Is this the end of colonoscopy screening for colorectal cancer? An Asia-Pacific perspective.

Abstract

The recent publication from the Nordic-European Initiative on Colorectal Cancer (NordICC) Study Group1 has generated considerable debate in the gastroenterology field on the role of colonoscopy for colorectal cancer (CRC) screening. This begs the question of whether colonoscopy screening for CRC is futile and a waste of time and resources, or is it the method of invitation the major reason for failure? Below, we will summarize the study findings and discuss the implications and the latest developments for CRC screening in the Asia-Pacific (APAC) region. In summary, the study investigators set out to explore the efficacy of a colonoscopy screening program in several European countries where participants were randomly assigned in a 1:2 ratio to receive either an invitation (by mailing a letter) to undergo a single screening colonoscopy or to receive no invitation or screening. The pre-specified primary endpoints were the risks of CRC and CRC-related death, and the secondary endpoint was death from any cause. From 2009 to 2014, over 80 000 subjects were recruited. At 10 years, CRC risk was 0.98% in the invited group and 1.20% in the usual-care group, which corresponds to a significant 18% risk reduction in the invited group. However, there was no significant difference in the CRC deaths between the groups (0.28% in the invited group vs 0.31% in the usual-care group). The number needed to invite for screening to prevent one case of CRC was 455. Looking deeper into the numbers, only 42% of those invited to have a colonoscopy had the procedure performed. Subgroup analysis of those who had the colonoscopy done showed that the risk of developing and dying from CRC decreased by approximately 30% and 50%, respectively. We should be looking beyond the differences in locoregional practice, the somewhat heated arguments on news, editorials and social media, and the unprecedented issuance of statements regarding this study by gastroenterological and oncological societies. Colonoscopy saves lives but whether a colonoscopy screening program implemented at the population level does the same is an entirely different question. The results of this large, well-run, randomized controlled trial suggest that the answer would be no for a colonoscopy-based CRC screening program where subjects were invited by post, at least for the European countries that were involved. Even in the United States, there is no nationwide bowel screening program covering all citizens, though screening tests recommended by the United States Preventive Services Task Force (USPSTF)2 are required by law to be covered by both private insurers and Medicare with no out-of-pocket costs for patients. However, despite efforts to enhance health insurance coverage, latest data from 2021 by the US Census Bureau showed that there are still 27 million citizens (8.3% of the total population) that are uninsured.3 We have several comments regarding the study. First, the screening participation was much higher in Norway than the other European countries. Is this due to better CRC awareness and thus better acceptance for screening colonoscopies? Or could it be differences in the postal system or adherence to follow-up actions for those who did not reply initially? The importance of participation for any organized screening efforts cannot be overstated. Previously a Spanish trial already showed that fecal immunochemical test (FIT)-based screening has higher participation rates when compared with colonoscopy (34% vs 25%).4 Interestingly, the resultant detection rate for CRC was similar in both arms in the intention-to-screen analysis, though the colonoscopy arm had a higher detection rate for advanced adenoma in the first round. We can only speculate whether this is still the case in subsequent screening rounds because screening uptake may decline for colonoscopy. Outside of clinical trials, real-world participation rates for screening would likely fare even worse. Second, the quality of colonoscopy may have been suboptimal in some sites. Although the overall adenoma detection rate (ADR) was around 31%, which is in line with the current recommended minimal threshold of 25% (30% in men and 20% in women),5 breaking down the data by country showed that the ADR ranges from 14.4% in Sweden to 35.2% in Poland. A major critique would be the dilution of the beneficial effects of colonoscopy if these were not up to standard. We also know that ADRs have been steadily increasing throughout the years, with a recent US registry-based study reporting an average ADR of 39%.6 The factors leading to these improvements are not fully known, but increased emphasis on quality colonoscopy through education, awareness, recognition of the importance of detecting and removing adenomatous polyps, benchmarking and utilizing ADR as a quality indicator, improvements in bowel preparation, optimization of withdrawal times and techniques, and technological advances in equipment may all play a part. These recent advances would not have been accounted for in this study. Also, apart from bowel preparation, caecal intubation rate and ADR, other quality indicators for colonoscopy were not described. Third, whether contamination of the usual-care group, that is, uninvited subjects may have had the opportunity to undergo colonoscopy or other CRC screening tests, was not addressed nor explored. Fourth, the detection of early-stage (Dukes' A or B) and late-stage (Dukes' stage C or D) was 0.38% and 0.40% in the invited group, respectively. There seem to be higher numbers of late-stage cancers than would be expected in a screening trial. Is it possible that some of the participants had symptoms but were nonetheless still recruited into the study? Finally, follow-up data from the Netherlands was not available due to a newly introduced Dutch law based on the European Union General Data Protection Regulation. Although the Dutch participants would have only accounted for approximately 10% of the entire study population, there is a possibility that this omission may have altered the final outcome. Is this the end of colonoscopy screening for CRC? We eagerly await the results of the CONFIRM trial, which compares screening colonoscopy vs annual FIT in 50 000 average-risk individuals recruited at Veteran Affairs medical centers across the United States,7 and the Japanese Akita pop-colon trial, which compares annual FIT alone vs annual FIT combined with one-time colonoscopy for the final verdict on colonoscopy for CRC screening.8 Pivoting back to the APAC region, despite being trend-setters, that is, implementing the world's first national CRC screening program based on FIT in Japan since 1992,9 and arguably the world's first screening program with distribution of FIT kits by postal mail in Australia,10 organized screening has not been universally adopted across the entire region (Fig. 1). In contrast to the United States where screening is largely reimbursed by commercial insurance with a preference for colonoscopy, in APAC, screening is largely funded by the government with the primary screening tool being FIT (Table 1). The seminal work by Wilson and Jungner11 highlighted the principles of whether population-based screening for a disease is worthwhile. Colorectal cancer, as with all cancer diagnoses, can lead to substantial morbidity or mortality, but whether it is an important enough health problem to screen will require the consideration of the societal impact and epidemiological trends. The large burden of colorectal cancer in East Asia and Oceania, which is largely on par with Europe and North America, supports the implementation of organized CRC screening,12 but this may not be a straightforward decision for other countries and regions in Southeast Asia and South Asia with varying CRC incidence rates. The natural history of CRC is well understood, and there is a relatively long and recognizable asymptomatic stage. If lesions are detected, there are effective treatments for precancerous lesions and CRC, ranging from endoscopic resection to surgery. There are also cost implications, as the cost of case-finding that includes diagnosis and treatment should be economically balanced in relation to the possible expenditures on medical care as a whole. Again, it would be difficult to implement a one-size-fits-all strategy for the entire APAC region due to the significant differences in medical access and heterogenous socioeconomic development. Studies have also shown that screening tailored to the age standardized rate (ASR) of CRC seems to be a cost-effective approach. A Markov modeling analysis from Singapore for a population cohort aged 50–70 years and using the threshold of US$50 000 per quality-adjusted life-years showed that in a population with an ASR > 21/100 000 (e.g., United States and Singapore) any screening modality would be cost-effective; if the population ASR of CRC decreases to 14–21/100 000, only FIT-based screening is cost-effective, as the incremental cost effectiveness ratio of a colonoscopy every 10 years exceeds the threshold; if the population ASR of CRC is further decreased to <14/100 000, then even FIT-based screening will not be cost-effective.13 Thus, whether CRC screening is worthwhile would really depend on the APAC region or country in question, given the vast differences in the overall risk of developing CRC, income status, and socioeconomic development. No nationwide screening program Cancer screening program in place for some rural (234 counties) and urban areas (42 cities) for high-risk individuals selected by risk assessment questionnaires Even for regions and countries where organized CRC screening programs have been implemented, the Wilson and Jungner criteria highlight the importance of adopting screening strategies that are easy to administer, effective, and acceptable to society (ideally noninvasive), clearly define whom to treat as patients, and be implemented in a way where the screening process is streamlined. Against this backdrop, the NordICC Study provides further evidence to support the statements of the recently published Third Asia-Pacific Consensus Recommendations on CRC Screening and Postpolypectomy Surveillance.14 Cognizant of the huge population in our region, the aging population with more individuals reaching screening age, enormous implications on the utilization of healthcare resources, shortage of endoscopists, low public awareness for colorectal cancer, and suboptimal uptake rates for existing screening programs, it was clearly stated that offering direct colonoscopy screening for the entire eligible population would not be a feasible nor pragmatic approach in many parts of APAC. Colonoscopy-based screening would also result in more surveillance colonoscopies due to the detection of many small and diminutive polyps. On the other hand, FIT-based screening has become increasingly attractive as it has been shown to have higher sensitivity for CRC, better detection of advanced adenomas, and greater screening participation rates when compared with the now obsolete guaiac-based fecal occult blood test.15 There may also be a sociocultural context. A prior US study demonstrated that Asians (primarily Chinese) preferred a noninvasive stool test when given the choice.16 Moreover, repeated testing according to screening intervals provides the opportunity for FIT to catch up with colonoscopy in terms of program sensitivity, which is defined as the summation of neoplastic lesions being detected in all screening rounds. Indeed, a recent Italian study showed that participating in a FIT-based screening program reduced not only CRC mortality but also incidence.17 Though a retrospective study from Korea found that colonoscopy was more effective in reducing mortality for patients with CRC compared with FIT,18 a meta-analysis also showed that risk-tailored CRC screening was cost-effective when compared with no screening, but was inconclusive when compared with age-based screening given the many discrepancies in the included studies.19 Despite this, risk-stratified approaches utilizing FIT, validated risk-stratification systems (using clinical parameters such as age, male gender, family history of colorectal cancer, smoking, and body mass index)20 genetic risk scores,21, 22 or several of these strategies together23, 24 will likely be the pragmatic way forward in this part of the world. Even by adopting a risk-stratified approach where high-risk subjects have been preselected that should theoretically lead to a much higher compliance to colonoscopy, pooled results from observational studies showed that only 77% of subjects had their endoscopy performed in the end.25 In the future, existing screening strategies can become even more effective by improving available tests or by incorporating new screening tools. Improving direct visualization tests, that is, colonoscopy by using various evidence-based endoscopic techniques (withdrawal times, second forward view examination, feedback to endoscopists on quality indicators for colonoscopy), utilization of a variety of distal attachment devices, routine use of image-enhanced endoscopy such as narrowband imaging (NBI), blue light imaging (BLI), and linked color imaging (LCI) with or without magnification, and incorporating artificial intelligence (AI) technology. In a systematic review and meta-analysis of 14 studies with >10 000 patients, computer-aided detection (CADe) led to a 65% reduction in the adenoma miss rate and a 78% reduction in the sessile serrated lesion miss rate.26 AI tools possibly increase the ADR by roughly 50% and may contribute to a 7–20% reduction of colonoscopy-related costs.27 The software, hardware, and interface platforms are already available. A study showed that a commercially available desktop graphics processing unit can already run some of these algorithms, making it attractive for low-middle-income countries.28 This may allow for technology leapfrogging if colonoscopy-based AI-enhanced screening programs become more mainstream. New and emerging screening tools may also have a role in future CRC screening strategies.29 These can be largely divided into stool, blood, image-based screening tests, colon capsule, metabolomics, and breath tests (Fig. 2). For stools, multitarget stool DNA testing is FDA approved, but its widespread adoption has been hampered by high costs. In some countries, fecal pyruvate kinase isoenzyme type M2 (M2-PK test) that assesses perturbations in tumor metabolism is also available.30 Stool microbial DNA markers have also been extensively studied. Various studies fromHong Kong have shown that microbial markers from the bacteria Fusobacterium nucleatum,31 Lachnoclostridium sp. m3, Bacteroides clarus, and Clostridium hathewayi combined with FIT have been shown to achieve better detection rates of CRC and advanced adenomas than FIT alone.32 Liquid biopsies targeting cell-free DNA (cfDNA) such as methylated septin 9 (mSEPT9) +/− protein markers in the blood are currently being studied as a new screening strategy. Moving beyond a one test–one cancer toward a one test–many cancers paradigm, the final results of the prospective, interventional PATHFINDER study conducted in the United States will provide evidence on the performance of a cfDNA-based multi-cancer early detection test.33 Another approach that is being studied is a multi-omics blood-based platform for the early detection of CRC with the prospective, multicenter observational PREEMPT CRC study having completed enrolment of more than 35 000 participants in 2022.34 Currently available image-based screening tests such as computed tomography (CT) colonography (virtual colonoscopy) are not recommended as primary screening tools in the latest APAC Consensus. Magnetic resonance colonography is a newer technique that will likely require further validation in larger-scale studies. Another novel approach is the X-ray imaging capsule. The device emits low dose X-rays and is sensed by detectors in the capsule to generate high-resolution three-dimensional (3D) imagery of the colon without bowel cleansing.35 Colon capsule endoscopy is relatively noninvasive but still requires bowel preparation. Future developments may be the incorporation of AI for the detection and diagnosis of lesions. Detecting metabolomic signatures of CRC in serum36 and/or urine37 also seems promising, but larger studies are needed. Last but not least, CRC seems to have a distinct odor that can be identified by canine scent detection38 and more recently by a noninvasive breath test based on detecting endogenous volatile organic compounds.39 These emerging modalities will likely be more sensitive than existing noninvasive tests, less invasive than endoscopy, and more palatable for the general population. Just by enhancing screening participation with a comparably accurate test may already be sufficient in upending the perceived absolute advantage of direct colonoscopy screening. Although these new screening tests are currently not recommended as a primary screening tool in the latest APAC Consensus, this will be an exciting space to watch due to rapid advances in recent years. No screening test is perfect, but in the near future, utilizing multi-omics or multiple screening tests in conjunction may be the way forward to further enhance the performance of noninvasive CRC screening strategies. With the rise in early-onset CRC (EO-CRC),40 the age to initiate CRC screening has been vigorously debated. In the United States, microsimulation modeling analyses suggest that screening for CRC starting at age 45 years provides an efficient balance between colonoscopy burden and life-years gained.41 However, one caveat of these simulations was the assumption of perfect adherence to screening strategies, which is never seen in real life. In addition, modeling studies need to be based on assumptions of the specific region/country in question, tailoring socioeconomic factors such as costs of a screening test, CRC treatment costs, etc. Indeed, an Australian microsimulation modeling analysis concluded that initiating screening at age 45 years could be cost-effective but would increase colonoscopy demand and jeopardize the benefit of screening for those aged 50–74 years.42 This is a pressing issue in the APAC as the huge Baby Boomer Generation have now reached screening age, and Generation X is also gradually becoming eligible for screening. As such, lowering the age of screening will lead to the unavoidable expansion of the entire screening population and will put an immeasurable amount of stress on the already overstretched healthcare services and endoscopy manpower. With the data currently available, we believe that lowering the age to start CRC screening to <50 years is not cost-effective in Asia and in line with the latest APAC Consensus recommendations. One possible strategy to address the concern of EO-CRC would be targeted screening of high-risk subgroups such as men who smoke or have metabolic syndrome that may justify earlier screening.43 Recently, a personal risk score based on genetic and environmental risk factors has been developed, which may discriminate between true average-risk and high-risk individuals.44 In conclusion, regions and countries in the APAC region that already have systematic CRC screening programs in place should focus on getting the basics right. Quality assurance of such programs is of paramount importance. Physicians and policymakers need to work together to remove barriers for screening, promote awareness for CRC, encourage eligible individuals to participate in the available screening programs, emphasize the importance of having follow-up tests, and ensure timely colonoscopy is performed (i.e., <3 months from the positive FIT as suggested in the latest APAC Consensus14). These performance standards should also be regularly audited. For regions or countries where systematic screening is not yet available, there is no easy answer on when to start CRC screening. As many parts of the APAC become increasingly affluent, more urban, with a more Westernized diet and a more sedentary lifestyle, the age-standardized incidence rate of CRC will likely increase. Thresholds should be set and closely monitored, and when reached should trigger in-depth discussions and real action, and be tied in with policy implementation and healthcare resources. A caveat would be the relative difficulty in some regions and countries to accurately identify all cases of CRC. Ideally, these policies should be clearly and formally documented in regional or national cancer control plans and tailored to each region and country's needs. Each individual society will need to decide on whether implementing such screening programs would be in the best interests of their citizens. Preliminary groundwork such as engaging various stakeholders and building political will is essential if a successful screening program is to be implemented. How emerging biomarkers and technologies can be incorporated into existing screening algorithms awaits to be seen, but cost-effectiveness, availability, and external validation are significant obstacles for widespread adoption. At the end of the day, the contemporary adage of “the best screening test is the one that gets done” still rings true.