Screening methods and strategies of choice

Statement 1. Risk stratification and sequential offering of tests are reasonable approaches to CRC screening in Asia.

Level of agreement: A=54.1%, B=41.7%, C=4.2%, D=0%, E=0%

Quality of evidence: I

Classification of recommendation: A

The incidence of CRC increases significantly after the age of 50 in most of the developed countries. Offering direct colonoscopy screening for the entire eligible population above 50 years of age is, however, not feasible in the majority of the countries in Asia, where the single-payer healthcare system is the mainstay, with constraints in funding, clinical capacity and manpower, and where public awareness of CRC and its screening is still low.1 Moreover, considering the prevalence of 2 to 3 per 1000 for CRC and 3–10% for advanced adenoma in the general screening population, a risk-based stratification and sequential offering of a screening test is a more pragmatic and efficient approach.5 A faecal immunochemical test (FIT) is currently the most popular screening test used in organised screening programmes, including the Asia-Pacific region.5 The positivity rate of a FIT, which is largely affected by the population risk of colorectal neoplasm and the FIT cutoff value used to determine positivity (based, for example, on the colonoscopy capacity in individual programmes), is generally within the range of 4% to 10%. In a recent Polish study in which 12 485 individuals were invited to either direct colonoscopy screening or a sequential FIT/colonoscopy screening or a choice between FIT and colonoscopy screening, the participation rate was found to be significantly higher in the sequential or choice strategy.6 A randomised controlled trial from China has provided strong evidence that a risk-adapted screening strategy with colonoscopy and FIT in colorectal cancer screening can achieve much higher participation.7 Therefore, using risk stratification followed by colonoscopy if the test is positive, will significantly improve screening uptake and reduce the colonoscopy burden of a screening programme.

Statement 2. Multiple risk-stratification systems have been developed and no single stratification method has proved to be better than the others

Level of agreement: A=83.3%, B=16.7%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

In the Asia-Pacific region, FIT is the most popular test for identifying individuals in the general population at high risk for advanced neoplasia (AN). In countries with rising CRC incidence but with no organised screening programmes yet in place, risk-score systems using simple demographics, family history or other CRC risk factors for triaging whether these individuals warrant a colonoscopy would be good alternatives to FIT and could help to make the most efficient use of the constrained endoscopy resources. The Asia-Pacific Risk Score was developed and validated in 15 Asia-Pacific countries/regions.8–11 Using this risk-score system, subjects in the middle- and high-risk tiers had 2.6-fold and 4.3-fold increased prevalence of AN/CRC, respectively. When used in combination with FIT, the risk score could detect around 70% of subjects with AN and 95% of subjects with cancers, both significantly higher than FIT screening alone, and hence reducing colonoscopy demand by half compared with direct colonoscopy screening.12 Several other risk-score systems were also developed, both in Asia and in other continents, using similar risk factors such as age, gender, family history, smoking, drinking and body mass index in different combinations and weighting. No single risk-score system was proved to be better than the others in head-to-head comparison.13 As the American Gastroenterological Association (AGA) white paper stated, “a one-size-fits-all approach to CRC screening has not and is unlikely to result in increased screening uptake or desired outcomes owing to barriers stemming from behavioural, culture and socioeconomic technologies".14 Six centres from China including 19 546 subjects also showed that a risk-adapted screening strategy (ie, high-risk subjects for colonoscopy and low-risk subjects for a FIT) had a much better participation rate.7 In the Asia-Pacific region, there is a broad variety of disease prevalence, healthcare funding models, capacity and manpower issues, which may affect the choice of risk stratification approach and selection of CRC screening strategies. The Asia-Pacific Working Group recommends that any of these risk stratification systems should be adopted after validation in local regions.

Statement 3. Quantitative faecal immunochemical test (every year or every 2 years) or colonoscopy (every 10 years) are the preferred screening tests in Asia.

Level of agreement: A=45.9%, B=45.8%, C=8.3%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

Accumulating evidence demonstrates the effectiveness of a FIT and screening colonoscopy to reduce CRC incidence and death.15–19 The time intervals for FIT and colonoscopy are based on large-scale, long-duration, follow-up cohort studies. The use of an annual faecal occult blood test significantly reduces the incidence of colorectal cancer.20 The effect of faecal occult blood testing on CRC is long lasting. The reduced mortality related to CRC is found to persist after 30 years of follow up.21 A simulation model of CRC screening by the US Preventive Services Task Force (USPSTF) also confirmed that 1–2 yearly screening is better than 3-yearly screening.22 A retrospective cohort study from northern California that included over 1 million individuals confirmed that a negative colonoscopy results in reduced risk of colorectal cancer and related death for more than 12 years after examination compared with unscreened patients.23 A Polish observational study also confirmed that after a single negative screening colonoscopy, profound and stable reduction in CRC incidence and mortality was documented for up to 17 years.24

In the Asia-Pacific region, organised nationwide screening programmes are in place in Australia, Brunei, Hong Kong, Japan, South Korea, New Zealand, Singapore and Taiwan, and all these programmes use FIT as the primary screening test. In the opportunistic screening settings, colonoscopy is more often used as primary screening tool.5 A recent cohort study with 10-year follow-up from Taiwan revealed that biennial FIT screening could significantly reduce advanced stage CRC incidence by 29% and CRC mortality by 35%.16 Flexible sigmoidoscopy is currently not used in organised screening programmes in this region, although its effectiveness in reducing CRC incidence and mortality in the distal colon is demonstrated by randomised trials. This is probably related to increased accessibility to colonoscopy and the concern about increasing prevalence of proximal colorectal neoplasms.25–28 Guaiac-based faecal occult blood tests are no longer used and have been largely replaced by FIT owing to its convenience of use, high screening uptake, and better sensitivity and specificity for CRC. Therefore, the working group recommends quantitative FIT (every year or every 2 years) or colonoscopy (every 10 years) as the preferred screening tests in Asia.

Statement 4. There are insufficient data for stool- or blood-based molecular tests and capsule colonoscopy and CT colonoscopy in Asia to be recommended as a primary screening tool.

Level of agreement: A=87.5%, B=12.5%, C=0%, D=0%, E=0%

Quality of evidence: III

Classification of recommendation: A

A multitarget stool-based DNA test was developed in the United States and approved by the Food and Drug Administration. Its sensitivity in detecting CRC and adenomas with high-grade dysplasia was 92.3% and 69.2%, respectively, both significantly higher than with a FIT.29 A recent comprehensive economic appraisal comparing different CRC screening strategies also showed that its cost-effectiveness may be comparable to that of annual FIT screening.22 There are, however, still significant concerns about using it for large-scale population screening owing to the unclear screening uptake by the public, inconvenience of stool collection and shipping, high cost and the lack of evidence on its programme sensitivity demonstrating its superiority over FIT.22 To date, no validation studies on multitarget stool-based DNA testing have been conducted in Asia. Blood-based tests using cell-free DNA, proteins, messenger RNA, micro-RNA, or even circulating tumour cells as the target biomarkers were also developed, and some of the studies have shown promising results but validation studies in large screening populations are still lacking which hinders their roles as a frontline screening test at this moment.30

In a recent meta-analysis that included studies conducted in screening populations, capsule colonoscopy detected 93% of the CRCs (using colonoscopy as the gold standard), and its sensitivity for detecting polyps ≥10 mm was 84% to 97% with a specificity of 91% to 99%, showing its potential as an alternative to colonoscopy.31 However, the variation in completion rate (57%–92%) of capsule colonoscopy examination of the entire colon is a concern. In comparison with colonoscopy, CT colonography (CTC) has a comparable detection rate for AN.32 However, higher screening participation rates of CTC led to diagnostic yields comparable to colonoscopy in a randomised controlled trial.33 CTC is now accepted in North Amercian and European countries as an alternative examination to colonoscopy, as the diagnostic test after a positive FIT. In the United States, CTC is used as the primary screening modality or when colonoscopy is incomplete. An important limitation of CTC is the need for excellent bowel preparation. CTC has also had significant false-positive reults due to preparation problems, which reduces the positive predictive values. A study from nationwide multicentre prospective controlled study from Japan compared the performances of CTC and colonoscopy on the same day in a non-screening population with images interpreted by trained gastroenterologists and radiologists. Results showed that CTC is not sufficiently sensitive to detect non-polypoid neoplasms.34 Most cost-effectiveness studies show that CTC is not cost-effective relative to other screening programmes.35 In the Asia-Pacific region, many countries have limited access to CTC. Therefore, the working group cannot recommend a stool-based DNA test, capsule colonoscopy or CTC as a first-line screening tool for CRC in the Asia-Pacific region.

Age of screening for CRC

Statement 5. There is a rising trend of young patients (<50 years of age) with CRC in Asia, especially in men and for cancers located in the rectum.

Level of agreement: A=70.8%, B=25.0%, C=4.2%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

The rising trend for young patients below the age of 50 years to develop CRC, which was first reported by Siegel et al,36 is no longer a western phenomenon. Studies in Asia have reported a similar trend of early-onset CRC. In a multinational cohort study involving four Asian countries/regions—namely, Taiwan, Korea, Japan and Hong Kong, a similar trend of increasing incidence of early-onset CRC was reported37 (table 2). The most pronounced change was observed for male rectal cancer, although a similar trend was also evident among women and in the proximal colon. Subsequently, the rising trend of early-onset CRC has also been reported in other Asia-Pacific countries, including Korea,38 Australia39 and India.40 Using a population-based, age-period, cohort study to compare the incidence of CRC in Asia with that in the West, again showed an increasing risk of patients with early-onset CRC in both Western and Asian populations.41 The rising trend of early-onset CRC is more evident in countries of high economic status, although a global phenomenon.42 Risk factors associated with early-onset CRC include male gender, inflammatory bowel disease, family history of CRC, obesity, smoking and diabetes.43 44 The recent rapid increase in incidence may also be related to lifestyle changes, such as intake of simple sugar and sweetened beverage during adolescence.45 This is an alarming trend that should alert healthcare systems in the Asia-Pacific region to take note and be prepared. It is imperative to look into the effects of lifestyle changes in preventing early-onset CRC.46 Discussion of the age at which CRC screening should be started is naturally becoming a hot topic of debate.

Statement 6. Lowering the age to start CRC screening to <50 years has not been proved to be cost-effective in Asia.

Level of agreement: A=37.5%, B=45.8%, C=4.2%, D=12.5%, E=0%

Quality of evidence: III

Classification of recommendation: B

With the increasing global incidence of early-onset CRC, the need for lowering the age at which CRC screening is started has been a focus of discussion. The US Multi-Society Task Force on Colorectal Cancer recommended that screening should begin at the age of 50 in people at average risk, except in African Americans in whom limited evidence supported starting screening at age 45.47 The American Cancer Society was the first to propose lowering the screening age for all adults to 45 years.48 The most recent USPSTF recommendation and the US Multi-Society Task Force recommend that asymptomatic adults aged 45 years or older at average risk of CRC (ie, no prior diagnosis of CRC or advanced adenoma, inflammatory bowel disease and no family history of known genetic disorder predisposing to CRC) should receive CRC screening.49 50 This is based on age-period cohort analyses in birth cohorts in the United States and clinical trials of CRC screening that include adults younger than 50 years of age.51 52 Using screening colonoscopy in a Markov model analysis, a study from the United States showed that a screening programme starting at the age of 45 years is likely to be cost-effective.53 Furthermore, modelling performed by the Cancer Intervention and Surveillance Modelling Network suggests that starting CRC screening at age 45 years may moderately increase life-years gained and decrease CRC cases and deaths compared with beginning screening at the age of 50.22 However, modelling studies are based on assumptions that are tailored to specific socioeconomic factors—for example, life-years gained, costs of a screening test and treatment costs for CRC in a specific region or country. The microsimulation modelling analysis, which suggests that CRC screening using a stool test, colonoscopy or CT colonography starting at the age of 45 years provides an efficient balance of colonoscopy burden and life-years, gained may not apply to the Asia-Pacific region. The only study in the Asia-Pacific region comes from Australia, which used a microsimulation model to study the cost-effectiveness of starting screening at 40–45 years and cessation at 79–84 years.54 The study concluded that starting screening at 45 years could be cost-effective, but would increase colonoscopy demand, jeopardising the benefit of those subject to screening at 50–74 years. The working group chose to recommend that, in view of the vast population in this region, priority should be given to those over the age of 50 years as their risk of CRC is still significantly higher than that of the average-risk population below the age of 50.

There are two caveats to this recommendations. First, the modelling studies are mainly based on colonoscopy-based screening scenarios. It is not clear whether it also applies in a two-tier screening setting using FIT and colonoscopy as a sequential screening tool. Also, given the possibility that young-onset CRC continues to increase in the Asia-Pacific region, further studies in this region using FIT-based screening starting at a younger age may be worthwhile.

Statement 7. Older age per se should not be a reason to exclude healthy individuals from starting screening.

Level of agreement: A=54.2%, B=29.2%, C=8.3%, D=8.3%, E=0%

Quality of evidence: III

Classification of recommendation: C

There are few data addressing the issue of the upper age limit for CRC screening. Most studies did not include people over the age of 75 years. The risk of screening colonoscopy increases with age and therefore, it is not recommended after age 75. The incidence of CRC and adenomas increases with age, but the use of non-invasive screening tests and a sequential screening approach (eg, FIT plus colonoscopy) might mitigate the risk in the older subjects. To discontinue safe, rapid, painless, cost-effective screening with a non-invasive faecal occult blood test, FIT, or stool-based multitarget DNA, or minimally invasive virtual colonoscopy and capsule colonoscopy for individuals older than 75 is neither supported, nor refuted by the literature. This strategy will, however, deny individuals the benefits of prevention and early diagnosis of cancers. CRC or adenomas identified by screening could be addressed with an evidence-based decision, considering the individual’s risk tolerance and desires of the patient and family. The Asia-Pacific Working Group considered the ageing population of Asia but also the general improved health status of adults over age of 75, and feels that age alone should not prohibit healthy subjects from screening. Using age alone as a cut-off point for CRC screening is insufficient.55 In determining whether CRC screening should be offered, patients and clinicians should discuss and make the decision together based on a patient’s overall health status, prior screening history and patient’s preferences.

Statement 8. Screening can be stopped at age 85 if an individual has a recent negative test as they are unlikely to benefit from further screening

Level of agreement: A=54.1%, B=33.3%, C=4.2%, D=4.2%, E=4.2%

Quality of evidence: III

Classification of recommendation: C

Typical pathogenesis of CRC is an adenomatous polyp slowly increasing in size and accumulating genetic mutations, leading to dysplasia and ultimately, cancer. Most CRCs arise from colorectal adenomas, but 20–30% of cases arise through pathways other than the adenoma-carcinoma sequence. Progression from adenoma to invasive cancer varies and may take an average of 15 years, and in certain instances progress within a shorter time. The American Cancer Society recommended that adults at average risk, in good health with a life expectancy of more than 10 years, should continue CRC screening until the age of 75 years.48 For individuals aged 76 to 85 years, decisions on CRC screening should be individualised. However, above the age of 85 years, clinicians should discourage individuals from continuing CRC screening. The US Preventive Services Task Force recommendation is to offer screening for adults aged 50 to 75 years. The USPSTF also recommends that clinicians can selectively offer screening for CRC to adults between 76 to 85 years of age.49 Although the current USPSTF guideline recommends against screening in patients older than 85 years, other competing health concerns should be considered, as there are no data to show that the benefits of screening cease at any particular age as the only factor. The working group accepts that screening should be stopped at age of 85 years if an individual has had a recent negative test.

Screening for individuals with a family history of CRC or advanced adenoma

Statement 9. Healthcare providers should always explore the family history of their patients with CRC for the possibility of hereditary CRC syndromes.

Level of agreement: A=91.7%, B=8.3%, C=0%, D=0%, E=0%

Quality of evidence: III

Classification of recommendation: A

Approximately 30% of patients with CRC have a family history of this disease.56 Some families with a history of CRC carry genetic variants that cause CRC with high or moderate penetrance, but these syndromes account for only around 5% of CRC cases. Recent germline cancer testings including more extensive panels of moderate risk genes suggest a higher association of 12–15%.57 Most families with a history of CRC and/or adenomas do not carry specific genetic variants associated with cancer syndromes. Family history of CRC includes three categories: (1) those with recognised familial genetic syndromes; (2) those who have first-degree relatives (FDRs) with CRC; their age at diagnosis may be young (<50 years) or older (>50 years) and (3) those having FDRs with documented advanced adenomas or SSLs. A positive family history increases the risk of developing CRC. Furthermore, those with a positive family history tend to develop CRC at a younger age.58

Many national and international guidelines on CRC screening include recommendations on screening of individuals with family history of CRC.47 59 (table 3). Where carriers of penetrant genes are identified (eg, Lynch syndrome, familial adenomatous polyposis, MUTYH- associated polyposis), syndrome-specific guidelines cover the age of onset and frequency of colonoscopic screening and address screening and risk-reducing strategies for non-colorectal syndrome specific cancers. For Lynch syndrome, MMR gene specific guidelines are now promulgated, recognising gender- and gene-specific differences in penetrance. Although the benefits of colonoscopy are currently being questioned owing to the recognition of a non-polypoid pathway to CRC in Lynch syndrome and the possibility of overdiagnosis,60 guidelines are still unanimous on the need for high-quality colonoscopy.

In Asia, very few genetic studies have been conducted specifically for CRC. Possibly a substantial number of high-risk asymptomatic individuals with family history of advanced neoplasia and CRC miss the opportunity of early diagnosis and treatment. Healthcare workers, including gastroenterologists, surgeons, family physicians, and nurses, should explore the family history of CRC in their patients. Studies have shown that both patient’s awareness of the importance of a family history of CRC as well as the recommendation of a family physician to undergo CRC screening are the two most powerful incentives for individuals at average risk to be enrolled in a screening programme.61 Guidelines for surveillance of people with a family history of CRC or advanced adenomas are very much needed.

Statement 10. Subjects having two FDRs with CRC or advanced adenoma diagnosed at any age should start screening at 10 years before the age at diagnosis of the youngest affected FDR or at age 40 (whichever is earlier) with colonoscopy every 5 years

Level of agreement: A=45.8%, B=41.7%, C=12.5%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

Statement 11. Subjects having one FDR with CRC or advanced adenoma diagnosed at age <60 years should start screening at 10 years before the age at diagnosis of the youngest affected FDR or age 40 (whichever is earlier) with colonoscopy every 5 years

Level of agreement: A=41.6%, B=41.7%, C=12.5%, D=0%, E=4.2%

Quality of evidence: II-2

Classification of recommendation: A

Statement 12. Subjects having one FDR with CRC diagnosed at age > 60 years should start screening at the age of 40 with the same test strategy and intervals as for average-risk individuals.

Level of agreement: A=50%, B=45.8%, C=4.2%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

Evidence suggests that having a FDR with advanced adenomas carries a similar increased risk for CRC as having an FDR with CRC. In a blinded, cross-sectional study, 200 asymptomatic siblings of patients with advanced adenomas (adenomas >10 mm, high-grade dysplasia and villous structure) and 400 age- and sex-matched siblings of subjects with normal findings from colonoscopies and no family history of CRC underwent colonoscopy at two hospitals in Hong Kong.62 The prevalence of advanced adenomas was 11.5% among subjects with one FDR diagnosed with CRC and 2.5% among subjects without such a family history (matched ORs=6.05, 95% CI 2.74 to 13.36). On the other hand, non-advanced adenoma do not carry the same risk of CRC to a FDR.63 Therefore, FDRs of those with CRC or advanced adenoma should start screening for CRC at an earlier age.

Starting age for screening depends on the age of diagnosis of the proband in the family and the number of FDRs with CRC or advanced adenoma. In a Utah population-based study, FDRs of patients with CRC had a 1.8-fold increased risk of CRC (hazard rate ratio 1.79, 95% CI 1.59 to 2.03), as did second-degree relatives (hazard rate ratio 1.32, 95% CI 1.19 to 1.47).64 It should be pointed out that the risk of relatives in the Utah study was not age related— that is, the risk, though elevated, was not at an early age if the proband was young. This risk was greater for FDRs when index patients developed CRC at less than 60 years of age than in those older than 60. Individuals who have two or more FDRs diagnosed with CRC or advanced adenoma also have a higher risk than those with only one FDR. Thus, the number of FDRs involved and the age of onset of disease are the two crucial factors to start screening earlier than at age 50.

Although there is clear evidence of an increased risk in this group, data are limited on its magnitude relating to the age of CRC onset of the proband. Therefore, there is a great variety across different guidelines from various national and professional societies (table 3). The definition of who should undergo intensified colonoscopy surveillance instead of average-risk screening also varies widely (table 3). Nevertheless, the consensus accepted starting screening at an age 10 years earlier than the earliest age of onset in the family as the overall risk at an earlier age in these circumstances surpasses the risk a decade or two later in average-risk populations. Thus, the number of FDRs involved and the age of onset of disease are the two crucial factors to start screening earlier than at age 50.

The Asia-Pacific Working Group considers that lacking large-scale cohort studies in the region to make alternative recommendations, we should follow the international guidelines—that is, using the number of FDRs and the age of diagnosis of CRC or advanced adenoma in the FDRs as risk stratification criteria. Subjects having two FDRs with CRC or advanced adenoma, diagnosed at any age, are recommended to start screening at 10 years before the age at diagnosis of the youngest affected FDR or at age 40 (whichever is earlier) with colonoscopy every 5 years. Subjects having one FDR with CRC or advanced adenoma diagnosed at age <60 years should start screening at 10 years before the age at diagnosis of the youngest affected FDR or age 40 (whichever is earlier) with colonoscopy. These two are considered the groups at highest risk. Subjects having one FDR with CRC diagnosed at age >60 years are considered lower risk and should start screening at the age of 40 with the same test strategy and intervals as for average-risk individuals—that is, colonoscopy or FIT is acceptable.

Surveillance for individuals with one or more adenomas

Statement 13. Surveillance colonoscopies represent a heavy burden to the healthcare system and to individuals and the optimal method for surveillance and intervals should be studied in Asia.

Level of agreement: A=87.5%, B=12.5%, C=0%, D=0%, E=0%

Quality of evidence: III

Classification of recommendation: A

Individuals with adenomas of different endoscopic and pathological features at colonoscopy carry different risks of developing metachronous advanced neoplasm. Surveillance colonoscopy provides these individuals with additional protection for incident CRC. Occurrence of CRC after colonoscopy could be secondary to inadequate quality of colonoscopy or be due to specific tumour biological behaviour (eg, MSI tumours). Surveillance colonoscopy provides an opportunity to detect overlooked or newly developed neoplasms before they become invasive or advanced-stage CRC.65 The frequency of undergoing surveillance colonoscopy should be tailored according to the individual risk of developing metachronous CRC, thereby making the most efficient use of the constrained resources such as colonoscopy capacity. A remarkable degree of overuse or underuse of surveillance colonoscopy exists and needs to be optimised based on the current evidence and the clinical guidelines.66–68 In addition, surveillance guidelines in the Asia-Pacific also require revision based on the most updated data on colonoscopy use and outcomes.69

A large body of literature has shown that those with three or more adenomas or adenomas with advanced histology are at higher risk of developing metachronous advanced neoplasms after colonoscopy.70–73 These tumours are more often flat, proximally located and likely to be missed or incompletely resected.74 More recent studies show that individuals with low-risk adenomas (defined as one or two tubular adenomas <10 mm) followed up for 7–15 years have an incidence of CRC comparable to those with no adenoma, and significantly lower than for those with high-risk adenoma (defined as three or more adenomas or adenomas with villous or high-grade dysplasia histology or size >10 mm).75–78 The European guideline (ESGE) recommends that five low-risk adenomas should be considered as high-risk instead of three, which further reduce the frequency of surveillance colonoscopy.67 Surveillance intervals could be tailored according to the risk levels to avoid unnecessary colonoscopies. High-quality baseline colonoscopy (complete caecal intubation, adequate bowel preparation, appropriate withdrawal time, adequate adenoma detection rates, complete polyp resection and low adverse events and complications)79 is the premise of recommending different surveillance intervals based on colonoscopic findings. Adenoma detection rates (ADRs), one of the most important metrics of colonoscopy quality, have been shown to be inversely associated with the risk of incident CRC after baseline colonoscopy, incident advanced-stage CRC and even CRC death in a colonoscopist’s practice.80–82 Incomplete endoscopic resection of neoplasms detected at baseline colonoscopy or new-onset neoplasms are other causes for the development of postcolonoscopy CRC.83–86

Statement 14. After removal of high-risk adenomas, surveillance colonoscopy should be done at 3 years

Level of agreement: A=58.3%, B=25%, C=8.3%, D=4.2%, E=4.2%

Quality of evidence: II-2

Classification of recommendation: A

It is well studied that individuals with one or more high-risk adenomas (high-grade dysplasia (HGD) or villous component in histology, 10 mm or larger in size, and three or more in number) are at a higher risk of developing metachronous advanced adenomas or CRC after colonoscopy.87 Approximately 4% to 10% of the colonoscopic screening participants belong to this category and currently, 3-year surveillance is recommended for this subgroup.76–78 88 Such a surveillance interval is based on the natural history of the colorectal neoplasia, with the average dwelling time of 3 to 4 years for an advanced adenoma to progress to invasive cancer, and the results of previous randomised trials, including the result from the National Polyp Study.89–92 A recent study by Wieszczy et al demonstrated that those subjects with an advanced adenoma ≥2 cm or HGD are at higher risk of metachronous CRC than those with other adenomas. Such data may help to further reduce the proportion of subjects who require intensive surveillance by threefold.77 A prospective multicentre trial was conducted in 11 Japanese institutions, randomising patients to receive follow-up colonoscopy at 1 and 3 years or at 3 years only after detecting advanced neoplasia at baseline colonoscopy.92 In this study of 3926 patients with a mean age of 57 years, no difference in the detection of adenomas >10 mm, HGD or cancers was found between the two strategies. Based on this study, a 3-year interval is considered appropriate for Asian individuals. The recommended surveillance interval by the Asia-Pacific Working Group after the removal of high-risk adenomas is generally in line with the 3-year surveillance interval recommendations by the Japanese and Korean guidelines.93 94

Statement 15. Surveillance for low-risk adenomas (<3) should be done by colonoscopy and/or FIT at intervals as used for average-risk individuals.

Level of agreement: A=70.8%, B=29.2%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: B

Five-year surveillance intervals have been recommended by major guidelines for individuals with low-risk adenomas. Based on recent studies showing a similar risk of incident CRC after colonoscopy in this low-risk group compared with a negative screening colonoscopy in the average-risk population,76–78 current recommendations suggest prolonging the surveillance interval to 7–10 years, or even the same as the interval for those individuals without adenomas (ie, 10- year interval). It should be noted that such a long surveillance interval is based on a high-quality baseline colonoscopy, and inadequate quality may lead to the occurrence of postcolonoscopy colorectal cancer (PCCRC).

In some FIT screening programmes or guidelines, regular FIT screening is recommended after negative colonoscopy.72 A recent study from Taiwan showed that subsequent FIT after a negative colonoscopy had positivity rate of 11.3%, and for every 12 colonoscopies one CRC was detected. This was associated with a reduction of incident CRC by 53% and ameliorated the risk of PCCRC caused by inadequate colonoscopy quality.95 Therefore, regular FIT in between colonoscopy seems to be a prudent strategy and may add further reassurance about the prevention and early diagnosis of CRC. Using serial FIT in the intermediate-risk group (three to four small adenomas or on e≥10 mm) has also been discussed. Cross et al reported that annual FIT after colonoscopy in such risk groups reduced the volume of colonoscopy by 71% but missed 30–40% of CRCs and 40–70% of advanced adenomas, which would have been detected by colonoscopic surveillance.96 In Asia, where the volume of colonoscopy is high and resources are limited, FIT as an alternative to colonoscopy in surveillance is definitely appealing. The use of a FIT for surveillance instead of colonoscopy-based surveillance needs further evaluation. For the current recommendation, the working group proposes surveillance for low-risk adenomas (<3), which should be performed by colonoscopy and/or FIT at intervals as used for average-risk individuals. Further specific recommendations on FIT surveillance require more data.

Screening and surveillance for sessile serrated lesions

Statement 16. As sessile serrated lesions are premalignant lesions, they should be detected and removed.

Level of agreement: A=83.3%, B=16.7%, C=0%, D=0%, E=0%

Quality of evidence: II-3

Classification of recommendation: B

For a long period of time, colorectal adenomas were thought to be the only precursor lesions for CRC. However, recent studies show that, despite a high adenoma detection rate, PCCRCs still occur, which are more likely to be related to serrated lesions mostly located in the proximal colon.97 New knowledge in histopathology has now clarified that serrated lesions are a heterogeneous group that includes both premalignant lesions and clinically insignificant small hyperplastic polyps. Almost a decade after publishing the fourth edition of WHO classification of tumours of the digestive system, the fifth edition WHO classification in 2019 has adopted the new term SSLs, which refers to serrated premalignant lesions that were formerly called sessile serrated adenomas (SSAs) or sessile serrated polyps (SSPs)98 99 (table 4). The change of terminology from adenoma/polyp to lesions highlighted that these serrated lesions might not show morphologic dysplasia or a polypoid structure, and hence calling them adenomas or polyps might be inappropriate and confusing. Hyperplastic polyps (HPs) are still considered to be benign lesions with no risk of malignant transformation. The subtype mucin-poor hyperplastic polyp has been deleted leaving only microvascular and goblet cell-rich subtypes.100 Yet, the malignant potential of SSL is confirmed and may contribute significantly to PCCRCs.101 Studies comparing prevalent CRC discovered at screening with incident CRC discovered during surveillance, found that the incident CRC tend to be more proximal and have characteristics compatible with SSL with CPG island methylation and MSI-H.97 102 Traditional serrated adenomas (TSAs) are rare but clearly also a precursor of malignancy. Both TSAs and SSLs may progress via the serrated pathway to CRC.100

The diagnostic criteria for SSLs have also been clarified. Instead of ‘two or three’ distorted serrated crypts in a polyp in the old classification,98 the updated 2019 WHO classification99 specified that as a minimum requirement, one distorted serrated crypt is sufficient for the diagnosis of SSL. The definition of an architecturally distorted serrated crypt is now specified as showing at least one of the following histologic features: (1) horizontal growth along the muscularis mucosa (L-shaped or inverted T-shaped crypt); (2) dilatation of the crypt base (basal one-third of the crypt); (3) serrations extending into the crypt base; (4) asymmetrical proliferation of the crypts. Although SSLs have been regarded as a premalignant lesion for CRC, only a small proportion of them show morphologic dysplasia (SSL-D). This suggests that once dysplasia develops within an SSL, the progression to cancer is accelerated.103 There are thought to be three types of dysplastic changes in SSLs104:

1. Intestinal dysplasia: similar to the dysplasia in adenomas, typified by elongated, pseudostratified, hyperchromatic nuclei with basal cytoplasm;

2. Serrated dysplasia: characterised by cuboidal cells with eosinophilic cytoplasm, increased nuclear prominence, atypia and mitosis, and its presence can be considered to represent progression to a TSA.

3. Minimal deviation dysplasia: with limited changes compared with SSL, but with the characteristic loss of MLH1.

Most cases of SSL-D have an undefined pattern of dysplasia, with the majority showing a loss of MLH1 expression. However, the clinical significance of these histologic differences in dysplasia is not fully known. Is the finding of SSL significant? In a matched cohort study through prospective record linkage in Sweden among patients aged 18 years or older with a first diagnosis of colorectal polyps in the nationwide gastrointestinal ESPRESSO histopathology cohort (1993–2016), CRC mortality rates of patients diagnosed with tubulovillous adenomas, villous adenomas as well as SSLs were higher than those in controls.105 Recently it has been shown in a FIT-positive screening population that, just like ADR and alongside the ADR, the proximal serrated polyp detection rate is inversely correlated with the incidence of interval postcolonoscopy CRC.100 Therefore, SSLs are indeed associated with CRC. With these clarifications in criteria of diagnosis and updated classification of SSLs, the Asia-Pacific Working Group feels that it is mandatory for gastroenterologists to detect and remove all SSLs to prevent CRC.

Statement 17. Sessile serrated lesions are likely to be underdiagnosed in Asia and should be carefully looked for during colonoscopy, especially in the right colon.

Level of agreement: A=79.2%, B=20.8%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: B

The serrated pathway is now believed to account for 30% of all cases of CRC but, as detection rates vary widely among endoscopists and pathologists, there is uncertainty about the prevalence of these lesions in our locality.106 Prevalence varies with study location, diagnostic criteria and examination quality. Relatively little is known about the epidemiology of these lesions (prevalence, location, family history) and risk of malignant transformation (timing, associated factors). With the changes in diagnostic criteria (see above) and awareness of the condition, new prevalence studies should be performed. A systematic review of 74 publications on SSPs was done, after screening 4462 papers extracted from the literature.106 Before 2010, few reports on the prevalence of SSPs were published, but this number increased dramatically after release of the WHO 2010 classification. Reported prevalence of SSPs was 3.9% in Europe, 5.1% in the USA and 2.6% in Asia. Australia reported an exceptionally high prevalence of 10.5%106 because of one statistical outlier.103 The US and European studies also showed substantial heterogeneity in estimates of SSL prevalence, probably due to varying detection rates between endoscopists and different baseline population risk. Under the old terminology, the prevalence of clinically relevant serrated polyps in US and European studies varied from 1.8% for large hyperplastic polyps to 3.3% for serrated adenomas to 8.5% for proximal serrated polyps.103

In Asia, only a few reports on SSLs have been published. Excluding one extreme statistical outlier, the mean prevalence of SSLs is much lower in Asia than in the United States—namely, less than 1%.106 On the other hand, a study from Japan recruiting 5218 asymptomatic subjects for CRC screening reported detection rates of serrated lesions of 23.3% and of right-sided serrated lesions of 7.6%, respectively.107 In that study, high-quality video endoscopes with narrow-band imaging and magnification were used with 0.4% indigo carmine dye to enhance the detection of flat lesions. On the other hand, in another CRC screening study from Hong Kong recruiting 6011 subjects, 486 (8.1%) subjects were reported to have HPs and only 85 (1.4%) SSLs.108 A retrospective study from China reported SSLs in only 0.7% among symptomatic and asymptomatic individuals who underwent colonoscopy.109 In Australia, the prevalence of SSLs in Chinese (2%) was lower than in Caucasian (7%) subjects.110 Taken together, SSLs are likely to be underdiagnosed in Asia and should be carefully looked for during colonoscopy, especially in the right side of the colon. Training for endoscopists and pathologists to identify SSLs will probably increase detection rates and improve the prevalence of estimates of these lesions.

Statement 18. Detection of sessile serrated lesions may be enhanced by improving awareness of the importance of these lesions and training in their detection.

Level of agreement: A=87.5%, B=12.5%, C=0%, D=0%, E=0%

Quality of evidence: I

Classification of recommendation: A

The high variability between studies for the prevalence of SSLs, is at least partly explained by varying detection rates of serrated lesions between endoscopists, as this rate appears highly operator dependent. In a study involving 15 experienced gastroenterologists at two academic endoscopy units between 2000 and 2009, during which a total of 11 049 polyps were detected in 6681 colonoscopies, the detection rate of serrated polyps in the proximal colon varied from 1% to 18%.111 Two recently published meta-analyses also reported the detection rates of SSLs. Results from 16 studies reporting SSL detection rates and 12 studies reporting proximal serrated polyp detection rates were pooled.112 Among 280 370 average-risk individuals undergoing screening colonoscopy, SSLs were detected in 2.5% and proximal serrated polyps in 10%. Heterogeneity and publication bias was noted in both this review and meta-analysis. In a similar meta-analysis by another group, including 17 studies and 129 000 colonoscopies, SSLs were detected in 2.5%.113 Detection of SSLs was higher among Caucasians (2.9%) than among Asians (0.7%). Subgroup analysis showed that SSLs, with an overall detection rate of 19%, were more common in men (22%) than in women (14%), and more common in Caucasians (25.9%) than in Asians (14.6%). Increasing both endoscopist and pathologist awareness, as well as interdisciplinary communication is one way to increase detection of SSLs.114 115

SSLs are sessile or flat lesions with an average size of 5–7 mm. They are usually larger than HPs. They may exhibit distinct endoscopic features, such as overlying mucus cap, cloud-like surface, ring of debris or stool around the lesion and obscured mucosal vasculature.116 117 During narrow-band imaging (NBI) endoscopy, they have a cloud-like appearance, irregular shape and dark spots inside the crypts. Better bowel preparation, longer withdrawal time and careful examination of the right colon (with repeated anterograde examination or retroflexion in the caecum) improved detection of SSLs.104 Chromoendoscopy such as NBI may marginally improve the detection of SSLs but is currently not recommended as mandatory practice as clear scientific evidence is lacking.

Dutch Workgroup serrAted polypS and Polyposis (WASP) developed a classification system for optical diagnosis of adenoma, HP and SSA/P118 by combining the NBI International Colorectal Endoscopic classification (1. brown colour, 2. brown vessel, 3. oval tubular or branched surface patter) with the WASP criteria for differentiation of SSLs (1. clouded surface, 2. indistinct borders, 3. irregular shape, 4. dark spots inside the crypts) in a stepwise approach. After training, the accuracy of optical diagnosis of SSL improved.119 Just a brief and simple educational classroom intervention to trained and untrained endoscopists significantly improved the detection of serrated polyps and the improvement was long term.119 The working group recommends setting up proper training opportunities for endoscopists to optimise SSL detection rates.

Statement 19. After complete removal of one SSL >10 mm or SSL with cytological dysplasia, colonoscopy should be recommended at a 3-year interval.

Level of agreement: A=54.1%, B=37.5%, C=4.2%, D=4.2%, E=0%

Quality of evidence: III

Classification of recommendation: B

PCCRCs are defined as cancers diagnosed after a colonoscopy in which no cancer was found before the next due surveillance procedure. They might arise from a specific biologic type of CRC, missed cancer as well as missed or incompletely resected benign lesions.120 Studies have shown that PCCRCs occur in 1–9% of cases and are more commonly found in the right colon.121 122 Molecular profiling studies have found that PCCRCs more often display the molecular features CIMP (OR 2.19) and MSI (OR 2.1), indicating that they often followed the serrated pathway and more often arise from serrated polyps.19 123

There is convincing evidence that patients with serrated polyps have an increased risk of CRC. A study using the Danish database including 272 342 individuals who underwent colonoscopies showed that having SSAs/Ps at baseline colonoscopy increased the risk of metachronous CRC about threefold.124 In the New Hamsphire Colonoscopy Registry study, 65 patients with large serrated polyps (HP, SSP or TSA) underwent two colonoscopies with a median time surveillance of 4.9 years, of whom 3.1% had a high-risk adenoma.125 SSA/Ps with cytological dysplasia have an even higher risk of CRC of 4.76-fold. SSLs with a nodule/protusion and those larger than 10 mm in size are more likely to have cytological dysplasia.126 However, all SSLs should be removed and all those individuals carefully surveilled after removal. Size and multiplicity are the two most important risk factors for metachronous malignancy.127

During colonoscopy, SSLs can be removed (piecemeal) by cold snare polypectomy, (piecemeal) by endoscopic mucosal resection and under water endoscopic mucosal resection. A study in four Australian centres showed that, irrespective of the method of removal, around 4% of SLLs recur within 18 months.128 A similar recurrence rate of 5% within 18 months was reported in Japan.129 Another study reported a 3.6% recurrence within 18 months.130 A meta-analysis pooling nine studies with 34 084 participants confirmed that individuals with serrated polyps have a twofold risk of having advanced adenoma.131 The presence of proximal serrated polyps and large serrated polyps further increases the risk of advanced adenomas by 2.7-fold and 4.1-fold, respectively.

Based on this evidence, the US Multi-Society Task Force on Colorectal Cancer issued a series of recommendations for follow-up after colonoscopy and polypectomy of SSLs.47 For patients with low-risk SSLs (one or two lesions, <10 mm in size and completely removed), repeat colonoscopy in 5–10 years is recommended. For patients with medium-risk SSLs (three or four SSLs,<10 mm in size) repeat colonoscopy is recommended in 3–5 years. For those with 5–10 SSLs, <10 mm in size or any SSL >10 mm, the US Multi-Society Task Force recommends repeat colonoscopy in 3 years. As the task force mentioned in their guidelines, these are weak recommendations based on very low-quality of evidence (mostly cohort studies, with relatively small sample size and limited follow-up). On the other hand, the ESGE guideline update in 2020 suggests that any serrated polyps <10 mm without dysplasia do not require endoscopic surveillance and those individuals should return to routine screening.67 An individual with any SSL >10 mm or SSL with dysplasia should receive surveillance colonoscopy after 3 years. The European guideline graded the evidence as moderate quality and classified the recommendation as strong. There is obvious discrepancy between the two major international guidelines. The Asia-Pacific Working Group feels that it is important to emphasise the importance of high-quality examination, excellent bowel preparation and complete removal of SSLs. Under these circumstances, after complete removal of any SSL >10 mm or with cytological dysplasia, colonoscopy should be recommended at a 3-year interval.

Quality assurance of CRC screening programme

Statement 20. The quality of a FIT programme depends on adherence to initial testing and follow-up of the testing results.

Level of agreement: A=79.2%, B=20.8%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

CRC screening is not a single diagnostic test but a programme with steps requiring adherence and follow-up of the testing results. Participation in FIT screening and compliance with diagnostic examinations after a positive FIT is crucial to secure the effectiveness of a FIT screening programme, and these parameters should be regularly audited.132 133 Any barrier to FIT screening and subsequent diagnostic examinations should be removed. Education of the public and reminders of serial FITs should be emphasised.

Currently, participation rates of the screening population in the Asia-Pacific region are 43.5% in Australia (2019), 47.8% (men) and 40.9% (women) in Japan (2019), 58.4% (ever-screening rate) in Korea and 38% in Taiwan (2019).16 134–136 The rate of diagnostic colonoscopy after a positive FIT is approximately 70% in Japan and Taiwan, 62% in Australia, and 61.3% in Korea.16 134 137 Studies from Taiwan showed that CRC death was 64% higher in FIT-positive individuals who were not compliant with the diagnostic colonoscopy.138 Both FIT screening participation and colonoscopy rate in the Asia-Pacific region were much lower than those in the FIT-based Dutch and UK programmes, which have a higher than 60% FIT participation and around 85% colonoscopy adherence rates.139 In the Asia-Pacific region, CRC screening programmes very much rely on a FIT. Adherence to a programme of regular FITs and follow-up actions when testing positive should be strengthened.

Statement 21. Timely colonoscopy (<3 months) after a positive FIT in a CRC screening programme should be offered and monitored.

Level of agreement: A=75%, B=25%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: B

Subjects with a positive non-invasive screening test are at higher risk of having advanced adenoma and CRC. Taking FIT as an example, subjects with positive FIT are at a four- to fivefold risk of having an advanced adenoma and 20- to 30-fold risk of CRC compared with the general population of screening age, and also have a higher likelihood of harbouring large and/or synchronous neoplasias.140 The risk of CRC and advanced adenoma is dependent on the test-positivity threshold.141 A study from China using the Asia-Pacific CRC Screening Score combined with a FIT showed improvement in positivity of colonoscopy.142 Delay in diagnostic colonoscopy after a positive FIT might lead to higher CRC incidence, more advanced stages of CRC, and more CRC deaths.143 The study by Lee et al demonstrated that the risk for CRC was 31% higher when colonoscopy was delayed by more than 6 months, and 109% higher for advanced-stage disease. Those risks continuously increased when colonoscopy was delayed by more than 12 months: for any CRC a 117% increase, and for advanced-stage CRC 184%.144 Similar results were also reported from other programmes.143 145 146 It is worthwhile mentioning that non-compliance or delay in colonoscopy became remarkable after the outbreak of the COVID-19 pandemic, and these screening and colonoscopy backlogs should aggressively be dealt with to minimise the negative impact on the effectiveness of entire screening programmes and the psychological distress of the FIT-positive subjects.147–149 In view of this, the working group recommends that timely colonoscopy after a positive non-invasive screening test in a CRC screening programme should be offered, preferably within 3 months after a positive FIT. The follow-up of a FIT should be monitored and used as a performance indicator of the CRC screening programme.

Statement 22. Barriers to screening test uptake and compliance with colonoscopy in CRC screening should be investigated and those present removed.

Level of agreement: A=91.7%, B=8.3%, C=0%, D=0%, E=0%

Quality of evidence: III

Classification of recommendation: B

It is important to identify facilitators and remove barriers to participation in CRC screening. A meta-analysis demonstrated that awareness of CRC affects views, attitudes and motivation for CRC screening.150 Factors enhancing awareness included public education to address misconceptions, enlisting the efforts of primary care physicians to recommend screening, and enhancing the influence of friends and family of patients with CRC. Significant barriers to participation included barriers to access, logistical challenges to attending screening and cultural beliefs.150 Previous studies in the Asia-Pacific countries/region showed that insufficient public awareness was the main cause of the low rates of screening participation. Recommendation by a physician is the strongest factor motivating people to go for screening in the opportunistic setting.151 152 As for the compliance to follow-up procedures after a positive FIT, lack of comprehension about the test was the strongest predictor of non-adherence, whereas higher socioeconomic status, higher perceived threat, higher cues for action, lower perceived barriers, and higher health behaviour scores were identified as factors associated with a greater willingness to undergo diagnostic colonoscopy after a positive FIT.153 154 Individual programmes should increase efforts to remove or overcome those barriers to follow-up colonoscopy, thereby maximising the effectiveness of the screening.

Statement 23. Quality control of a FIT and colonoscopy is mandatory for CRC screening programmes, and benchmarks should be determined.

Level of agreement: A=91.7%, B=8.3%, C=0%, D=0%, E=0%

Quality of evidence: II-2

Classification of recommendation: A

Assurance of the quality of a FIT and colonoscopy is crucial in securing effectiveness of the entire screening programme. A false sense of security can be generated by a screening programme not properly run.

The performance of a FIT may be affected by the design of the collection tube, the buffer that is used, the ambient temperature, the duration from stool collection to assaying in the laboratory and different FIT brands.155 In a study from Taiwan, two brands of FITs used in the screening programme were compared and it was demonstrated that one FIT outperformed the other, with better detection rates for CRC and lower rates of interval CRC after a negative FIT, even using the same cut-off point.156 In the Asia-Pacific, some programmes use more than one brand of FIT kits. In Korea, it was demonstrated that positivity rate of a FIT was lower in the summer, with subsequently higher risk of interval CRC in this season.157 Regular monitoring of the performance of a FIT at a regional and national level using short-term indicators (positivity rate, detection rate for CRC) and long-term indicators (interval CRC after a negative FIT) is important.

Inadequate quality of colonoscopy is related to PCCRC. It is well-known that the quality of colonoscopy varies between individual programmes, institutes and endoscopists in the rate of complete colonoscopy, detection rates of neoplastic lesions and polypectomy techniques.121 158 The ADR, one of the most important quality metrics of colonoscopy, is associated with the risk of PCCRC, advanced- stage CRC and CRC death.80–82 High ADR has been proved to be associated with a reduced risk of interval CRC and CRC death.159 A multinational study in the Asia-Pacific region has revealed that overall ADR, male-specific ADR, female-specific ADR and overall advanced ADR were 1.93-, 2-, 1.94- and 8.2-fold higher, respectively in the FIT-positive colonoscopy cohort in comparison with the direct screening colonoscopy cohort.160 Therefore, ADR benchmarks should be higher among FIT-positive individuals. Monitoring and auditing, together with continued training programmes and accreditation systems, will be indispensable to sustain high-quality colonoscopy screening for CRC.161 The Asia-Pacific Working Group feels strongly that quality control of colonoscopy and FIT is mandatory for CRC screening programmes, and benchmarks should be determined.