colonoscopies, are needed to confirm these CRC risk estimates as consideration is given to further extension of surveillance intervals and noninvasive options for surveillance. Fourth, while categories of low- and high-risk polyps are generally agreed upon, worldwide standardization of terminology based on size, number, and histology may more accurately inform risk estimates in pooled analyses. Finally, risk estimates in these observational studies are determined based on follow up after resection of neoplasms that were identified during index and surveillance colonoscopies. Determining factors that increase the risk of progression is perhaps just as essential as determining risk for metachronous CRC risk based on resected polyp features.

5. Risk prediction for surveillance

To date, risk-stratification for surveillance has largely been based on polyp features including size, number, and histology. However, such an approach lacks predictive accuracy for future risk of advanced neoplasia [39]. A growing body of literature shows that factors such as patient- and endoscopist-related characteristics, in addition to polyp characteristics, could improve predictive accuracy and thus the ability to tailor surveillance to individuals who are most likely to benefit [40]. Much like models that predict future CRC risk to guide screening decision-making [41], several risk prediction models are emerging as tools that could guide surveillance decision-making [40,42].

One such model was recently developed to estimate the likelihood of non-CRC related mortality among individuals who had advanced adenomas on index colonoscopy in order to identify those unlikely to benefit from ongoing surveillance [43]. Using a cohort of 2,943 veterans ages 40 years or older who underwent colonoscopy for diagnostic and screening indications at 13 Veterans Affairs Medical Centers between 2002 and 2009, advancing age, Acute Physiologic and Chronic Health Evaluation (APACHE) score 7 or greater, specific comorbidities such as cirrhosis, congestive heart failure, stroke, end-stage renal disease, and hospitalization within one year prior to index colonoscopy were found to be independently associated with non-CRC related mortality within 5 years of colonoscopy. This model was well-calibrated (p = 0.41) to its population and had good discrimination (c-statistic 0.74; 95% CI 0.71–0.76) to predict 5-year non-CRC mortality for individuals with history of advanced adenomas (Table 2). Points were then assigned for each factor and added for a total possible score ranging from 0 to 20 to create low-, intermediate-, and high-risk categories for risk of non-CRC related mortality, with the goal of determining which individuals were most and least likely to benefit from surveillance. Compared to individuals in the low-risk category (0–1 points), patients in the intermediate- (2–4 points) and high-risk categories (≥5 point) had a significantly greater risk for non-CRC mortality (OR 7.19, 95% CI 6.21–8.31 and OR 2.37, 95% CI 2.03–2.76, respectively). Therefore, patients in the high-risk category could forego surveillance colonoscopy while those in the low-risk category would continue surveillance. Application of this model in practice was estimated to reduce unnecessary colonoscopy resource use by 22% without a significant increase in CRC risk.

Table 2.

Selected surveillance risk prediction models.

Study (1st author, year)	Study population	Model variables	Outcome	Conclusions
Lee et al. (2024) [43]	95,001 patients with adenoma on index colonoscopy in the Kaiser health system between 2004–2016	• Age • Diabetes • Index colonoscopy indication • Polyp features	CRC risk prediction by a model including patient and polyp characteristics compared to polyp features only	Model improved predictive accuracy for post-polypectomy CRC (AUC 0.71; 95% CI 0.68–0.74) vs model with polyp features only (AUC 0.61; 95% CI 0.58–0.64)
Levin et al. (2024) [45]	52,848 patients who underwent index colonoscopy in the Kaiser health system between 2014–2019	• Age • Polyp size ≥10 mm • Endoscopist ADR <32.5% • Smoking history	CRC risk as predicted by a model with patient, endoscopist, and polyp features compared to polyp features only	Model improved predictive accuracy for metachronous CRC (AUC 0.73; 95% CI 0.66–0.81) vs model with polyp features only (AUC 0.52; 95% CI 0.45–0.60)
Gupta et al. (2023) [46]	30,897 individuals from 73 VA Medical Centers who had surveillance between 2004–2016	• Age • Sex • Diabetes • Smoking • Endoscopist ADR • Adenoma characteristics	Risk for metachronous advanced neoplasia	Model with patient characteristics, ADR, and adenoma characteristics improved post-polypectomy risk stratification (AUROC 0.63 (95% CI 0.62–0.65)) compared to polyp features alone
Guo et al. (2023) [47]	4,696 with CRC compared to 3,709 controls from a population-based screening study in Germany	• Age • Polyp features • Validated CRC-related SNPs	Metachronous CRC risk	Polygenic risk score can further risk stratify individuals with history of neoplasia by identifying those at high risk for metachronous CRC who would benefit from surveillance
Kahi et al. (2022) [43]	2,943 patients ≥40 years of age who had a colonoscopy at one of 13 VA Medical Centers between 2002–2009	• Age • APACHE ≥7 • Cirrhosis • CHF • Stroke • ESRD • Hospitalization within 1 year prior to colonoscopy	5-year non-CRC mortality for individuals with advanced adenomas	Model well calibrated, p = 0.41, with good discrimination (c-statistic 0.74, 95% CI 0.71–0.76) to predict individuals with high risk for non-CRC mortality who would likely not benefit from ongoing surveillance

VA = Veterans Affairs; APACHE = Acute Physiologic and Chronic Health Evaluation; CHF = Congestive heart failure; ESRD = End-stage renal disease; CRC = Colorectal cancer; CI = Confidence interval; AUC = area under the curve; ADR = adenoma detection rate; SNP = single-nucleotide polymorphism.

Several additional models have been developed to predict future risk of neoplasia using cohorts from large, integrated health systems in the US [44,45]. In one study based on the Kaiser health system, data from 95,001 patients with an adenoma on index colonoscopy performed for screening or diagnostic indications between 2004 and 2016 were used to develop and internally validate a model that could improve risk stratification for surveillance. In their model, Lee and colleagues found that patient age, diabetes, index colonoscopy indication, and polyp features improved the predictive accuracy for post-polypectomy CRC (AUC 0.71; 95% CI 0.68–0.74) as compared to a model using only polyp features (AUC 0.61; 95% CI 0.58–0.64) [44]. Using a separate cohort of 52,848 patients who also underwent index colonoscopy in the Kaiser health system between 2014 and 2019, Levin and colleagues found that a model comprising factors including polyp size ≥10 mm, endoscopist ADR <32.5%, patient age, and smoking history improved performance as compared to a model based on polyp features alone (AUC 0.73; 95% CI 0.66–0.81 vs AUC 0.52; 95% CI 0.45–0.60, respectively) [45].

Finally, another study based on VA data included a cohort of 30,897 individuals who underwent index colonoscopy at one of 73 VA sites, with the finding that increasing age, male sex, diabetes, smoking, and lower endoscopist ADR, in addition to adenoma characteristics (3–4 adenomas, polyp location, size ≥10 mm, villous features) were independently associated with increased risk for metachronous advanced neoplasia [46]. Performance characteristics of this model were more limited with an AUC 0.63 (95% CI 0.62–0.65) as compared to the Lee and Levin models.

In addition to patient characteristics and endoscopist related factors, genetic predisposition has also been studied for personalized surveillance risk stratification. To investigate the impact of integrating a polygenic risk score to post-polypectomy surveillance, Guo and colleagues identified a cohort of 4,696 individuals with CRC and 3,709 controls from an ongoing population-based screening study in Germany, and calculated a polygenic risk score based on a validated set of 140 CRC-related single-nucleotide polymorphisms (SNPs) [47]. Using the Freedman et al model [48] and Pfeiffer et al methodology [49] they determined absolute CRC risk based on the PRS and prior colonoscopy findings. Within categories of low, medium, and high PRS, CRC risk was estimated based on polyp features and interval since index colonoscopy, and absolute CRC risk was then determined for 3-, 5-, and 10-years for men and women 50 years of age with either no polyps, HPs, LRA, or HRA on index exam. Among several findings, notable was that the absolute CRC risk for 50-year-old men and women with no polyps was 0.2%. This risk estimate was equivalent to several subgroups of 50-year-old men in their study including: those with LRA and low PRS at 10 years, LRA and high PRS at 5 years, HRA and medium PRS at 5 years, and HRA and high PRS at 3 years. For 50-year-old women, the CRC risk estimate of 0.2% was equivalent for those with LRA and low PRS at 10 years, LRA and high PRS at 5 years, and HRA and high PRS at 3 years. This study also found that the highest CRC risk, 2.1%, was among 50-year-old men with no prior colonoscopy and high PRS at 10 years. While PRS has the potential to estimate CRC risk more precisely for post polypectomy surveillance at an individual level, the complexity of such an approach and limited availability of genetic predisposition would need to be overcome prior to large scale integration into practice. Further, whether and how much discrimination the PRS adds to polyp-specific and phenotypic features needs to be determined. Table 2 summarizes these selected studies aimed at predicting metachronous CRC for patients undergoing post-polypectomy surveillance.

6. Surveillance considerations for older adults

In addition to polyp features, age is strongly and consistently associated with risk of advanced neoplasia in risk prediction models for post-polypectomy surveillance. Advancing age is also associated with increased risk of complications related to colonoscopy and sedation, morbidity, and non-CRC related mortality, yet there is considerable variation in surveillance recommendations in practice [43,50,51]. In the US, estimated life expectancy for men and women combined is 77.5 years [52], and while screening those who have not undergone screening and are in otherwise good health at this age or older has been proven beneficial [53], post-polypectomy surveillance, even for those with low-risk neoplasia, continues for many far beyond this age [50].

When comparing risk of CRC at time of surveillance colonoscopy among older adults (>70 years of age) with those 50–70 years of age, Williams et al found that older adults were at increased risk compared to younger individuals (absolute risk 2.1% vs 1.4%, respectively; risk ratio 1.5, 95% CI 1.1–2.2) in the pooled analysis [54]. In a population-based study of individuals 70–85 years of age who underwent post polypectomy surveillance between 2017 and 2019 in a large, integrated health system, overall CRC incidence was 0.3% at time of surveillance with no significant increase based on advancing age group 70–74 years (0.2%), 75–79 years (0.4%), 80–85 years (0.4%) (p = 0.12) [55]. As expected, CRC risk was higher among individuals with history of advanced adenoma as compared to non-advanced adenoma. Sessile serrated polyps were not included in these risk estimates. Patients with a personal history of CRC, hereditary CRC syndromes, IBD, colectomy, or prior colonoscopy that was incomplete or showed a poor bowel preparation, were excluded. Limitations of this study include potential underestimation of CRC incidence in this cohort as a result of lack of inclusion of patients who may have been due for surveillance, but did not undergo follow-up due to advanced comorbidities, or alternatively, age may be less important and no longer related to CRC incidence after index colonoscopy findings are known and categorized. In contrast, a smaller, retrospective cohort study of 574 adults 75 years of age or older in a South Australian CRC surveillance program found CRC incidence of 1.6% with advanced neoplasia diagnosed in nearly 40% of participants undergoing surveillance [56]. Patients with a history of IBD, and those with a prior colonoscopy that was either incomplete or had a poor bowel preparation, were excluded. Previous personal history of CRC, age ≥85 years, and active smoking were independently associated with CRC. Based on these findings, the authors concluded that continuation of surveillance beyond age 75 years may be considered for individuals with a history of CRC and for active smokers.

To better understand patient and provider perspectives for stopping surveillance for older adults, Schifferdecker and colleagues performed a single-center, qualitative study comprising semi-structured interviews of patients ≥65 years of age enrolled in surveillance, primary care providers, and gastroenterologists [57]. Overall, they found that patients and providers were open to the concept of stopping surveillance. They found that patients and PCPs generally considered age and comorbidity in decision-making while gastroenterologists also considered these factors along with prior polyp characteristics. Providers generally agreed with stopping surveillance for older adults in poor health and that PCPs were in a better position to advise on the decision to stop surveillance, but system-level barriers such as automated surveillance scheduling would need to be overcome.

Thus far, among the various professional societies that have developed post-polypectomy surveillance guidelines, the BSG/ACGBI/PHE guidelines explicitly state to stop surveillance for estimated life-expectancy <10 years or if age is greater than 75 years [19]. The expert panel further recommended use of life-expectancy calculators, such as ePrognosis (https://eprognosis.ucsf.edu/index.php), to guide shared decision-making for surveillance cessation. The ESGE also recommends stopping surveillance for individuals at age 80 years, or earlier if life expectancy is limited by comorbidities, acknowledging that this recommendation is based on low quality evidence [18]. However, de-intensification of surveillance, even among those of older age and greater comorbidity who have low-risk neoplasia [58], and the uptake of life-expectancy calculators in cancer prevention decision-making have been met with some resistance [59]. Further efforts are needed to ensure that such tools are accurate, acceptable and generalizable to the population to which they are applied.

7. Clinical trials of surveillance

The National Polyp Study, organized in 1978, was the first randomized trial in the US to evaluate post-polypectomy surveillance [10]. This trial showed no difference in high-risk features of metachronous polyps found at 1- vs 3-year surveillance intervals for individuals with adenomas on index exam. Until these results were published in 1993, surveillance intervals were variable and often performed at 1-year intervals. With growing evidence for low risk of future CRC at based on observational data, additional clinical trials are ongoing to determine the optimal surveillance approach for individuals with advanced and non-advanced neoplasia.

The first of these is the European Polyp Surveillance (EpOS) trial, which comprises three studies in which more than 20,000 participants ages 40–74 years were randomized to different surveillance intervals across 8 countries in Europe between 2015 and 2024 [9]. The three trials included are the following: 1) EpOS I in which patients with LRAs were randomized to undergo surveillance at 5- and 10-years, or surveillance at 10 years, after index colonoscopy; 2) EpOS II in which patients with HRAs were randomized to undergo surveillance at 3, 5, and 10 years, or at 5 and 10 years, after index colonoscopy; and 3) EpOS III in which patients with SPs undergo surveillance at 5 and 10 years after index colonoscopy. The primary outcome for each trial is CRC incidence at 10 years of follow-up. Recruitment is now complete for the EpOS trials with results forthcoming.

Another ongoing trial in Europe is the Study to Compare Fecal Immunochemical Test with Endoscopic Surveillance After Advanced Adenoma Resection in Fecal Immunochemical Test Colorectal Cancer Screening Programs (Polyprev), in multiple centers in Spain [60]. Included in this trial are individuals 50–65 years of age who underwent CRC screening with FIT which led to a colonoscopy in which at least one AA (size ≥10 mm, tubulovillous/villous histology, or high-grade dysplasia) or at least 3 non-advanced adenomas were completely resected. The comparison is surveillance colonoscopy vs annual FIT for up to 10 years. Participants in the FIT arm will be referred for colonoscopy for fecal hemoglobin ≥10 mcg/g. If the colonoscopy is incomplete or a lesion is requiring endoscopy surveillance is completely resected, FIT-based surveillance will resume in one year. If the follow-up colonsoscopy after an abnormal FIT is complete or showed 1–2 non-advanced adenomas, surveillance will resume beginning with colonoscopy in 5 years. The primary outcome will be 10-year CRC incidence.

In the US, similar to EpOS I, the Five or Ten Year Colonoscopy for 1–2 Non-Advanced Adenomatous Polyps (FORTE) trial is ongoing across 165 sites in the US [7]. In this trial, individuals 50–70 years of age who have non-advanced adenomas or SSLs during a colonoscopy performed within 4 years of enrollment, are randomized to undergo surveillance in 5 and 10 years (arm 1) or 10 years (arm 2). The primary outcome includes CRC incidence as determined at 10 years along with incidence of advanced neoplasia at time of follow-up colonoscopies and CRC-related mortality. The second ongoing clinical trial in the US is the recently initiated, Colonoscopy vs Stool Testing for Older Adults with Colon Polyps (COOP) trial [8]. In this non-inferiority, pragmatic clinical trial, individuals 65–82 years of age with a history of non-advanced adenomas or SSLs are randomized to annual FIT for 5 years, or colonoscopy, followed by a 5-year observational phase to determine incidence of advanced neoplasia up to 10 years following enrollment.

In Australia, the FIT For Purpose trial is an ongoing randomized controlled trial to personalize surveillance colonoscopy for individuals at increased risk of CRC [61]. In this trial, individuals 18 years of age or older enrolled in the Southern Cooperative Program for the Prevention of Colorectal Cancer (SCOOP) program, who have a low FIT fecal hemoglobin of <2 mcg/g within the last 3 years, and due for surveillance colonoscopy will be invited to participate. Patients enrolled in the SCOOP program are those that are considered to be high-risk for CRC with either a personal history of colon neoplasia (adenoma, SSL, and/or CRC), a family history of first degree relative (FDR) with CRC diagnosed at age <55 years, and/or 2 FDR, or 1 FDR and 1 second degree relative with CRC diagnosed at any age. For the trial, those who are enrolled will be randomized to either the control arm (surveillance colonoscopy to be performed as scheduled) vs the intervention group, in which surveillance colonoscopy will be delayed either by 1 or 2 years for those due for 3-year or 5-year surveillance, respectively. For those in the intervention arm who were due for a 3-year surveillance and have been delayed for surveillance colonoscopy by 1 year, an interval FIT will be offered at the time of the original surveillance due date. For those in the intervention arm who were due for a 5-year surveillance colonoscopy, two interval FIT will be completed, one on the original survieillance due date and the second 12 months later. If the interval FIT is positive (fecal Hb >20 mcg/g) an early colonoscopy will be performed. The primary outcome of this trial is advanced neoplasia (advanced adenoma or CRC) incidence.

While these clinical trials are ongoing, a recently published study highlights the potential for noninvasive post-polypectomy surveillance strategies [62]. In their cross-sectional observational study of individuals 50–75 years of age entered in post-polypectomy surveillance, Carvalho and colleagues found that stool-based tests (FIT and multitarget stool DNA) were non-inferior to colonoscopy when modeling long-term CRC incidence and mortality while substantially reducing the number of colonoscopies needed.

8. Conclusion

Considerable progress has been made in CRC prevention, from screening to surveillance since the inception of colonoscopy and polypectomy (Table 3). Colorectal cancer risk based on polyp features is well-established but further refinement of optimal surveillance strategies continues. As evidenced by evolving guidelines, there has been a shift toward risk-stratified surveillance intervals that are based on number, size and histology of the polyps alone. While most current guideline recommendations are based only on polyp features, some countries have started to incorporate noninvasive surveillance for those who have low-risk neoplasia. As data integrating additional factors such as patient phenotype and potentially genotype, combined with endoscopist related factors and colonoscopy quality become available, personalized surveillance will likely continue to evolve to better delineate who should continue surveillance, and for older adults, who no longer requires surveillance. Such a shift would allow for an increase in colonoscopy resource availability to individuals who are at highest risk and most likely to benefit, while reducing unnecessary costs and decreasing exposure to unnecessary risks for those who are least likely to benefit.

Table 3.

Takeaway points.

• Worldwide, post-polypectomy guidelines consistently recommend intensive surveillance colonoscopy for advanced colorectal neoplasia but show variation in features for non-advanced neoplasia, in recommended surveillance frequency, and in test choice for non-advanced neoplasia surveillance.

• Several, large observational cohort studies show high-risk of metachronous CRC among individuals with advanced adenomas at baseline colonoscopy, but not among individuals with low-risk adenomas.

• Ongoing clinical trials in Europe, the US, and Australia will contribute to optimizing post-polypectomy surveillance intervals and test modalities based on polyp features at baseline colonoscopy.

• Further studies are needed to determine whether, and how, patient specific characteristics and life expectancy may be used to personalize post-polypectomy surveillance beyond polyp features and the effectiveness of noninvasive tests for surveillance.

9. Future perspectives

While we await results from ongoing clinical trials, evidence based on observational studies supports intensive, and in some cases (i.e., LRAs) less intensive post-polypectomy surveillance and continues to impact practice. The ongoing post-polypectomy surveillance trials will further add to our knowledge of optimal surveillance intervals and provide evidence for potential noninvasive alternatives to surveillance. However, several areas of uncertainty remain and warrant further research. These include obtaining a better understanding of risk of metachronous advanced neoplasia after SPs are removed, especially among individuals who have synchronous SPs and adenomas. Recognition of the serrated pathway in colorectal carcinogenesis is well-established, but cohort studies for post-polypectomy surveillance have largely focused on adenomas. Data are also needed for the risk of metachronous advanced neoplasia among individuals under the age 50 (e.g., early onset) who have neoplasia on index colonoscopy, to determine optimal surveillance intervals. And finally, among the various factors that have been studied such as polyp characteristics, patient phenotype (demographics, comorbidities, lifestyle risk factors), genotype, and endoscopist-specific factors, further studies are needed to determine which of these are most informative for surveillance decision-making and how they could be integrated into risk prediction models. As risk prediction models gain increasing attention, further efforts are needed to ensure their validity, generalizability, and incremental discrimination when compared with polyp factors alone, as well as their integration into clinical practice. These considerations are important for determining optimal intervals and modalities (invasive vs noninvasive) for individuals who continue surveillance and for determining when to stop surveillance, particularly among older adults. Finally, with increasing technological advancments, our ability to detect small polyps has improved leading to an increased proportion of patients entering surveillance. With evidence showing the low risk of metachronous CRC among individuals with low-risk polyps, further research is needed to risk-stratify individuals who would benefit from surveillance. Table 4 highlights a framework for future research efforts to further optimize post-polypectomy surveillance.

Table 4.

Agenda for future research.

• Improve risk-stratification based on polyp features including:

• Multiplicity

• Sessile serrated lesions

• Synchronous sessile serrated lesions and adenomas

• Further characterize individuals most likely to benefit from surveillance based on:

• Patient-specific factors (demographics, genotype, lifestyle risk factors, life expectancy)

• Yield of second surveillance exam

• Validation of existing risk-prediction models to determine the utility of personalized surveillance recommendations

• Determine optimal surveillance intervals and test options based on individualized risk.

• Establish a standardized approach to decision-making for surveillance cessation.

10. Article highlights

10.1. Introduction

• Up to one third of patients undergoing colonoscopy are found to have neoplasia that requires ongoing follow-up, or surveillance.

10.2. Current guidelines

• Surveillance guideline recommendations assume that a high-quality colonoscopy with a good bowel preparation has been performed.

• Surveillance recommendations are based on polyp size, number, and histology.

• Colonoscopy is currently the only recommended surveillance test in the U.S.

• Some regions in Canada recommend the fecal immunochemical test in 5-years for individuals with non-advanced adenomas while other regions in Canada, Europe, and Asia recommend a return to screening for individuals with small, non-advanced adenomas and SSL.

• The global trend has been to de-intensify surveillance for individuals with low-risk polyps.

10.3. Neoplasia and long-term colorectal cancer risk

• Data from observational cohort studies show that individuals with 1–2 non-advanced adenomas are at low risk for metachronous CRC, while those with advanced adenomas are at higher risk, supporting more intensive surveillance for the latter.

• Evidence for long-term CRC risk for individuals with SSL is more limited than for adenomas; however, those who have large or dysplastic SSLs are at high risk.

10.4. Risk prediction for surveillance

• In addition to polyp features, several models that include patient- and endoscopist-specific factors, have been developed to personalize surveillance colonoscopy.

• Further validation and testing in the real-world setting are needed prior to the integration of these models into practice.

10.5. Surveillance considerations for older adults

• While age is a risk factor for neoplasia, advancing age is also associated with greater risks with colonoscopy and sedation, and mortality related to causes other than CRC.

• There is limited guidance on when to stop surveillance for older adults.

• British and European guidelines recommending stopping surveillance for individuals who are 75 or 80 years of age, respectively, or earlier based on life expectancy.

10.6. Clinical trials

• There are several ongoing randomized controlled trials related to colorectal neoplasia surveillance around the world: 3 in Europe, 2 in the U.S., and 1 in Australia.

• These trials will advance our knowledge about optimal frequency of surveillance for colorectal neoplasia, and whether noninvasive tests (compared to colonoscopy) can be used for surveillance, particularly for individuals with non-advanced neoplasia.

10.7. Future perspectives

• Studies are needed to better understand long-term CRC risk after removal of SSLs, and after removal of synchronous adenomas and SSLs.

• Studies are needed to understand metachronous CRC risk for individuals who have early onset (i.e., age <50 years) of neoplasia.

• Studies are also needed to identify risk factors for neoplasia progression.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/1758194X.2025.2525070.