Cost-effectiveness of colonoscopy and related procedures: population screening perspectives

INTRODUCTION

Colorectal cancer (CRC) is the second most common cancer in women and the third most common cancer in men worldwide^[1]. Early detection of CRC by effective screening strategies and removal of screening-detected adenomatous polyps is the most effective way to reduce the mortality associated with an incidence of CRC^[2]. Colonoscopy is the essential process and common pathway of CRC screening and plays a pivotal role in a screening program. Furthermore, patients who have adenomas detected during a colonoscopy are at increased future risk of CRC even after removal of the lesions; therefore, periodic surveillance is indispensable for providing further protection against metachronous CRC^[3,4]. The major guidelines currently recommend surveillance intervals based primarily on colonoscopic findings and place special emphasis on the number, size, and histology of detected neoplasms^[5-7]. Along with the increasing CRC epidemiology and the launch of population CRC screening in many developed countries, colonoscopy has nowadays become one of the most frequently performed clinical procedures^[8]; colonoscopies have become a dominant workload, and the costs of colonoscopies and their related procedures have become a significant financial burden for healthcare systems^[9]. The majority of colonic lesions detected by screening, however, are asymptomatic precancerous lesions and early CRCs; thus, subsequent treatment costs could be saved compared with that in the symptomatic clinical stage. Hence, there is a trade-off between the costs of colonoscopies and their related procedures and the treatment of colonic lesions including CRCs. Evolving endoscopic technologies such as endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) provide less invasive curative treatment options. The appropriate position of ESD in clinical practice is worthy of consideration. In general, ESD is more costly than the traditional polypectomy or EMR, but the expense is much lower than surgery and related expenses (admission fee, anesthesia, etc.). If compared with conventional polypectomy or EMR, it can reduce local recurrence or even post-colonoscopy colorectal cancer (PCCRC); and, if compared with surgery, it can reduce post-treatment morbidity, shorten the duration of admission, and improve the quality of life of the patients. These therapeutic options should be precisely tailored to appropriate lesions that would benefit from individual treatment, thereby balancing effectiveness (PCCRC and CRC mortality), cost, and willingness-to-pay (WTP) levels.

From the population screening perspective, screening logistics involve multiple steps, and each incurs various expenditures. Colonoscopy is the costliest part of the entire screening program, even when considering its related procedures. In this article, we systematically review the current evidence regarding economic appraisals pertaining to colonoscopic screening, therapeutic procedures for screening-detected neoplasms, and subsequent surveillance and provide a balanced view from the perspective of population screening programs.

METHODS

We conducted this systematic review based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines^[10] and Cochrane Handbook for systematic reviews^[11]. We conducted an electronic search on PubMed and Cochrane databases using the following terms: colonoscopy; colorectal cancer; cost-effectiveness; colorectal cancer screening; surveillance; polypectomy; resect-and-discard; endoscopic mucosal resection; endoscopic submucosal dissection. Our search was restricted to studies written in English and published between January 2012 and December 2021. The titles and abstracts were screened by W.F.H. and H.M.C. to exclude irrelevant studies. The full articles of potentially eligible studies were further reviewed by W.F.H. and H.M.C. independently.

DECISION ANALYSES AND ECONOMIC APPRAISALS OF DIFFERENT CRC SCREENING STRATEGIES: COLONSCOPY VS. VARIOUS NONINVASIVE TESTS

Many previous studies have investigated the economic impact of various CRC screening strategies^[12,13]. There are two main CRC screening methods: direct colonoscopy and two-tier noninvasive screening strategies using stool tests, such as guaiac fecal occult blood tests, fecal immunochemical test (FIT), or stool DNA tests, as the primary screening test followed by colonoscopy if positive test results are obtained. Colonoscopy has a higher ability to detect colonic neoplasms, especially early-stage CRC or precancerous neoplasms, than other noninvasive tests but is associated with higher risk and related costs^[14]. Moreover, the effectiveness of CRC screening may be offset by the lower participation rate^[14]. In addition to colonoscopy, there are several other noninvasive modalities for CRC screening, and these modalities decrease the demand for colonoscopic capacity, hence reducing the incidence of colonoscopy-related complications. Although those modalities may enhance screening participation rate, their ability to detect adenoma and early CRC is inferior to colonoscopy. Eleven studies analyzed the cost-effectiveness of different screening strategies [Table 1]^[15-26]. Compared with no screening strategy, screening with a 10-year colonoscopy was cost-saving in four studies^{[16,19,21,23]} and was cost-effective in the other seven studies^{[15,17,18,20,22,24,26]}.

The two-tier screening strategy with the FIT as the primary screening test is currently the most preferred way of providing CRC screening service in regions or countries where endoscopy capacity is constrained. It has nowadays become the mainstay of the stool-based screening approach, as it has the advantage of absolving the need for dietary restriction, has a more user-friendly platform with higher screening participation by the public, and provides higher test performance compared with that of guaiac FOBT^[27]. When 10-year colonoscopy was compared with FIT screening strategies (annual/biennial), two studies from the United States reported that colonoscopy screening had an incremental cost-effectiveness ratio (ICER) below $50,000/per quality-adjusted life-year (QALY) and was dominant with 100% certainty^[16,25]. Six other studies showed that colonoscopy every 10 years was more costly but with more life-years (LYs)/QALYs gained; hence, it was more cost-effective than the annual FIT screening strategy^{[15,18,19,23,24.26]}. Four studies reported that annual FIT screening was more effective and less costly than 10-year colonoscopy^{[17,20,21,24]}. The results mainly indicate that adherence to the noninvasive screening tests was presumed to be higher than that to the colonoscopy. In two studies, when 10-year colonoscopy was compared with biennial FIT, the former was less cost-effective, and the ICER was higher than $200,000/per LY/QALY^[17,24]. One study from Hong Kong reported that biennial FIT was less costly and more effective than colonoscopy every 10 years under the assumption of the same compliance rate of 60% for all screening strategies^[22]. Some studies also assessed mixed-method screening with combined FIT and colonoscopy or sigmoidoscopy. In the analysis conducted by Dinh and colleagues, colonoscopy every 10 years was more cost-effective than initial annual FIT/colonoscopy screening once and subsequent annual FIT/sigmoidoscopy screening and had ICERs of $35,000 and $51,000 per QALY gained, respectively^[19]. In Ladabaum’s study, hybrid strategies with biennial FIT and colonoscopy had greater effectiveness and lower costs than 10-year colonoscopy^[21].

Ladabaum et al.^[21] also compared 10-year colonoscopy with blood-based screening using the methylated septin 9 DNA (mSEPT9) test. Compared with colonoscopy every 10 years, annual ^mSEPT9-based strategies cost more and were less effective with fewer QALYs gained. Biennial ^mSEPT9-based strategies were cost-effective compared with colonoscopy every ten years, but the ICER was $167,570-975,788 per QALY gained. Multitarget stool DNA (MT-sDNA) every three years was approved by the US Food and Drug Administration (FDA) in 2014^[28]. Although the sensitivity of MT-sDNA for CRC was 92.3%, which was much higher than that for FIT, the specificity of MT-sDNA was lower than that of FIT, resulting in an increased demand for colonoscopy and reducing its cost-effectiveness^[29]. Another study by Ladabaum et al.^[24] showed that, compared with 10-year colonoscopy, 3-year MT-sDNA was more expensive with fewer QALYs gained. In Peterse’s study, compared with colonoscopy every 10 years, ^mSEPT9 (annual/biennial) and MT-sDNA (annual/3-year) cost more and were less effective with fewer QALYs gained^[26].

Virtual colonography with computed tomography colonography (CTC) or MR colonography (MRC) is another noninvasive approach that can potentially be an alternative examination for subjects for whom colonoscopy is not feasible or where colonoscopy capacity or manpower is constrained. Four studies showed that 10-year colonoscopy was more effective and less costly than CTC^[16,18,25]. One study showed that 5-year CTC was less cost-effective than 10-year colonoscopy and displayed an ICER of $372,727 per LY gained^[17], and another study displayed an ICER of $8500 per QALY gained^[26]. Notably, in these five studies, the same compliance rates for CTC and colonoscopy were assumed. Nevertheless, concerns have been raised about CTC because of the potential radiation risks, especially the risk of radiation-related cancer, and the compliance regarding CTC by the public in Asia is rarely reported^[30,31]. The compliance rate for CTC may affect screening uptake and overall effectiveness.

Different age groups are associated with different risks of CRC. Precision screening strategies based on different CRC risks are important issues for balancing between effectiveness and usage of relevant resources. Ladabaum et al.^[21] compared 10-year colonoscopy, a hybrid strategy with FIT years for ages 50-54 and then colonoscopy starting at age 55 (FIT/COLO 55,65), and a hybrid strategy with FIT years for ages 50-59 and then colonoscopy starting at age 60 (FIT/COLO 60,70). The FIT/COLO 55,65 hybrid strategy was the most effective strategy, followed by FIT and FIT/COLO 60,70 hybrid strategy. At a willingness-to-pay threshold of €50,000 per QALY gained, the FIT/COLO 55,65 hybrid strategy was the optimal strategy. These results reveal that a hybrid strategy with FIT and colonoscopy based on the different CRC risks was a cost-effective approach. Dan et al.^[18] reported that a hybrid strategy using FIT for lower-risk patients aged 50-60 years and colonoscopy from ages 60 to 72 years showed it is potentially the most cost-effective strategy. It saves almost as many lives as 10-year colonoscopy but reduces costs by 25% and the number of colonoscopies and screening-related deaths by 35%. These results show that the hybrid strategy with FIT and colonoscopy based on the different risks for CRC was cost-effective. However, there was no relevant cost-effectiveness analysis for the precision screening strategies by the combination of age and other risk factors, and further studies are needed.

The optimal strategies at different WTP levels are shown in Table 1. Two models from the US^[16,25] and one model from Canada^[15] reported that the optimal strategy was 10-year colonoscopy even at the threshold of $20,000. One model from Japan reported the same results, which may be due to the much lower fee per colonoscopy in Japan than in the US or other countries^[23]. Two modeling studies from the US and Singapore reported that the optimal strategy was 10-year colonoscopy at the threshold of $100,000^[18,26]. Three models from the US, Hong Kong, and France, however, reported that annual or biennial FIT screening strategies were optimal strategies at the threshold of $50,000^[17,22,24].

Therefore, there is no “most recommended” test over the others at this time because the optimal strategy may vary along with the healthcare scenario, medical resources and expenses in individual countries, the WTP threshold of the healthcare payer, and local epidemiology of colorectal neoplasms collectively affect the choice of screening strategies. According to Dan’s study, the cost of colonoscopy and the incidence of CRC were the essential parameters for determining the optimal strategy. When the cost of colonoscopy was less than $300, regardless of the risk of CRC, the strategy of 10-year colonoscopy was the most cost-effective^[18]. When the cost of colonoscopy was above $300, FIT was considered the dominant technique at lower incidence levels of CRC. At a higher WTP threshold, the strategy of 10-year colonoscopy was the most effective but required a huge budget and extensive logistic resources.

The quality of colonoscopy also largely impacts the effectiveness of the screening. The adenoma detection rate (ADR), the proportion of subjects with neoplastic lesions detected by colonoscopy among all subjects who undergo colonoscopy, is associated with the risk of incident CRC or CRC death and has been recommended as a key quality benchmark of colonoscopy^[32-34]. Hassan et al.^[35] reported that colonoscopies performed by endoscopists with low ADRs resulted in a 7% absolute reduction in the CRC incidence prevention rate compared to colonoscopies performed by those with an average ADR (70% vs. 77%). The difference in CRC incidence prevention rate increased to 21% when comparing endoscopists with an average ADR to those with a high ADR. The substantial reduction in the long-term colorectal cancer prevention rate also resulted in substantial losses in LY and economic resources^[35].

Cost-effectiveness studies have also compared different scenarios of repeat screening. Greuter et al.^[36] reported that virtual colonography as CTC and MRC with more than three screening rounds were cost-effective alternatives for colonoscopy screening. Five rounds of CTC screening were even more effective at lower costs, and the ICER of five rounds of MRC was €3498 per LY gained compared with three rounds of colonoscopy^[36]. In Aronsson’s study, repeated colonoscopic screening strategies were more cost-effective than FIT when lifelong effects and costs were considered^[37]. Therefore, although a single screening colonoscopy yielded the lowest cost per QALY, 10-year colonoscopy gained additional QALYs at a reasonable cost.

COST-EFFECTIVENESS ANALYSES OF DIFFERENT COLONOSCOPIC TREATMENT STRATEGIES FOR SMALL ADENOMA

A substantial cost of a CRC screening program is represented by endoscopic polypectomy. Furthermore, polypectomy costs are partially related to the cost of pathologic examination. The cost-effectiveness analysis in this regard mainly relates to the “resect-and-discard” strategy^[38]. The rationale for the resect-and-discard strategy is based on the low prevalence of advanced histology in small (≤ 1 cm) adenomas, and discarding the specimen without sending it for pathological inspection could substantially reduce the cost. In the colonoscopy-based screening program, Lieberman et al.^[39] reported that the proportions of patients in a screening cohort with advanced histology were only 1.7% in the 1-5 mm group and 6.6% in the 6-9 mm group. Gupta et al.^[40] also reported that only 0.7% and 0.9% of polyps < 6 mm in size in the right and left colon were advanced neoplasms, respectively.

Image-enhanced endoscopy, such as narrow-band imaging (NBI), outperforms conventional white light colonoscopy in discriminating adenomatous from non-adenomatous colonic polyps with high accuracy^[41]. In patients with small polyps (< 10 mm), the application of NBI and the resect-and-discard strategy might correctly guide post-polypectomy surveillance intervals without histopathology information. The American Society for Gastrointestinal Endoscopy (ASGE) developed the Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI) initiative that has set the thresholds for real-time endoscopic assessment of the histology of diminutive colorectal polyps^[41]. The threshold in the PIVI document regarding the application of the resect-and-discard strategy for colorectal polyps < 5 mm in size is that endoscopic technology should provide at least 90% agreement in the assignment of the appropriate post-polypectomy surveillance interval. In this context, according to Hassan’s study, the application of the resect-and-discard strategy for diminutive polyps resulted in a savings of $25/person without reducing the screening efficacy. This approach would result in undiscounted annual savings of $33 million for the US population^[38]. However, due to the thresholds in the PIVI document being at least 90% in agreement in the assignment of the appropriate post-polypectomy surveillance interval, whether each endoscopy achieves the thresholds with high confidence is still unclear. As Patel et al.^[42] reported, previous studies revealed a discrepancy in agreement in surveillance intervals between endoscopists at academic centers and community-based endoscopists. This discrepancy may hinder the applicability of the resect-and-discard strategy and make its cost-effectiveness uncertain. Whether the aforementioned resect-and-discard strategy is similarly applicable to FIT-positive colonoscopy is not clear because the distributions of size and advanced histology of detected adenomas in the FIT-positive population are largely different from the screening colonoscopy cohort (general population of screening age). Hsu et al.^[43] analyzed 3920 neoplastic lesions in a FIT-positive colonoscopy cohort and compared them with 9789 neoplastic lesions in a direct screening colonoscopy cohort. Subjects who underwent colonoscopy for positive FIT had a lower prevalence of diminutive polyps (32.2%) than that in the screening colonoscopy cohort (60.5%); therefore, the resect-and-discard strategy in FIT-colonoscopy may not result in a similar degree of cost reduction that observed in the colonoscopy-based screening strategy. Although FIT positivity is associated with an increased likelihood of detecting neoplasms with advanced histology, those diminutive adenomas being detected in FIT-positive subjects were associated with a low risk of advanced histology (4.3%), and none of these patients had invasive cancer; hence, the risk of exposing the patient to undertreatment of invasive cancer by the resect-and-discard strategy is nearly zero. Moreover, compared with histopathological assessment, the agreement in determining surveillance interval by applying the resect-and-discard strategy was as high as 96.5%. Collectively, the resect-and-discard strategy for diminutive polyps in FIT-colonoscopy is feasible and could still save substantial costs making it a cost-effective approach.

COST-EFFECTIVENESS ANALYSIS OF COLONOSCOPIC TREATMENT STRATEGIES FOR LARGE NEOPLASMS

The cost-effectiveness analyses on colonoscopic treatment for large colorectal neoplasms mainly pertain to large flat lesions - laterally spreading tumors (LSTs). The LST was originally proposed by Kudo for tumors that spread extensively and circumferentially along the colonic wall and are larger than 10 mm in diameter^[44,45]. EMR is a useful therapeutic technique for LST; however, the maximum diameter of the lesion for en bloc resection is approximately 20 mm, owing to the size of the snare, the technical difficulty, and the perforation risk^[46]. For LSTs larger than 20 mm, the piecemeal EMR (pEMR) technique is commonly used, but it may be associated with a higher risk of recurrence^[47-49]. In Yamada’s study, LSTs with the presence of a large nodule, a depression, and an invasive pit pattern under magnified chromoendoscopy had a higher risk of deep submucosal invasion^[50]. Due to the substantial risk of submucosal invasion and multifocal invasion, en bloc resection with ESD techniques could be reserved for some specific lesions as an alternative procedure to more invasive surgery^[51]. Universal treatment with ESD for all LSTs larger than 20 mm may not be cost-effective because of its technical difficulty with a long procedure time, higher cost for the devices, and higher risk of perforation complications. In Bahin’s study, the cost-effectiveness of three strategies for removing large LSTs > 20 mm was compared: (1) wide-field endoscopic mucosal resection (WF-EMR); (2) universal endoscopic submucosal dissection (U-ESD); and (3) selective ESD (S-ESD)^[52]. In the WF-EMR strategy, lesions classified as highly suspicious for submucosal invasive cancer (SMIC) were referred for surgical management, suspicious rectal lesions classified as low-risk SMIC were referred for transanal excision, and all other lesions were treated via WF-EMR. In the U-ESD strategy, all lesions were initially treated by ESD regardless of SMIC suspicion. In the S-ESD strategy, lesions with high suspicion for SMIC were treated with ESD; others were treated with WF-EMR. S-ESD was the least expensive strategy and more effective than WF-EMR by preventing 19 additional surgeries per 1000 patients. Compared with S-ESD, U-ESD prevented another 13 surgeries at an incremental cost per surgery of $210,112^[52]. Therefore, S-ESD is the preferred treatment strategy for large LSTs [Table 2].

LSTs are macroscopically classified into granular (LST-G) and nongranular types (LST-NG). LST-G is further subclassified into homogeneous (LST-G-H) and nodular mixed (LST-G-M) types. LST-NG is further subclassified into flat elevated (LST-NG-F) and pseudo-depressed (LST-NG-PD) types. Such a classification is clinicopathologically relevant, and one meta-analysis showed that LST-NG-PD had an extraordinarily higher risk of SMIC (31.6%) than LST-NG-FE (4.9%), LST-G-H (0.5%), and LST-G-NM (10.5%), and SMIC was more common in distally located LSTs than in proximally located LSTs (odds ratio: 2.50; 95%CI: 1.24-5.02)^[53]. Sekiguchi et al.^[54] reported that, for colon/rectal LST-NG ≥ 2 cm and LST-G-M ≥ 3 cm, when compared with pEMR, ESD was more cost-effective with a higher cost that was offset by fewer recurrences and surgeries. The ICERs for avoided recurrence and surgery for ESD against pEMR were $3575-4521 and $69,604-77,689, respectively. The probability of ESD being chosen as a more cost-effective option than pEMR was > 50% at willingness-to-pay values of ≥ $3795-4744 for avoiding recurrence and ≥ $90,143-99,631 for avoiding surgery^[53]. Due to the higher cost of ESD, whether the ESD strategy is cost-effective depends not only on cost settings but also on the skill level of the endoscopist, which largely affects the time cost of ESD as well as the WTP value for avoiding recurrence/surgery. Further study is warranted.

COST-EFFECTIVENESS ANALYSES OF DIFFERENT COLONOSCOPIC SURVEILLANCE STRATEGIES

Determination of the surveillance interval after colonoscopy is currently based on the risk level as determined by colonoscopic findings^[5]. Sonnenberg’s study reported that the most common reason for performing colonoscopy is surveillance of previously discovered colorectal neoplasia, either adenoma or cancer^[8]. Five studies assessed the cost-effectiveness of colonoscopic surveillance strategies with different surveillance intervals in patients with adenomas [Table 3]^[55-59]. Four studies reported that, compared with no surveillance strategy, colonoscopic surveillance strategies were more costly but had more LY/QALY gain, making them more cost-effective strategies^[55-58]. Only one study showed that colonoscopic surveillance strategies were less costly with more QALY gain^[59]. In Arguedas’ study, colonoscopic surveillance every three years for patients with large adenoma and five years for patients with small adenoma or without adenoma were cost-effective and had an ICER of $27,970 per LY gained compared to no surveillance^[55]. Shaukat et al.^[56] also reported that colonoscopy surveillance every three years for patients with large adenoma and every five years for patients with small adenoma or without adenoma was cost-effective and had an ICER of $20,600 per LY gained. Saini et al.^[58] found that, compared with no surveillance (colonoscopy every 10 years for patients with any endoscopic finding), the strategy of colonoscopy surveillance every 3 years for high-risk patients and 10 years for low-risk patients was highly cost-effective, with an ICER of $5743 per QALY gained. In contrast, the strategy of colonoscopic surveillance every three years for high-risk patients and five years for low-risk patients was relatively expensive, with an ICER of $296,266 per QALY gain. The strategy of colonoscopy surveillance every three years for all patients only resulted in additional cost and harm. Hassan et al.^[57] simulated 60-year-old patients with adenoma and found that single colonoscopy surveillance after one year was cost-effective, with an ICER of $66,136 per LY gained. These results reveal the relatively high prevalence of CRC one year after clearing polypectomy and a relative deficiency of the current guidelines to exclude a clinically meaningful risk of CRC after polypectomy^[57]. Meester et al.^[59] conducted a cost-effectiveness analysis of different screening strategies. In patients with low-risk adenomas in the FIT screening program, the risk of CRC incidence and mortality was higher than those with LRAs in the colonoscopy screening program. Thus, the cost-effectiveness ratio for high-intensity surveillance (5 years) versus low-intensity surveillance (10 years) was more favorable (ICER = $11,100 per QALY gained) than in patients with low-risk adenomas in the colonoscopy screening program. However, the overall risk and benefit of surveillance in patients with high-risk adenomas in the FIT screening program versus in the colonoscopy screening program were similar. The ICERs for high-intensity surveillance (three years) versus low-intensity surveillance for patients with high-risk adenomas were similar ($11,100 per QALY gained) in the FIT and colonoscopy screening programs^[59].

Recent studies have shown the usefulness of using FIT in detecting advanced neoplasms during colonoscopic surveillance intervals. For example, Cross et al.^[60] reported that in intermediate-risk patients (with three to four adenomas < 10 mm or at least one ≥ 10 mm), the cumulative sensitivity and specificity for CRC of three rounds of annual FIT with low threshold levels for fecal hemoglobin (10 µg/g) were 91.7% (95 %CI: 73.0 - 99.0) and 69.8 % (95 %CI: 68.5 - 71.1), respectively. However, in this study, the three-year program sensitivities for CRC and advanced adenomas of annual FIT at the threshold level of 10 µg/g were 72% and 57%, respectively. This strategy would result in 28% of CRCs and 43% of advanced adenomas being missed compared with three-year colonoscopic surveillance^[60]. A recent study by Peng et al.^[61] from the Taiwan CRC Screening Program revealed that those who received subsequent FIT after negative colonoscopy had a significantly lower risk of incident CRCs than those who did not, which was assumed to be mainly attributable to the detection of noncancerous advanced adenoma that might have been missed at the initial colonoscopy. Whether such an “interval-FIT” approach could become a cost-effective alternative or an adjunctive approach to colonoscopic surveillance is largely unknown and requires further study.

EMERGING ISSUES THAT REQUIRE FURTHER EXPLORATION

Although colonoscopy and related procedures are cost-effective for CRC screening and post-polypectomy surveillance, there are still some issues that require further exploration [Table 4]. First, the application of new technologies such as artificial intelligence (AI) in colonoscopy is now attracting substantial attention, and its cost-effectiveness is worthy of investigation^[62]. The major roles of CAD during colonoscopy include computer-aided detection (CADe) and computer-aided characterization (CADx)^[63]. CADe helps endoscopists detect polyps that could be overlooked visually, and CADx improves accurate characterization by optical biopsy^[63]. Mori et al.^[64] demonstrated that the use of AI enabled the “diagnose-and-leave” strategy and resulted in substantial cost reductions for colonoscopy compared with the resect-all-polyps strategy. However, there remains no comprehensive cost-effectiveness analysis conducted for the application of AI in colonoscopy. This is because this technology is still in its primitive stage, and many clinical and cost parameters remain unclear. For example, AI tools can increase the detection of small adenoma, and it may result in a substantial increment in the pathologic examination cost. Moreover, it remains unclear for which level of endoscopists AI is most beneficial; therefore, using universal ADR parameters may lead to an erroneous conclusion. Further study is anticipated. Second, exploration of the cost-effectiveness of CRC screening with different strategies in young (40-49 years) adults is an emerging issue due to the increasing incidence of young-onset CRC in many countries^[65,66]. According to the recent modeling study by the US Preventive Services Task Force, if screening were to begin at age 45 years instead of age 50, the models of CRC screening strategies with stool tests, colonoscopy, or CTC showed 22-27 additional LY gained (8%-9% increase) and required 161-784 additional colonoscopies (10%-23% increase) per 1000 persons, and these results suggest that CRC screening starting at age 45 years provides an efficient balance of colonoscopy burden and life-years gained^[67]. Given the different young-onset CRC epidemiology and the healthcare facts across different countries, further cost-effectiveness analysis studies based on local scenarios are needed. Third, omics technology has provided integrated proteogenomic analysis for precision medicine, especially for cancer prevention. Targeted population screening for CRC would reduce the burden and demand for colonoscopy^[68]. Comprehensive assessment of the risk of CRC and the development of reliable and verifiable risk stratification tools that can be used in the clinic are essential. Three studies reported that the current uniform colonoscopy screening strategy appeared more cost-effective than personalized risk-based screening strategies based on polygenic risk profile and family history^[69-71]. However, cost-effectiveness is highly dependent on the costs of determining risk, and such costs may decrease in the future along with the further advance of omic technologies. Furthermore, cost-effectiveness analysis studies for personalized CRC screening with different strategies are needed. Fourth, endoscopic full-thickness resection (EFTR) has been demonstrated to be effective and safe for non-lifting and difficult lesions^[72]. Only one cost-effectiveness analysis study reported that the mean costs per R0 resection were €3708.98 for EFTR, €3115.10 for standard endoscopic resection, and €8924.05 for surgical treatment. EFTR is cost-effective in comparison with surgical and endoscopic treatment for colonic non-lifting and difficult lesions^[73]. Moreover, long-term follow-up is also needed to further assess the cost-effectiveness of EFTR. Fifth, many assumptions in the cost-effectiveness analyses limit the validity of those studies. For example, colonoscopy quality or skill level of endoscopic diagnosis (ability to accurately diagnose neoplasms and estimate invasion depth) or treatment such as polypectomy/EMR/ESD by individual endoscopists were assumed to be consistent in most of the cost-effectiveness models. Furthermore, the payment for colonoscopy varies greatly from country to country and ranges from $45.2 to $1695. Colonoscopy payments tend to be higher in countries with higher gross domestic products (GDPs) per capita, but a significant discrepancy still exists even within the same category of GDP level [Figure 1 and Supplementary Table 1], which has a significant impact on the results of cost-effectiveness analyses and the optimal strategies. When the payment for colonoscopy is low, colonoscopy-based screening or more frequent surveillance strategies will become the dominant or more cost-effective strategy over other noninvasive or less intensive strategies. More comprehensive cost-effectiveness analysis studies that take into account the diversity of neoplasms, heterogeneity of the clinical practice scenario, or the existing comorbidities of the patients are needed.

Figure 1. Payment of colonoscopy worldwide in 2020.

CONCLUSIONS

Gaps between limited healthcare resources and expanding healthcare service demands, and the development of new technologies have become a significant pressure on many healthcare systems worldwide and generate the demand for comprehensive health technology assessment, including economic appraisals, for scientific policymaking. Our current review provides updates on the economic appraisal of colonoscopic practices and their related procedures in the context of population CRC screening and surveillance. Many of the cost-effective strategies have been implemented in our current practice, but some have not, which may be due to the insufficient diversity being accommodated in the parameters or assumptions used in the analyses and the limited number of high-quality studies or clinical trials from which parameters were derived. Finally, whether a cost-effective strategy is applicable to the individual program still needs careful and specific consideration, which must include the healthcare context and the availability of certain services in the individual countries or healthcare systems, patient preference, and WTP levels.