Characterizing the epidemiology and natural history of colorectal cancer using fecal immunochemical test data from screening programs: a modelling study

Characterizing the epidemiology and natural history of colorectal cancer using fecal immunochemical test data from screening programs: a modelling study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Characterizing the epidemiology and natural history of colorectal cancer using fecal immunochemical test data from screening programs: a modelling study Kathy Leung, Zhenyu Wang, Joseph Wu, Horace C. W. Choi, Wai Keung Leung, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8110710/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Colorectal cancer (CRC) epidemiology remains insufficiently characterized in many settings, limiting optimal prevention strategies. Using quantitative fecal immunochemical test (FIT) data from 248,692 first-time participants (aged 49–77; 56% female) in the Hong Kong CRC Screening Programme, we developed a natural history model incorporating adenoma and serrated pathways with stage-specific FIT distributions. Colonoscopy referral was triggered if either of two submitted samples exceeded 100 ng/mL (13% positivity). We estimated that 37% (95% credible interval = 36–39%) of males and 27% (26–29%) of females had colorectal neoplasms at age 50; ~8% had advanced colorectal neoplasms (advanced adenoma, serrated lesions, or CRC). Prevalence of advanced neoplasms increased ~ 0.5% per year after age 50. Annual progression to CRC was ~ 4% for advanced adenoma and 1–2% for serrated lesions. Preclinical CRC advanced from stages I-II to III-IV within 3–4 years. At the 100 ng/mL threshold, FIT demonstrated 88–97% sensitivity for CRC. The positive predictive value for advanced neoplasms rose from ~ 20% at age 50 by ~ 1% annually, while the negative predictive value remained > 90%. Males with advanced neoplasms had higher FIT values than females. Quantitative FIT data thus enables robust characterization of CRC epidemiology and progression, providing a foundation for evaluating screening strategies and cost-effectiveness. Health sciences/Gastroenterology/Gastrointestinal diseases/Gastrointestinal cancer/Colorectal cancer Health sciences/Oncology/Cancer/Cancer screening Health sciences/Oncology/Cancer/Cancer prevention Health sciences/Health care/Public health/Population screening Health sciences/Medical research/Epidemiology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction ​Colorectal cancer (CRC) is a major global health threat​​, ranking third in incidence and second in mortality among all cancers worldwide. 1 Early detection through screening significantly reduces CRC burden, 2,3 but the optimal strategy remains a subject of ongoing debate. Screening initiation age varies across countries: in Europe and the Asia Pacific, the most common recommendation is to begin at 50 years, whereas in the United States, recent epidemiological trends have prompted a lowering of the recommended starting age to 45 years. 4 – 6 In addition to age considerations, population-based CRC screening programs primarily employ two categories of screening methods: stool-based tests and direct visualization techniques, such as flexible sigmoidoscopy and colonoscopy. 6 Among stool-based tests, the fecal immunochemical test (FIT), which detects the concentration of occult blood in the stool, is widely implemented in large-scale screening programs in countries including the United Kingdom, Australia, South Korea, and Japan. 7 , 8 In most of these programs, individuals undergo FIT as the primary screening modality, and those with positive results are referred for colonoscopy to confirm diagnosis. In contrast, the United States predominantly utilizes direct visualization techniques, particularly colonoscopy, as the primary screening modality. Despite its high diagnostic accuracy, colonoscopy has disadvantages as a primary screening modality: it is invasive, resource-intensive, and less accessible, leading to lower adherence. As such, FIT has gained prominence as a scalable and non-invasive alternative to triage individuals who would benefit from colonoscopy and further diagnosis procedures. 9 , 10 With comparable performance in cancer detection, FIT-based strategies have demonstrated improved participation and helped mitigate disparities in screening uptake. 10 Importantly, emerging evidence suggests such FIT-based screening is non-inferior to primary colonoscopy screening in reducing CRC mortality. 11 Nevertheless, FIT data remains underutilized for optimizing population-based screening. Most screening programs apply a binary threshold: individuals with FIT concentrations above a predefined cut-off are referred for colonoscopy, while those below are re-tested at regular intervals, typically every two years. 12 Emerging evidence links higher FIT concentrations to more severe colorectal neoplasms, 13,14 indicating that FIT concentrations encode information on disease severity and hence could inform inference on local epidemiology and the natural history of CRC. A locally derived, data-driven model that fully leverages quantitative FIT values could therefore be valuable for optimizing screening strategies. Guided by this rationale, we developed a modeling framework based on routine quantitative FIT data from the Hong Kong Colorectal Cancer Screening Programme (HKCRCSP). 15 Using this framework, we aimed to characterize local CRC epidemiology and progression, encompassing both the conventional adenoma-carcinoma and serrated pathways. Results Our dataset comprised 108,493 (44%) males and 140,199 (56%) females aged 49–77 years who were first-time participants of the HKCRCSP. Among them, 17,717 (16%) male and 14,868 (11%) female screenees tested FIT-positive and were referred for colonoscopy. Of these FIT-positive screenees, 84% had completed their colonoscopy with diagnosis (Fig. 1 a). Overall, 13,027 (12%) male and 9,761 (7.0%) female screenees were diagnosed with colorectal neoplasms (i.e., HP, NA, SL, AA, or CRC). Specifically, when considering advanced colorectal neoplasms (i.e., SL, AA, or CRC), the case counts and the proportions relative to the overall gender-specific population were: (i) 685 cases of SL (0.63%), 2,107 of AA (1.9%), and 1,068 of CRC (0.98%) in males; and (ii) 659 cases of SL (0.47%), 1,282 of AA (0.91%), and 678 of CRC (0.48%) in females (Table S1 ). These statistics did not include (undiagnosed) colorectal neoplasms among individuals who were either FIT-negative or FIT-positive but had not yet completed their colonoscopy diagnosis. The proportion of FIT-positive individuals increased with age in both genders and laboratories (Figure S2). However, when stratified by disease stage, FIT concentrations showed no significant association with age (Figure S5; Kruskal-Wallis test, p > 0.05). Given that FIT values tended to rise with the severity of neoplasms across both genders and laboratories (Fig. 2 ), the age-related increase in FIT positivity likely reflects a higher prevalence of advanced colorectal neoplasms in older individuals, rather than a direct effect of age on FIT concentrations. In other words, the higher FIT positivity observed among older individuals simply reflected a shift in disease burden toward more advanced lesions. This interpretation is supported by our model-based estimates (Figure S9). The fitted model was congruent with the data on stage-specific FIT distribution, and CRC incidence and mortality (Figure S10). Tables S3 and S4, Figs. 3 – 5 , and S12 summarize our estimates of the key epidemiologic parameters and the FIT value distributions. Particularly, we used the maximum two-sample FIT values (2-FIT-max) as the summary statistic of the empirical and estimated two-sample FIT distributions. Prevalence and natural history. The estimated prevalence of colorectal neoplasms increased with age (Fig. 4 ). By age 50 (the start age of HKCRCSP), the prevalence of colorectal neoplasms was 37.4% (95% credible interval [CrI] = 35.6%-39.1%) in males and 27.4% (25.8%-29.1%) in females. Between the ages of 50 and 70, the prevalence of NA and HP increased substantially in both genders. This is mainly driven by (i) rapidly increasing incidence of NA and HP with age along both pathways, and (ii) slow annual progression rates (around 1%) of NA and HP before age 70 in both genders (Fig. 3 ). The prevalence of advanced colorectal neoplasms also increased notably after age 50. At age 50, the prevalence of advanced colorectal neoplasms was 3.1% (2.9%-3.3%) in males and 2.2% (2.1%-2.4%) in females. By age 70, the prevalence of advanced colorectal neoplasms would have increased three- or four-fold compared with age 50, reaching 11% (11%-12%) in males and 7.8% (7.3%-8.2%) in females (Figure S14). Among those aged 70 with advanced colorectal neoplasms, 16% (15%-17%) of males and 12% (11%-14%) of females had CRC (Figure S15). Disease progression accelerated notably as lesions advanced along both pathways (Table S3). In the conventional pathway, the annual progression rate from AA to preclinical (undiagnosed) CRC was approximately 5–6 times higher than that from NA to AA. In the serrated pathway, the progression rate from SL to preclinical CRC was 30%-60% higher than that from HP to SL. The annual progression rate to preclinical CRC was around 4% from AA and 1–2% from SL (Table S3). The public health benefit of CRC screening can be clearly deduced from the age- and gender-specific prevalence and progression durations of preclinical CRC. These epidemiological parameters provide a quantitative basis for estimating opportunities for timely case detection and the potential mortality reduction through screening. At age 50, the prevalence of preclinical CRC was 0.40% (0.37%-0.42%) in males and 0.25% (0.23%-0.28%) in females, with an absolute increase of approximately 0.09% in males and 0.06% in females per year from this age onwards. For those aged 50, the mean durations of preclinical CRC stages 1–4 were: (i) 1.16 (1.02–1.32), 1.87 (1.66–2.08), 4.78 (4.18–5.37) and 3.64 (2.89–4.70) years for males; and (ii) 1.43 (1.17–1.78), 1.33 (1.10–1.60), 4.45 (3.92–5.14) and 3.28 (2.51–4.17) years for females. Stage progression also accelerated with age. For example, the mean sojourn times for all stages at age 75 were approximately 15% shorter than at age 50 for both genders (Fig. 3 ). FIT test performance. Males with advanced colorectal neoplasms generally exhibited higher FIT values compared to females (Fig. 5 ). Consequently, the FIT-positivity proportion among those with advanced colorectal neoplasms was substantially higher in males (Figure S9). At a threshold of 100 ng/mL for CRC detection, FIT demonstrated high sensitivity in both genders: (i) 96% (96%-97%) for males and 88% (88%-90%) for females in laboratory A, and (ii) 97% (96%-98%) for males and 97% (95%-98%) for females in laboratory B. Specificity decreased with age because the prevalence of colorectal neoplasms among non-CRC cases increased with age. At age 50, specificity was (i) 93% (93%-93%) for males and 96% (96%-96%) for females in laboratory A, and (ii) 90% (89%-90%) and 92% (92%-92%), respectively, in laboratory B. Between the ages of 50 and 75, specificity decreased by approximately 0.4% per year. The negative predictive value (NPV) exceeded 99% for CRC across all ages and remained above 90% for advanced colorectal neoplasms up to age 75. In comparison, the positive predictive value (PPV) was notably lower: approximately 5% for CRC and 20% for advanced colorectal neoplasms in both genders at age 50, with an absolute increase of about 0.2% and 0.7% per year from this age onwards, respectively (Figure S17). As expected, decreasing the FIT positivity threshold improved sensitivity and NPV at the expense of reduced specificity and PPV (Figures S18 and S19). Discussion To our knowledge, our study is the first to characterize the local epidemiology and natural history of CRC using FIT data from a screening program. Our model integrates quantitative FIT results with local epidemiologic data to infer the gender- and age-specific prevalence of colorectal neoplasms. Such locally derived estimates are crucial, as CRC epidemiology varies widely across populations due to differences in demographics, genetics, lifestyle, and healthcare quality. 16 Our findings also show that screening test performance, including sensitivity, specificity, PPV, and NPV, can vary substantially depending on local epidemiology, screenee characteristics (e.g., gender), FIT-positivity threshold, and processing laboratory (Figures S17-S19). These results highlight the need for local evidence to guide screening policies. By combining local FIT and epidemiologic data, our study addresses this gap and provides a transferable framework. Our findings suggested that a “one-size-fits-all approach” may not be optimal for CRC screening. For instance, we observed a higher prevalence of advanced colorectal neoplasms in males (Fig. 4 ), which suggests that men should start screening earlier than women, consistent with Germany’s recommended start screening age of 45 and 50 for males and females, respectively. 17 Moreover, our results suggest that at the same disease stage, females had lower FIT concentrations than males (Fig. 5 ), which resulted in comparatively lower sensitivity in females for any gender-agnostic threshold (Figure S17). We also found that the specificity of FIT decreased with age, indicating higher false-positive rates among older individuals. Taken together, adjusting FIT thresholds by gender and age can potentially improve equity and effectiveness. This is consistent with studies in Sweden, the Netherlands, and Denmark, which advocate age- and gender-specific thresholds to enhance efficiency. 18 – 20 More broadly, a data-driven approach for risk-based CRC prevention requires the incorporation of all relevant and available risk factors, including age, gender, and previous screening results. This capability enables the starting age to screen, screening intervals, and colonoscopy referrals to be tailored according to an individual’s estimated risk of advanced colorectal neoplasms. In particular, older individuals are associated with a higher risk of colorectal neoplasms, 21 along with higher FIT concentrations in prior FIT-negative tests, which are linked to an increased likelihood of subsequent colorectal neoplasm detection. 22 Therefore, colonoscopy referrals might also be considered for older FIT-negative screenees whose FIT values are just marginally below the positivity threshold. For instance, in our study, we estimated that among FIT-negative screenees with 2-FIT-max values between 50 and 100 ng/ml, around 5%-15% had undiagnosed advanced colorectal neoplasms and 0.1%-0.5% had undiagnosed CRC (Figure S20), with the proportions increasing with higher FIT values and advancing ages. Our estimated progression rates of CRC highlight the public health importance of CRC screening. We estimated that the average time from early-stage preclinical CRC (stage 1 or 2) to late-stage disease (stage 3 or 4) was approximately 3 to 4 years (Fig. 3 ). Given the marked difference in 5-year survival rates following clinical diagnosis—96%, 87%, 69%, and only 9% in Hong Kong for stages 1 through 4, respectively— the rapid progression of preclinical CRC underscores the narrow window for effective intervention before the disease advances. 23 Without timely diagnosis, individuals with undetected early-stage preclinical CRC may quickly advance to late-stage disease, where treatment options are limited and outcomes are substantially worse. By quantifying these progression dynamics locally, our study offers a generalizable framework for optimizing the design and cost-effectiveness of screening. Our study has several limitations. First, our study lacks information on other risk factors associated with CRC, such as family history of CRC, smoking history, body mass index, history of diabetes, and other predictors of advanced colorectal neoplasms (e.g., excessive alcohol use and physical inactivity). 21 , 24 Addressing this gap would require secure data linkage between the screening program (managed by the Department of Health) and the electronic health records of the screenees (managed by the Hospital Authority). Second, our epidemiology inferences regarding CRC are based on the HKCRCSP screenees and hence may not be fully representative of the general population. For example, cancer screening services might be less accessible to individuals from low socioeconomic and educational backgrounds, as well as ethnic minority groups, due to insufficient awareness and time constraints. 25 – 27 Third, the disease stages of FIT-negative screenees were unknown in our dataset because they did not undergo colonoscopy. Although their disease stage could be inferred within our framework, the robustness of the resulting estimates of disease prevalence and progression rates depends on the validity of our model assumptions (e.g., stage-specific FIT value distributions are independent of age, FIT values are stochastically larger at more advanced stages, etc.). Future colonoscopy-based studies that estimate disease prevalence among FIT-negative screenees could help validate our results. Finally, we did not distinguish between colon and rectal cancers, despite evidence of divergent local incidence trends and potentially heterogeneous etiologies and natural histories. 28 More granular data would be needed to calibrate a refined model that more accurately characterizes the epidemiology of CRC. 29 In conclusion, quantitative FIT data from screening programs can be harnessed to characterize the epidemiology and natural history of CRC. Our novel, data-driven framework provides a robust and adaptable platform for evaluating and optimizing the cost-effectiveness of CRC screening programs. Methods Study design and population Launched as a pilot program in September 2016 and fully implemented in August 2018, the HKCRCSP provides biennial CRC screening to individuals aged 50-75. 15 Screenees are required to submit two fecal samples taken within 5 days apart. These sister samples were processed by laboratory A (using NS Prime, Alfresa Pharma Co., Ltd) before September 2019 and laboratory B (using OC-Sensor, Eiken Chemical Co., Ltd) thereafter. Screenees are regarded as FIT-positive if one or both of their FIT values are 100 ng/mL (equivalent to 20 µg/g) or above, and FIT-negative otherwise. FIT-positive screenees are referred for colonoscopy to confirm diagnosis, and FIT-negative screenees are invited to screen at a two-year interval (Fig. 1 a). Please see the Supplementary Appendix (SA, pp 2–7) for more details about the HKCRCSP. According to the 2024 population in Hong Kong, 30 8.6% of the eligible population has participated in the HKCRCSP. We utilized three data sources to estimate the epidemiology and natural history parameters of CRC and the FIT value distributions by disease stages: (i) records from first-time screenees of the HKCRCSP who enrolled between September 2016 and September 2021, including their age, two-sample FIT values, and colonoscopy diagnoses for those who tested FIT-positive and completed colonoscopy; 15 (ii) population-level CRC incidence and mortality statistics for 2012–2016 from the Hong Kong Cancer Registry (HKCaR); 23 and (iii) all-cause mortality statistics for 2012–2016 from the Hong Kong Census and Statistics Department. 30 In this study, colonoscopy outcomes were defined as follows: (i) normal (N; no abnormality, no polyps, or only non-significant histology); (ii) hyperplastic polyps (HP; hyperplastic lesions < 10 mm); (iii) non-advanced adenoma (NA; tubular adenomas < 10 mm with low-grade dysplasia); (iv) serrated lesions (SL; sessile serrated lesions or traditional serrated adenomas of any size, or hyperplastic polyps ≥ 10 mm); (v) advanced adenoma (AA; tubular adenomas ≥ 10 mm or high-grade dysplasia, or adenomas with villous features); (vi) colorectal cancer (CRC; histologically confirmed adenocarcinoma); and (vii) unknown disease stage (Unk; those who were FIT-negative or FIT-positive without completing colonoscopy). 31 See Figure S4 for detailed definitions of colonoscopy findings. We assumed that, before the first screening, the underlying epidemiology and natural history of CRC among screenees reflected that of the general population. This study was approved by the Institutional Review Board of The University of Hong Kong / Hospital Authority Hong Kong West Cluster (UW 17–214). The model Our data-driven model comprises two components: (1) a natural history model that simulates lifetime development and progression of CRC; and (2) a statistical model that characterizes the probability distribution of FIT values across different disease stages. Here, we briefly describe the two model components. Please see the SA (pp 8–25) for more details. Natural history . We developed a gender-specific natural history model with two pathways of colorectal neoplasm (Fig. 1 b): 1. Conventional adenoma pathway: N → NA → AA → pCRC → dCRC 2. Serrated pathway: N → HP → SL → pCRC → dCRC where pCRC and dCRC are referred to as preclinical (undiagnosed) and diagnosed CRC, respectively. pCRC represents those individuals who have already developed CRC but remain undiagnosed (e.g., due to the absence of symptoms or delay in care-seeking). The prevalence of each disease stage \(\:s\) for \(\:s\in\:\left\{N,HP,\:NA,SL,AA,CRC\right\}\) at age 20 and the age-specific disease progression rates were subject to statistical inference (Table S3). Based on previous studies, we assumed that the serrated pathway accounts for 10–30% of all CRC cases and inferred the corresponding proportion through model calibration. 32 – 34 The model simulated the epidemiology of colorectal neoplasms from age 20 through death or up to age 100, whichever occurs first. Furthermore, individuals diagnosed with CRC experienced stage-specific relative survival as reported from local empirical data. Please see the SA (pp 8–11) for more details. FIT value distribution . We modeled the probability distribution of FIT values across disease stages based on the patterns observed in the empirical screening data (Figs. 2 , S2, and S3). The screening data indicated no statistically significant association between FIT values and age within any disease stage (Figure S5). As such, we assumed that for each screenee, the FIT value of a given sample \(\:V\) depended on an underlying FIT value \(\:U\) such that: 1. the mean and variance of \(\:U\) are stage-specific; 2. \(\:V=0\) if \(\:U=0\) ; and 3. \(\:\text{ln}\left(1+V\right)=\text{ln}\left(1+U\right)+\epsilon\:\) if \(\:U>0\) , where \(\:\epsilon\:\) is a normal random variable with mean 0 and standard deviation \(\:\omega\:\left(U\right)\) , i.e., \(\:\epsilon\:\sim\:\text{N}(0,\omega\:(U\left)\right)\) . Please refer to SA (pp 12–17) for a full description of the statistical inference of the parameters for the FIT value distributions. Parameter estimation . We used \(\:\theta\:\) to denote the set of model parameters subject to statistical inference (see Fig. 3 , Tables S3 and S4). Let \(\:p\left(s\right|a,\theta\:)\) be the prevalence of disease stage \(\:s\in\:\{N,HP,\:NA,SL,AA,CRC\}\) among individuals of age \(\:a\) . Let \(\:f\left({v}_{1},{v}_{2}|s,l,\theta\:\right)\) be the probability density function of the two-sample FIT values ( \(\:{v}_{1},{v}_{2})\) for individuals with disease stage \(\:s\) and processed by laboratory \(\:l\) . Let \(\:{\lambda\:}_{xj}^{HKCaR}\) and \(\:{\lambda\:}_{xj}^{model}\) be the incidence of CRC stage \(\:x\) ( \(\:x\) = 1, 2, 3, 4, and “unstaged”) in the j th age group in the HKCaR database and our model, respectively. Similarly, let \(\:{z}_{j}^{HKCaR}\) and \(\:{z}_{j}^{model}\) be the corresponding CRC mortality statistics. We assumed that annual CRC incidence and mortality follow Poisson distributions. Let \(\:{C}^{+}\) be the group of FIT-positive screenees who had received their colonoscopy diagnosis and \(\:{C}^{-}\) be the remaining screenees (who were either FIT-negative or FIT-positive but had not yet completed their colonoscopy diagnosis). The likelihood function was: $$\:L\left(\theta\:\right)={L}_{{C}^{+}}\left(\theta\:\right)\times\:{L}_{{C}^{-}}\left(\theta\:\right)\times\:\underset{\text{CRC\:incidence}}{\underset{⏟}{\prod\:_{x}^{}{\prod\:}_{j}Poisson({\lambda\:}_{xj}^{HKCaR},{\lambda\:}_{xj}^{model})}}\times\:\underset{\text{CRC\:mortality}}{\underset{⏟}{{\prod\:}_{j}Poisson({z}_{j}^{HKCaR},{z}_{j}^{model})}}$$ where $$\:{L}_{{C}^{+}}\left(\theta\:\right)=\prod\:_{i\in\:{C}^{+}}p\left({s}_{i}\right|{a}_{i},\theta\:\left)f\right({v}_{i1},{v}_{i2}|{s}_{i},{l}_{i},\theta\:)$$ corresponded to screenees in group \(\:{C}^{+}\) and $$\:{L}_{{C}^{-}}\left(\theta\:\right)=\prod\:_{i\in\:{C}^{-}}\sum\:_{s\in\:\{N,HP,NA,SL,AA,CRC\}}p\left(s\right|{a}_{i},\theta\:\left)f\right({v}_{i1},{v}_{i2}|s,{l}_{i},\theta\:)$$ corresponded to screenees in group \(\:{C}^{-}\) . Note that although screenees in group \(\:{C}^{-}\) (primarily FIT-negative) did not have colonoscopy outcomes, their FIT values contained information on the likelihood of their disease stages. The likelihood component \(\:{L}_{{C}^{-}}\left(\theta\:\right)\) encapsulated such information where \(\:s\) is a dummy variable denoting the possible disease stages. We estimated \(\:\theta\:\) by fitting the model to data using Markov chain Monte Carlo (MCMC) methods. Please see the SA (pp 18–25) for details on parameter inference and the assessment of model identifiability. Data availability The data that support the findings of this study are available from the Hong Kong Hospital Authority, but restrictions apply to the availability of these data, which were used only for the current study, and so are not publicly available. Code availability Code supporting the findings of this study is available upon request. Abbreviations AA advanced adenoma CRC colorectal cancer CLN colonoscopy CrI credible interval d CRC diagnosed colorectal cancer FIT fecal immunochemical test FIT-neg FIT-negative FIT-pos FIT-positive HKCRCSP Hong Kong Colorectal Cancer Screening Programme HP hyperplastic polyps N normal NA non-advanced adenoma p CRC preclinical colorectal cancer SL serrated lesions Unk unknown disease stage 2-FIT-max maximum two-sample FIT values. Declarations We confirmed that informed consent was obtained from all participants of the Hong Kong Colorectal Cancer Screening Programme by the Hong Kong Department of Health, and the need for consent to conduct this research with the anonymised data from participants was approved by the Institutional Review Board of The University of Hong Kong / Hospital Authority Hong Kong West Cluster (Ref: UW 17-214). Funding Health and Medical Research Fund (HMRF), Hong Kong Special Administrative Region, China. The funding bodies had no role in the design of the study or in the analysis and interpretation of data. Correspondence and requests for materials should be addressed to Joseph T. Wu and Kathy Leung. Competing interests: All authors report no conflicts of interest. Writing assistance: No writing assistance was provided in the preparation of the manuscript. Statement of authorship contributions: JTW, WKL, WLL, GML, and KL designed the study, guided the writing of the manuscript, and reviewed the manuscript. KL cleaned and managed data. JTW initially developed the model, performed the analyses, and drafted the manuscript. ZW and JTW further refined the analyses. ZW, JTW, HCWC, and KL extended the analyses and wrote the manuscript. JTW, KL, ZW, and HCWC have directly accessed and verified the underlying data. All authors interpreted the data and revised the manuscript. All authors read and approved of the final manuscript. References Bray F et al (2024) Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. https://doi.org/10.3322/caac.21834 Navarro M, Nicolas A, Ferrandez A, Lanas A (2017) Colorectal cancer population screening programs worldwide in 2016: An update. World J Gastroenterol 23:3632–3642. https://doi.org/10.3748/wjg.v23.i20.3632 Levin TR et al (2018) Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large Community-Based Population. Gastroenterology 155, 1383–1391.e1385. https://doi.org/10.1053/j.gastro.2018.07.017 Cardoso R, Guo F, Heisser T, Hoffmeister M, Brenner H (2020) Utilisation of Colorectal Cancer Screening Tests in European Countries by Type of Screening Offer: Results from the European Health Interview Survey. Cancers (Basel) 12. https://doi.org/10.3390/cancers12061409 Sung JJ et al (2015) An updated Asia Pacific Consensus Recommendations on colorectal cancer screening. Gut 64:121–132. https://doi.org/10.1136/gutjnl-2013-306503 U. S. Preventive Services Task Force et al. Screening for Colorectal Cancer: US Preventive Services Task Force Recommendation Statement. JAMA 325, 1965–1977 (2021) https://doi.org/10.1001/jama.2021.6238 Schreuders EH et al (2015) Colorectal cancer screening: a global overview of existing programmes. Gut 64:1637–1649. https://doi.org/10.1136/gutjnl-2014-309086 Young GP, Rabeneck L, Winawer SJ (2019) The Global Paradigm Shift in Screening for Colorectal Cancer. Gastroenterology 156, 843–851 e842 https://doi.org/10.1053/j.gastro.2019.02.006 Kim SY, Kim H-S, Park HJ (2019) Adverse events related to colonoscopy: Global trends and future challenges. World J Gastroenterol 25:190–204. https://doi.org/10.3748/wjg.v25.i2.190 Robertson DJ, Rex DK, Ciani O, Drummond MF (2024) Colonoscopy vs the Fecal Immunochemical Test: Which is Best? Gastroenterology 166, 758–771 https://doi.org/10.1053/j.gastro.2023.12.027 Castells A et al (2025) Effect of invitation to colonoscopy versus faecal immunochemical test screening on colorectal cancer mortality (COLONPREV): a pragmatic, randomised, controlled, non-inferiority trial. Lancet 405:1231–1239. https://doi.org/10.1016/S0140-6736(25)00145-X Shaukat A, Levin TR (2022) Current and future colorectal cancer screening strategies. Nat Rev Gastroenterol Hepatol 19:521–531. https://doi.org/10.1038/s41575-022-00612-y Senore C et al (2020) Faecal haemoglobin concentration among subjects with negative FIT results is associated with the detection rate of neoplasia at subsequent rounds: a prospective study in the context of population based screening programmes in Italy. Gut 69:523–530. https://doi.org/10.1136/gutjnl-2018-318198 Digby J et al (2025) Combining faecal haemoglobin, iron deficiency anaemia status and age can improve colorectal cancer risk prediction in patients attending primary care with bowel symptoms: a retrospective observational study. Gut. https://doi.org/10.1136/gutjnl-2024-334248 Hong Kong Colorectal Cancer Screening Program https://www.colonscreen.gov.hk/en/public/index.html ( Collaborators GBDCC (2022) Global, regional, and national burden of colorectal cancer and its risk factors, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Gastroenterol Hepatol 7:627–647. https://doi.org/10.1016/S2468-1253(22)00044-9 Sergeev D, Heisser T, Hoffmeister M, Brenner H (2025) Potential for enhancing efficacy of screening colonoscopy by lowering starting ages and extending screening intervals: A modelling study for Germany. Int J Cancer 156:2303–2310. https://doi.org/10.1002/ijc.35322 Blom J et al (2019) Gender-specific cut-offs in colorectal cancer screening with FIT: Increased compliance and equal positivity rate. J Med Screen 26:92–97. https://doi.org/10.1177/0969141318804843 Njor SH, Rasmussen M, Friis-Hansen L, Andersen B (2022) Varying fecal immunochemical test screening cutoffs by age and gender: a way to increase detection rates and reduce the number of colonoscopies. Gastrointest Endosc 95:540–549. https://doi.org/10.1016/j.gie.2021.09.038 Harlass M et al (2025) Benefits of colorectal cancer screening using FIT with varying positivity thresholds by age and sex. JNCI: J Natl Cancer Inst. https://doi.org/10.1093/jnci/djaf149 Kastrinos F, Kupfer SS, Gupta S (2023) Colorectal Cancer Risk Assessment and Precision Approaches to Screening: Brave New World or Worlds Apart? Gastroenterology 164, 812–827 https://doi.org/10.1053/j.gastro.2023.02.021 van den Berg DMN, van den Puttelaar R, de Jonge L, Lansdorp-Vogelaar I, Toes-Zoutendijk E (2025) Fecal Hemoglobin Levels in Prior Negative Screening and Detection of Colorectal Neoplasia: A Dose-Response Meta-Analysis. Gastroenterology 168:587–597. https://doi.org/10.1053/j.gastro.2024.10.047 Hong Kong Cancer Registry https://www3.ha.org.hk/cancereg/hkcar.html ( Wong MC et al (2014) A validated tool to predict colorectal neoplasia and inform screening choice for asymptomatic subjects. Gut 63:1130–1136 Moss SM et al (2012) Performance measures in three rounds of the English bowel cancer screening pilot. Gut 61:101–107. https://doi.org/10.1136/gut.2010.236430 Bozhar H et al (2022) Socio-economic inequality of utilization of cancer testing in Europe: A cross-sectional study. Prev Med Rep 26:101733. https://doi.org/10.1016/j.pmedr.2022.101733 Carethers JM, Doubeni CA (2020) Causes of Socioeconomic Disparities in Colorectal Cancer and Intervention Framework and Strategies. Gastroenterology 158:354–367. https://doi.org/10.1053/j.gastro.2019.10.029 Zhang B, Xie SH, Yu IT (2018) Differential incidence trends of colon and rectal cancers in Hong Kong: an age-period-cohort analysis. Cancer Commun (Lond) 38:42. https://doi.org/10.1186/s40880-018-0311-2 DeYoreo M, Rutter CM, Ozik J, Collier N (2022) Sequentially calibrating a Bayesian microsimulation model to incorporate new information and assumptions. BMC Med Inf Decis Mak 22:12. https://doi.org/10.1186/s12911-021-01726-0 Hong Kong Census and Statistics Department https://www.censtatd.gov.hk/home/index.jsp ( Lokuhetty D, Organization WH, Cancer IA (2019) f. R. o. WHO Classification of Tumours of the Digestive System Tumours. IARC East JE et al (2017) British Society of Gastroenterology position statement on serrated polyps in the colon and rectum. Gut 66:1181–1196. https://doi.org/10.1136/gutjnl-2017-314005 Szylberg L, Janiczek M, Popiel A, Marszalek A (2015) Serrated Polyps and Their Alternative Pathway to the Colorectal Cancer: A Systematic Review. Gastroent Res Pract 2015 https://doi.org/Artn 573814 1155/2015/573814 East JE, Vieth M, Rex DK (2015) Serrated lesions in colorectal cancer screening: detection, resection, pathology and surveillance. Gut 64:991–1000. https://doi.org/10.1136/gutjnl-2014-309041 Additional Declarations There is NO Competing Interest. Supplementary Files 2Appendix1114.docx Supplemental Appendix Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8110710","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":548861712,"identity":"3eef50cf-56d0-496e-bb20-c35f5a9cbf15","order_by":0,"name":"Kathy Leung","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAw0lEQVRIiWNgGAWjYBACCQYGxgMgBr8EG5DZkECUFgawFskZJGsxuEGsFsn+0wkHPrbdsdt8uy1NgnFHGmEt0hK5Gw7ObHuWvO3OsWMSjGdyCGuRk+DdcJi37XCy2Y30NgnGtgoitPCf3XD4L1CL8QxitUgz5G44zNh22M5AIg3osDYiHCY5A+iXnnOHEyRupCVbJLYR4X2J82c3PvhRdtief0aa4Y2PbcmEtcBAYgOITCBeAwODPSmKR8EoGAWjYIQBAKhEQLPEGDgmAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-4777-388X","institution":"The University of Hong Kong","correspondingAuthor":true,"prefix":"","firstName":"Kathy","middleName":"","lastName":"Leung","suffix":""},{"id":548861713,"identity":"1d6bd074-3609-4b47-90b1-d4405320d968","order_by":1,"name":"Zhenyu Wang","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Zhenyu","middleName":"","lastName":"Wang","suffix":""},{"id":548861714,"identity":"e24c5509-be96-4673-be54-db77f68a052c","order_by":2,"name":"Joseph Wu","email":"","orcid":"https://orcid.org/0000-0002-3155-5987","institution":"HKU","correspondingAuthor":false,"prefix":"","firstName":"Joseph","middleName":"","lastName":"Wu","suffix":""},{"id":548861715,"identity":"8982ffe5-d795-474d-89d3-4b0e659879c9","order_by":3,"name":"Horace C. W. Choi","email":"","orcid":"https://orcid.org/0000-0001-8387-8780","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Horace","middleName":"C. W.","lastName":"Choi","suffix":""},{"id":548861716,"identity":"344f4966-7500-4f59-8964-6ca3eb743922","order_by":4,"name":"Wai Keung Leung","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Wai","middleName":"Keung","lastName":"Leung","suffix":""},{"id":548861717,"identity":"5eb64a64-c4a7-42e9-ad25-120d58745623","order_by":5,"name":"Wai Lun Law","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Wai","middleName":"Lun","lastName":"Law","suffix":""},{"id":548861718,"identity":"98f2555c-f2b9-4efe-b204-0d49fafe6545","order_by":6,"name":"Gabriel Leung","email":"","orcid":"https://orcid.org/0000-0002-2503-6283","institution":"School of Public Health, LKS Faculty of Medicine, The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Gabriel","middleName":"","lastName":"Leung","suffix":""}],"badges":[],"createdAt":"2025-11-14 04:55:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8110710/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8110710/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97299135,"identity":"59ec6802-2faf-4fa2-a4c1-447321632197","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":160490,"visible":true,"origin":"","legend":"","description":"","filename":"1Manuscript1114.docx","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/0aa9e3f0fd16f1bf62bcb860.docx"},{"id":97369510,"identity":"28ae9284-e19a-46d4-959a-d26b85187641","added_by":"auto","created_at":"2025-12-03 16:25:03","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1018642,"visible":true,"origin":"","legend":"","description":"","filename":"4Figures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/5d093ab224ad0349e8d52a29.pdf"},{"id":97299130,"identity":"c5960c9f-6187-4e92-bf7d-9e6b77749a48","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"json","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8221,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2594200T.json","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/976d9d77b529fb59b2f0e31c.json"},{"id":97299143,"identity":"0be091dc-6cc9-4287-ae28-4046ad5c57bf","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7021712,"visible":true,"origin":"","legend":"","description":"","filename":"2Appendix1114.docx","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/9499f80fa7261e7cccfea0eb.docx"},{"id":97299138,"identity":"7b786672-4acf-47dc-9916-902bfd13e0e4","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102354,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2594200T0enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/fbdc63b7f7f8640c83d090fb.xml"},{"id":97299141,"identity":"cca2c222-8db1-4dbf-bf20-13fba9e2358d","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"pdf","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1018642,"visible":true,"origin":"","legend":"","description":"","filename":"4Figures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/81927b3156940da4a9f3dc58.pdf"},{"id":97299140,"identity":"4d77a101-086e-4f2c-899e-d9ff16192838","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"xml","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97357,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2594200T0structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/9f6033bb31889e2240607a55.xml"},{"id":97299142,"identity":"7ea161c4-7bcf-4c44-a558-db4e4ce9e4c6","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"html","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":114902,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/ad032c8c1ea9196c9d969a2c.html"},{"id":97370590,"identity":"66946977-661b-4b4b-b006-6e1c79baf656","added_by":"auto","created_at":"2025-12-03 16:27:39","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":353977,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStudy participants and the natural history model. \u003c/strong\u003e(a). Study participants were stratified by gender, FIT-positivity, and colonoscopy completion status. (b). In the natural history model, individuals develop CRC via either the conventional adenoma or the serrated pathway withage- and gender-specific incidence rates. Disease progression rates depend on age, gender, and disease stage. \u003cstrong\u003eAbbreviations:\u003c/strong\u003e AA, advanced adenoma; \u003cem\u003edCRC\u003c/em\u003e, diagnosed colorectal cancer; FIT, fecal immunochemical test; HKCRCSP, Hong Kong Colorectal Cancer Screening Programme; HP, hyperplastic polyps; N, normal; NA, non-advanced adenoma; \u003cem\u003epCRC\u003c/em\u003e, preclinical colorectal cancer; SL, serrated lesions.\u003c/p\u003e","description":"","filename":"4Figures1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/03b03f5af94026ab9fca3425.jpg"},{"id":97370306,"identity":"396c087a-152b-406b-a68a-b66eaa6d0c9c","added_by":"auto","created_at":"2025-12-03 16:27:07","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":339342,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe observed one and two-sample FIT distribution in two laboratories\u003c/strong\u003e. There were three possible outcomes for each sample: invalid results, a FIT value of 0, and a positive FIT value. For positive FIT values, the laboratories applied different censoring thresholds: Laboratory A did not apply left-censoring, but the right-censoring threshold was 120,000 ng/mL. In contrast, laboratory B applied a left-censoring threshold at 9 ng/mL and a right-censoring threshold at 1,000 ng/mL. The ‘Unk’ stage comprises screenees who (i) were FIT-negative or (ii) were FIT-positive but whose colonoscopy examinations were not yet performed or were invalid. The ‘tail’ below 9 ng/mL in laboratory B was driven by FIT values of 0, an artefact of the data. \u003cstrong\u003eAbbreviations:\u003c/strong\u003e AA, advanced adenoma; CRC, colorectal cancer; CLN, colonoscopy; FIT, fecal immunochemical test; FIT-neg, FIT-negative; FIT-pos, FIT-positive; HP, hyperplastic polyps; N, normal; NA, non-advanced adenoma; SL, serrated lesions; Unk, unknown disease stage; 2-FIT-max, maximum two-sample FIT values.\u003c/p\u003e","description":"","filename":"4Figures2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/8a0629eadc184c93114ebb0a.jpg"},{"id":97368932,"identity":"3fd8fab8-2915-4ce8-89cd-dd7a26dd4092","added_by":"auto","created_at":"2025-12-03 16:23:15","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":297403,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe posterior estimates of key epidemiological and natural history parameters. Abbreviations:\u003c/strong\u003e AA, advanced adenoma; CRC, colorectal cancer; CrI, credible interval; HP, hyperplastic polyps; NA, non-advanced adenoma; SL, serrated lesions.\u003c/p\u003e","description":"","filename":"4Figures3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/61ceb049017fd36ac6951556.jpg"},{"id":97369747,"identity":"db1bde18-609b-4668-8022-241cf0606f17","added_by":"auto","created_at":"2025-12-03 16:25:40","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":264819,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEstimated age-specific prevalence of different disease stages. Abbreviations: \u003c/strong\u003eAA, advanced adenoma; CRC, colorectal cancer; CrI, credible interval; HP, hyperplastic polyps; N, normal; NA, non-advanced adenoma; SL, serrated lesions.\u003c/p\u003e","description":"","filename":"4Figures4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/b46ad9552ba0589691358e5c.jpg"},{"id":97368963,"identity":"c01cdee7-8a63-4fe1-bee0-2fb538e14854","added_by":"auto","created_at":"2025-12-03 16:23:22","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":242882,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCumulative distribution function (CDF) of the estimated 2-FIT-max stratified by health states.\u003c/strong\u003eThe black vertical dashed lines indicate the threshold of 100 ng/mL for FIT-positivity. \u003cstrong\u003eAbbreviations:\u003c/strong\u003e AA, advanced adenoma; CRC, colorectal cancer; CrI, credible interval; FIT, fecal immunochemical test; HP, hyperplastic polyps; N, normal; NA, non-advanced adenoma; SL, serrated lesions; 2-FIT-max, maximum two-sample FIT values.\u003c/p\u003e","description":"","filename":"4Figures5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/50825f7b4d8e77b1a76bb6b6.jpg"},{"id":97664504,"identity":"04684974-9c51-4a5c-bcde-d2ab63f9496d","added_by":"auto","created_at":"2025-12-08 09:06:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2225605,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/b6bc017a-fcc6-4161-a4b0-5b78adaf7e7b.pdf"},{"id":97299137,"identity":"b87c7eeb-895e-4707-804b-eea94855d2cb","added_by":"auto","created_at":"2025-12-03 00:48:57","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":7021712,"visible":true,"origin":"","legend":"Supplemental Appendix","description":"","filename":"2Appendix1114.docx","url":"https://assets-eu.researchsquare.com/files/rs-8110710/v1/fa6950a4621e53136834c73e.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Characterizing the epidemiology and natural history of colorectal cancer using fecal immunochemical test data from screening programs: a modelling study","fulltext":[{"header":"Introduction","content":"\u003cp\u003e​Colorectal cancer (CRC) is a major global health threat​​, ranking third in incidence and second in mortality among all cancers worldwide.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e Early detection through screening significantly reduces CRC burden,\u003csup\u003e2,3\u003c/sup\u003e but the optimal strategy remains a subject of ongoing debate. Screening initiation age varies across countries: in Europe and the Asia Pacific, the most common recommendation is to begin at 50 years, whereas in the United States, recent epidemiological trends have prompted a lowering of the recommended starting age to 45 years.\u003csup\u003e\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e In addition to age considerations, population-based CRC screening programs primarily employ two categories of screening methods: stool-based tests and direct visualization techniques, such as flexible sigmoidoscopy and colonoscopy.\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e Among stool-based tests, the fecal immunochemical test (FIT), which detects the concentration of occult blood in the stool, is widely implemented in large-scale screening programs in countries including the United Kingdom, Australia, South Korea, and Japan.\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e In most of these programs, individuals undergo FIT as the primary screening modality, and those with positive results are referred for colonoscopy to confirm diagnosis. In contrast, the United States predominantly utilizes direct visualization techniques, particularly colonoscopy, as the primary screening modality.\u003c/p\u003e\u003cp\u003eDespite its high diagnostic accuracy, colonoscopy has disadvantages as a primary screening modality: it is invasive, resource-intensive, and less accessible, leading to lower adherence. As such, FIT has gained prominence as a scalable and non-invasive alternative to triage individuals who would benefit from colonoscopy and further diagnosis procedures.\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e With comparable performance in cancer detection, FIT-based strategies have demonstrated improved participation and helped mitigate disparities in screening uptake.\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e Importantly, emerging evidence suggests such FIT-based screening is non-inferior to primary colonoscopy screening in reducing CRC mortality.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eNevertheless, FIT data remains underutilized for optimizing population-based screening. Most screening programs apply a binary threshold: individuals with FIT concentrations above a predefined cut-off are referred for colonoscopy, while those below are re-tested at regular intervals, typically every two years.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Emerging evidence links higher FIT concentrations to more severe colorectal neoplasms,\u003csup\u003e13,14\u003c/sup\u003e indicating that FIT concentrations encode information on disease severity and hence could inform inference on local epidemiology and the natural history of CRC.\u003c/p\u003e\u003cp\u003eA locally derived, data-driven model that fully leverages quantitative FIT values could therefore be valuable for optimizing screening strategies. Guided by this rationale, we developed a modeling framework based on routine quantitative FIT data from the Hong Kong Colorectal Cancer Screening Programme (HKCRCSP).\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e Using this framework, we aimed to characterize local CRC epidemiology and progression, encompassing both the conventional adenoma-carcinoma and serrated pathways.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eOur dataset comprised 108,493 (44%) males and 140,199 (56%) females aged 49\u0026ndash;77 years who were first-time participants of the HKCRCSP. Among them, 17,717 (16%) male and 14,868 (11%) female screenees tested FIT-positive and were referred for colonoscopy. Of these FIT-positive screenees, 84% had completed their colonoscopy with diagnosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Overall, 13,027 (12%) male and 9,761 (7.0%) female screenees were diagnosed with colorectal neoplasms (i.e., HP, NA, SL, AA, or CRC). Specifically, when considering advanced colorectal neoplasms (i.e., SL, AA, or CRC), the case counts and the proportions relative to the overall gender-specific population were: (i) 685 cases of SL (0.63%), 2,107 of AA (1.9%), and 1,068 of CRC (0.98%) in males; and (ii) 659 cases of SL (0.47%), 1,282 of AA (0.91%), and 678 of CRC (0.48%) in females (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). These statistics did not include (undiagnosed) colorectal neoplasms among individuals who were either FIT-negative or FIT-positive but had not yet completed their colonoscopy diagnosis.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe proportion of FIT-positive individuals increased with age in both genders and laboratories (Figure S2). However, when stratified by disease stage, FIT concentrations showed no significant association with age (Figure S5; Kruskal-Wallis test, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Given that FIT values tended to rise with the severity of neoplasms across both genders and laboratories (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), the age-related increase in FIT positivity likely reflects a higher prevalence of advanced colorectal neoplasms in older individuals, rather than a direct effect of age on FIT concentrations. In other words, the higher FIT positivity observed among older individuals simply reflected a shift in disease burden toward more advanced lesions. This interpretation is supported by our model-based estimates (Figure S9).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe fitted model was congruent with the data on stage-specific FIT distribution, and CRC incidence and mortality (Figure S10). Tables S3 and S4, Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e, and S12 summarize our estimates of the key epidemiologic parameters and the FIT value distributions. Particularly, we used the maximum two-sample FIT values (2-FIT-max) as the summary statistic of the empirical and estimated two-sample FIT distributions.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003ePrevalence and natural history.\u003c/em\u003e The estimated prevalence of colorectal neoplasms increased with age (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003e). By age 50 (the start age of HKCRCSP), the prevalence of colorectal neoplasms was 37.4% (95% credible interval [CrI]\u0026thinsp;=\u0026thinsp;35.6%-39.1%) in males and 27.4% (25.8%-29.1%) in females. Between the ages of 50 and 70, the prevalence of NA and HP increased substantially in both genders. This is mainly driven by (i) rapidly increasing incidence of NA and HP with age along both pathways, and (ii) slow annual progression rates (around 1%) of NA and HP before age 70 in both genders (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). The prevalence of advanced colorectal neoplasms also increased notably after age 50. At age 50, the prevalence of advanced colorectal neoplasms was 3.1% (2.9%-3.3%) in males and 2.2% (2.1%-2.4%) in females. By age 70, the prevalence of advanced colorectal neoplasms would have increased three- or four-fold compared with age 50, reaching 11% (11%-12%) in males and 7.8% (7.3%-8.2%) in females (Figure S14). Among those aged 70 with advanced colorectal neoplasms, 16% (15%-17%) of males and 12% (11%-14%) of females had CRC (Figure S15).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eDisease progression accelerated notably as lesions advanced along both pathways (Table S3). In the conventional pathway, the annual progression rate from AA to preclinical (undiagnosed) CRC was approximately 5\u0026ndash;6 times higher than that from NA to AA. In the serrated pathway, the progression rate from SL to preclinical CRC was 30%-60% higher than that from HP to SL. The annual progression rate to preclinical CRC was around 4% from AA and 1\u0026ndash;2% from SL (Table S3). The public health benefit of CRC screening can be clearly deduced from the age- and gender-specific prevalence and progression durations of preclinical CRC. These epidemiological parameters provide a quantitative basis for estimating opportunities for timely case detection and the potential mortality reduction through screening. At age 50, the prevalence of preclinical CRC was 0.40% (0.37%-0.42%) in males and 0.25% (0.23%-0.28%) in females, with an absolute increase of approximately 0.09% in males and 0.06% in females per year from this age onwards. For those aged 50, the mean durations of preclinical CRC stages 1\u0026ndash;4 were: (i) 1.16 (1.02\u0026ndash;1.32), 1.87 (1.66\u0026ndash;2.08), 4.78 (4.18\u0026ndash;5.37) and 3.64 (2.89\u0026ndash;4.70) years for males; and (ii) 1.43 (1.17\u0026ndash;1.78), 1.33 (1.10\u0026ndash;1.60), 4.45 (3.92\u0026ndash;5.14) and 3.28 (2.51\u0026ndash;4.17) years for females. Stage progression also accelerated with age. For example, the mean sojourn times for all stages at age 75 were approximately 15% shorter than at age 50 for both genders (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cem\u003eFIT test performance.\u003c/em\u003e Males with advanced colorectal neoplasms generally exhibited higher FIT values compared to females (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Consequently, the FIT-positivity proportion among those with advanced colorectal neoplasms was substantially higher in males (Figure S9). At a threshold of 100 ng/mL for CRC detection, FIT demonstrated high sensitivity in both genders: (i) 96% (96%-97%) for males and 88% (88%-90%) for females in laboratory A, and (ii) 97% (96%-98%) for males and 97% (95%-98%) for females in laboratory B. Specificity decreased with age because the prevalence of colorectal neoplasms among non-CRC cases increased with age. At age 50, specificity was (i) 93% (93%-93%) for males and 96% (96%-96%) for females in laboratory A, and (ii) 90% (89%-90%) and 92% (92%-92%), respectively, in laboratory B. Between the ages of 50 and 75, specificity decreased by approximately 0.4% per year. The negative predictive value (NPV) exceeded 99% for CRC across all ages and remained above 90% for advanced colorectal neoplasms up to age 75. In comparison, the positive predictive value (PPV) was notably lower: approximately 5% for CRC and 20% for advanced colorectal neoplasms in both genders at age 50, with an absolute increase of about 0.2% and 0.7% per year from this age onwards, respectively (Figure S17). As expected, decreasing the FIT positivity threshold improved sensitivity and NPV at the expense of reduced specificity and PPV (Figures S18 and S19).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e To our knowledge, our study is the first to characterize the local epidemiology and natural history of CRC using FIT data from a screening program. Our model integrates quantitative FIT results with local epidemiologic data to infer the gender- and age-specific prevalence of colorectal neoplasms. Such locally derived estimates are crucial, as CRC epidemiology varies widely across populations due to differences in demographics, genetics, lifestyle, and healthcare quality.\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e Our findings also show that screening test performance, including sensitivity, specificity, PPV, and NPV, can vary substantially depending on local epidemiology, screenee characteristics (e.g., gender), FIT-positivity threshold, and processing laboratory (Figures S17-S19). These results highlight the need for local evidence to guide screening policies. By combining local FIT and epidemiologic data, our study addresses this gap and provides a transferable framework.\u003c/p\u003e\u003cp\u003eOur findings suggested that a \u0026ldquo;one-size-fits-all approach\u0026rdquo; may not be optimal for CRC screening. For instance, we observed a higher prevalence of advanced colorectal neoplasms in males (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e4\u003c/span\u003e), which suggests that men should start screening earlier than women, consistent with Germany\u0026rsquo;s recommended start screening age of 45 and 50 for males and females, respectively.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e Moreover, our results suggest that at the same disease stage, females had lower FIT concentrations than males (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e), which resulted in comparatively lower sensitivity in females for any gender-agnostic threshold (Figure S17). We also found that the specificity of FIT decreased with age, indicating higher false-positive rates among older individuals. Taken together, adjusting FIT thresholds by gender and age can potentially improve equity and effectiveness. This is consistent with studies in Sweden, the Netherlands, and Denmark, which advocate age- and gender-specific thresholds to enhance efficiency.\u003csup\u003e\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eMore broadly, a data-driven approach for risk-based CRC prevention requires the incorporation of all relevant and available risk factors, including age, gender, and previous screening results. This capability enables the starting age to screen, screening intervals, and colonoscopy referrals to be tailored according to an individual\u0026rsquo;s estimated risk of advanced colorectal neoplasms. In particular, older individuals are associated with a higher risk of colorectal neoplasms,\u003csup\u003e21\u003c/sup\u003e along with higher FIT concentrations in prior FIT-negative tests, which are linked to an increased likelihood of subsequent colorectal neoplasm detection.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e Therefore, colonoscopy referrals might also be considered for older FIT-negative screenees whose FIT values are just marginally below the positivity threshold. For instance, in our study, we estimated that among FIT-negative screenees with 2-FIT-max values between 50 and 100 ng/ml, around 5%-15% had undiagnosed advanced colorectal neoplasms and 0.1%-0.5% had undiagnosed CRC (Figure S20), with the proportions increasing with higher FIT values and advancing ages.\u003c/p\u003e\u003cp\u003eOur estimated progression rates of CRC highlight the public health importance of CRC screening. We estimated that the average time from early-stage preclinical CRC (stage 1 or 2) to late-stage disease (stage 3 or 4) was approximately 3 to 4 years (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Given the marked difference in 5-year survival rates following clinical diagnosis\u0026mdash;96%, 87%, 69%, and only 9% in Hong Kong for stages 1 through 4, respectively\u0026mdash; the rapid progression of preclinical CRC underscores the narrow window for effective intervention before the disease advances.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e Without timely diagnosis, individuals with undetected early-stage preclinical CRC may quickly advance to late-stage disease, where treatment options are limited and outcomes are substantially worse. By quantifying these progression dynamics locally, our study offers a generalizable framework for optimizing the design and cost-effectiveness of screening.\u003c/p\u003e\u003cp\u003eOur study has several limitations. First, our study lacks information on other risk factors associated with CRC, such as family history of CRC, smoking history, body mass index, history of diabetes, and other predictors of advanced colorectal neoplasms (e.g., excessive alcohol use and physical inactivity).\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e Addressing this gap would require secure data linkage between the screening program (managed by the Department of Health) and the electronic health records of the screenees (managed by the Hospital Authority). Second, our epidemiology inferences regarding CRC are based on the HKCRCSP screenees and hence may not be fully representative of the general population. For example, cancer screening services might be less accessible to individuals from low socioeconomic and educational backgrounds, as well as ethnic minority groups, due to insufficient awareness and time constraints.\u003csup\u003e\u003cspan additionalcitationids=\"CR26\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e Third, the disease stages of FIT-negative screenees were unknown in our dataset because they did not undergo colonoscopy. Although their disease stage could be inferred within our framework, the robustness of the resulting estimates of disease prevalence and progression rates depends on the validity of our model assumptions (e.g., stage-specific FIT value distributions are independent of age, FIT values are stochastically larger at more advanced stages, etc.). Future colonoscopy-based studies that estimate disease prevalence among FIT-negative screenees could help validate our results. Finally, we did not distinguish between colon and rectal cancers, despite evidence of divergent local incidence trends and potentially heterogeneous etiologies and natural histories.\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e More granular data would be needed to calibrate a refined model that more accurately characterizes the epidemiology of CRC.\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eIn conclusion, quantitative FIT data from screening programs can be harnessed to characterize the epidemiology and natural history of CRC. Our novel, data-driven framework provides a robust and adaptable platform for evaluating and optimizing the cost-effectiveness of CRC screening programs.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003eStudy design and population\u003c/h2\u003e\u003cp\u003eLaunched as a pilot program in September 2016 and fully implemented in August 2018, the HKCRCSP provides biennial CRC screening to individuals aged 50-75.\u003csup\u003e15\u003c/sup\u003e Screenees are required to submit two fecal samples taken within 5 days apart. These sister samples were processed by laboratory A (using NS Prime, Alfresa Pharma Co., Ltd) before September 2019 and laboratory B (using OC-Sensor, Eiken Chemical Co., Ltd) thereafter. Screenees are regarded as FIT-positive if one or both of their FIT values are 100 ng/mL (equivalent to 20 \u0026micro;g/g) or above, and FIT-negative otherwise. FIT-positive screenees are referred for colonoscopy to confirm diagnosis, and FIT-negative screenees are invited to screen at a two-year interval (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Please see the Supplementary Appendix (SA, pp 2\u0026ndash;7) for more details about the HKCRCSP. According to the 2024 population in Hong Kong,\u003csup\u003e30\u003c/sup\u003e 8.6% of the eligible population has participated in the HKCRCSP.\u003c/p\u003e\u003cp\u003eWe utilized three data sources to estimate the epidemiology and natural history parameters of CRC and the FIT value distributions by disease stages: (i) records from first-time screenees of the HKCRCSP who enrolled between September 2016 and September 2021, including their age, two-sample FIT values, and colonoscopy diagnoses for those who tested FIT-positive and completed colonoscopy;\u003csup\u003e15\u003c/sup\u003e (ii) population-level CRC incidence and mortality statistics for 2012\u0026ndash;2016 from the Hong Kong Cancer Registry (HKCaR);\u003csup\u003e23\u003c/sup\u003e and (iii) all-cause mortality statistics for 2012\u0026ndash;2016 from the Hong Kong Census and Statistics Department.\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eIn this study, colonoscopy outcomes were defined as follows: (i) normal (N; no abnormality, no polyps, or only non-significant histology); (ii) hyperplastic polyps (HP; hyperplastic lesions\u0026thinsp;\u0026lt;\u0026thinsp;10 mm); (iii) non-advanced adenoma (NA; tubular adenomas\u0026thinsp;\u0026lt;\u0026thinsp;10 mm with low-grade dysplasia); (iv) serrated lesions (SL; sessile serrated lesions or traditional serrated adenomas of any size, or hyperplastic polyps\u0026thinsp;\u0026ge;\u0026thinsp;10 mm); (v) advanced adenoma (AA; tubular adenomas\u0026thinsp;\u0026ge;\u0026thinsp;10 mm or high-grade dysplasia, or adenomas with villous features); (vi) colorectal cancer (CRC; histologically confirmed adenocarcinoma); and (vii) unknown disease stage (Unk; those who were FIT-negative or FIT-positive without completing colonoscopy).\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e See Figure S4 for detailed definitions of colonoscopy findings. We assumed that, before the first screening, the underlying epidemiology and natural history of CRC among screenees reflected that of the general population.\u003c/p\u003e\u003cp\u003e This study was approved by the Institutional Review Board of The University of Hong Kong / Hospital Authority Hong Kong West Cluster (UW 17\u0026ndash;214).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eThe model\u003c/h3\u003e\n\u003cp\u003eOur data-driven model comprises two components: (1) a natural history model that simulates lifetime development and progression of CRC; and (2) a statistical model that characterizes the probability distribution of FIT values across different disease stages. Here, we briefly describe the two model components. Please see the SA (pp 8\u0026ndash;25) for more details.\u003c/p\u003e\u003cp\u003e\u003cem\u003eNatural history\u003c/em\u003e. We developed a gender-specific natural history model with two pathways of colorectal neoplasm (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb):\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e1. Conventional adenoma pathway: \u003cem\u003eN\u003c/em\u003e \u0026rarr; \u003cem\u003eNA\u003c/em\u003e \u0026rarr; \u003cem\u003eAA\u003c/em\u003e \u0026rarr; \u003cem\u003epCRC\u003c/em\u003e \u0026rarr; \u003cem\u003edCRC\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e2. Serrated pathway: \u003cem\u003eN\u003c/em\u003e \u0026rarr; \u003cem\u003eHP\u003c/em\u003e \u0026rarr; \u003cem\u003eSL\u003c/em\u003e \u0026rarr; \u003cem\u003epCRC\u003c/em\u003e \u0026rarr; \u003cem\u003edCRC\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cem\u003epCRC\u003c/em\u003e and \u003cem\u003edCRC\u003c/em\u003e are referred to as preclinical (undiagnosed) and diagnosed CRC, respectively. \u003cem\u003epCRC\u003c/em\u003e represents those individuals who have already developed CRC but remain undiagnosed (e.g., due to the absence of symptoms or delay in care-seeking). The prevalence of each disease stage \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:s\\)\u003c/span\u003e\u003c/span\u003e for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:s\\in\\:\\left\\{N,HP,\\:NA,SL,AA,CRC\\right\\}\\)\u003c/span\u003e\u003c/span\u003e at age 20 and the age-specific disease progression rates were subject to statistical inference (Table S3). Based on previous studies, we assumed that the serrated pathway accounts for 10\u0026ndash;30% of all CRC cases and inferred the corresponding proportion through model calibration.\u003csup\u003e\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e The model simulated the epidemiology of colorectal neoplasms from age 20 through death or up to age 100, whichever occurs first. Furthermore, individuals diagnosed with CRC experienced stage-specific relative survival as reported from local empirical data. Please see the SA (pp 8\u0026ndash;11) for more details.\u003c/p\u003e\u003cp\u003e\u003cem\u003eFIT value distribution\u003c/em\u003e. We modeled the probability distribution of FIT values across disease stages based on the patterns observed in the empirical screening data (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, S2, and S3). The screening data indicated no statistically significant association between FIT values and age within any disease stage (Figure S5). As such, we assumed that for each screenee, the FIT value of a given sample \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:V\\)\u003c/span\u003e\u003c/span\u003e depended on an underlying FIT value \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:U\\)\u003c/span\u003e\u003c/span\u003e such that:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e1. the mean and variance of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:U\\)\u003c/span\u003e\u003c/span\u003e are stage-specific;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e2. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:V=0\\)\u003c/span\u003e\u003c/span\u003e if \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:U=0\\)\u003c/span\u003e\u003c/span\u003e; and\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e3. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{ln}\\left(1+V\\right)=\\text{ln}\\left(1+U\\right)+\\epsilon\\:\\)\u003c/span\u003e\u003c/span\u003e if \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:U\u0026gt;0\\)\u003c/span\u003e\u003c/span\u003e, where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\epsilon\\:\\)\u003c/span\u003e\u003c/span\u003e is a normal random variable with mean 0 and standard deviation \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\omega\\:\\left(U\\right)\\)\u003c/span\u003e\u003c/span\u003e, i.e., \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\epsilon\\:\\sim\\:\\text{N}(0,\\omega\\:(U\\left)\\right)\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003ePlease refer to SA (pp 12\u0026ndash;17) for a full description of the statistical inference of the parameters for the FIT value distributions.\u003c/p\u003e\u003cp\u003e\u003cem\u003eParameter estimation\u003c/em\u003e. We used \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\theta\\:\\)\u003c/span\u003e\u003c/span\u003e to denote the set of model parameters subject to statistical inference (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Tables S3 and S4). Let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:p\\left(s\\right|a,\\theta\\:)\\)\u003c/span\u003e\u003c/span\u003e be the prevalence of disease stage \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:s\\in\\:\\{N,HP,\\:NA,SL,AA,CRC\\}\\)\u003c/span\u003e\u003c/span\u003e among individuals of age \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:a\\)\u003c/span\u003e\u003c/span\u003e. Let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:f\\left({v}_{1},{v}_{2}|s,l,\\theta\\:\\right)\\)\u003c/span\u003e\u003c/span\u003e be the probability density function of the two-sample FIT values (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{v}_{1},{v}_{2})\\)\u003c/span\u003e\u003c/span\u003e for individuals with disease stage \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:s\\)\u003c/span\u003e\u003c/span\u003e and processed by laboratory \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:l\\)\u003c/span\u003e\u003c/span\u003e. Let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\lambda\\:}_{xj}^{HKCaR}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\lambda\\:}_{xj}^{model}\\)\u003c/span\u003e\u003c/span\u003e be the incidence of CRC stage \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:x\\)\u003c/span\u003e\u003c/span\u003e (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:x\\)\u003c/span\u003e\u003c/span\u003e = 1, 2, 3, 4, and \u0026ldquo;unstaged\u0026rdquo;) in the \u003cem\u003ej\u003c/em\u003e\u003csup\u003eth\u003c/sup\u003e age group in the HKCaR database and our model, respectively. Similarly, let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{j}^{HKCaR}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{j}^{model}\\)\u003c/span\u003e\u003c/span\u003e be the corresponding CRC mortality statistics. We assumed that annual CRC incidence and mortality follow Poisson distributions. Let \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}^{+}\\)\u003c/span\u003e\u003c/span\u003e be the group of FIT-positive screenees who had received their colonoscopy diagnosis and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}^{-}\\)\u003c/span\u003e\u003c/span\u003e be the remaining screenees (who were either FIT-negative or FIT-positive but had not yet completed their colonoscopy diagnosis).\u003c/p\u003e\u003cp\u003eThe likelihood function was:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:L\\left(\\theta\\:\\right)={L}_{{C}^{+}}\\left(\\theta\\:\\right)\\times\\:{L}_{{C}^{-}}\\left(\\theta\\:\\right)\\times\\:\\underset{\\text{CRC\\:incidence}}{\\underset{⏟}{\\prod\\:_{x}^{}{\\prod\\:}_{j}Poisson({\\lambda\\:}_{xj}^{HKCaR},{\\lambda\\:}_{xj}^{model})}}\\times\\:\\underset{\\text{CRC\\:mortality}}{\\underset{⏟}{{\\prod\\:}_{j}Poisson({z}_{j}^{HKCaR},{z}_{j}^{model})}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:{L}_{{C}^{+}}\\left(\\theta\\:\\right)=\\prod\\:_{i\\in\\:{C}^{+}}p\\left({s}_{i}\\right|{a}_{i},\\theta\\:\\left)f\\right({v}_{i1},{v}_{i2}|{s}_{i},{l}_{i},\\theta\\:)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ecorresponded to screenees in group \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}^{+}\\)\u003c/span\u003e\u003c/span\u003e and\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:{L}_{{C}^{-}}\\left(\\theta\\:\\right)=\\prod\\:_{i\\in\\:{C}^{-}}\\sum\\:_{s\\in\\:\\{N,HP,NA,SL,AA,CRC\\}}p\\left(s\\right|{a}_{i},\\theta\\:\\left)f\\right({v}_{i1},{v}_{i2}|s,{l}_{i},\\theta\\:)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ecorresponded to screenees in group \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}^{-}\\)\u003c/span\u003e\u003c/span\u003e. Note that although screenees in group \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}^{-}\\)\u003c/span\u003e\u003c/span\u003e (primarily FIT-negative) did not have colonoscopy outcomes, their FIT values contained information on the likelihood of their disease stages. The likelihood component \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{L}_{{C}^{-}}\\left(\\theta\\:\\right)\\)\u003c/span\u003e\u003c/span\u003e encapsulated such information where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:s\\)\u003c/span\u003e\u003c/span\u003e is a dummy variable denoting the possible disease stages. We estimated \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\theta\\:\\)\u003c/span\u003e\u003c/span\u003e by fitting the model to data using Markov chain Monte Carlo (MCMC) methods. Please see the SA (pp 18\u0026ndash;25) for details on parameter inference and the assessment of model identifiability.\u003c/p\u003e\n\u003ch3\u003eData availability\u003c/h3\u003e\n\u003cp\u003eThe data that support the findings of this study are available from the Hong Kong Hospital Authority, but restrictions apply to the availability of these data, which were used only for the current study, and so are not publicly available.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eCode availability\u003c/h2\u003e\u003cp\u003eCode supporting the findings of this study is available upon request.\u003c/p\u003e\u003c/div\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eAA\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eadvanced adenoma\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eCRC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ecolorectal cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eCLN\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ecolonoscopy\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eCrI\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ecredible interval\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003e\u003cem\u003ed\u003c/em\u003eCRC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ediagnosed colorectal cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eFIT\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003efecal immunochemical test\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eFIT-neg\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eFIT-negative\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eFIT-pos\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eFIT-positive\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eHKCRCSP\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eHong Kong Colorectal Cancer Screening Programme\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eHP\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ehyperplastic polyps\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eN\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003enormal\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eNA\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003enon-advanced adenoma\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003e\u003cem\u003ep\u003c/em\u003eCRC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003epreclinical colorectal cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eSL\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eserrated lesions\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eUnk\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eunknown disease stage\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003e2-FIT-max\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003emaximum two-sample FIT values.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eWe confirmed that informed consent was obtained from all participants of the Hong Kong Colorectal Cancer Screening Programme by the Hong Kong Department of Health, and the need for consent to conduct this research with the anonymised data from participants was approved by the Institutional Review Board of The University of Hong Kong / Hospital Authority Hong Kong West Cluster (Ref: UW 17-214).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHealth and Medical Research Fund (HMRF), Hong Kong Special Administrative Region, China. The funding bodies had no role in the design of the study or in the analysis and interpretation of data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCorrespondence\u003c/strong\u003e and requests for materials should be addressed to Joseph T. Wu and Kathy Leung.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors report no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWriting assistance:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo writing assistance was provided in the preparation of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement of authorship contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJTW, WKL, WLL, GML, and KL designed the study, guided the writing of the manuscript, and reviewed the manuscript. KL cleaned and managed data. JTW initially developed the model, performed the analyses, and drafted the manuscript. ZW and JTW further refined the analyses. ZW, JTW, HCWC, and KL extended the analyses and wrote the manuscript. JTW, KL, ZW, and HCWC have directly accessed and verified the underlying data. All authors interpreted the data and revised the manuscript. All authors read and approved of the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBray F et al (2024) Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3322/caac.21834\u003c/span\u003e\u003cspan address=\"10.3322/caac.21834\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNavarro M, Nicolas A, Ferrandez A, Lanas A (2017) Colorectal cancer population screening programs worldwide in 2016: An update. World J Gastroenterol 23:3632\u0026ndash;3642. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3748/wjg.v23.i20.3632\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v23.i20.3632\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLevin TR et al (2018) Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large Community-Based Population. \u003cem\u003eGastroenterology\u003c/em\u003e 155, 1383\u0026ndash;1391.e1385. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2018.07.017\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2018.07.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCardoso R, Guo F, Heisser T, Hoffmeister M, Brenner H (2020) Utilisation of Colorectal Cancer Screening Tests in European Countries by Type of Screening Offer: Results from the European Health Interview Survey. Cancers (Basel) 12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/cancers12061409\u003c/span\u003e\u003cspan address=\"10.3390/cancers12061409\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSung JJ et al (2015) An updated Asia Pacific Consensus Recommendations on colorectal cancer screening. Gut 64:121\u0026ndash;132. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2013-306503\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2013-306503\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eU. S. Preventive Services Task Force et al. Screening for Colorectal Cancer: US Preventive Services Task Force Recommendation Statement. \u003cem\u003eJAMA\u003c/em\u003e 325, 1965\u0026ndash;1977 (2021) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jama.2021.6238\u003c/span\u003e\u003cspan address=\"10.1001/jama.2021.6238\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchreuders EH et al (2015) Colorectal cancer screening: a global overview of existing programmes. Gut 64:1637\u0026ndash;1649. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2014-309086\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2014-309086\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYoung GP, Rabeneck L, Winawer SJ (2019) The Global Paradigm Shift in Screening for Colorectal Cancer. \u003cem\u003eGastroenterology\u003c/em\u003e 156, 843\u0026ndash;851 e842 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2019.02.006\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2019.02.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim SY, Kim H-S, Park HJ (2019) Adverse events related to colonoscopy: Global trends and future challenges. World J Gastroenterol 25:190\u0026ndash;204. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3748/wjg.v25.i2.190\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v25.i2.190\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRobertson DJ, Rex DK, Ciani O, Drummond MF (2024) Colonoscopy vs the Fecal Immunochemical Test: Which is Best? \u003cem\u003eGastroenterology\u003c/em\u003e 166, 758\u0026ndash;771 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2023.12.027\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2023.12.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCastells A et al (2025) Effect of invitation to colonoscopy versus faecal immunochemical test screening on colorectal cancer mortality (COLONPREV): a pragmatic, randomised, controlled, non-inferiority trial. Lancet 405:1231\u0026ndash;1239. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0140-6736(25)00145-X\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(25)00145-X\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShaukat A, Levin TR (2022) Current and future colorectal cancer screening strategies. Nat Rev Gastroenterol Hepatol 19:521\u0026ndash;531. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41575-022-00612-y\u003c/span\u003e\u003cspan address=\"10.1038/s41575-022-00612-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSenore C et al (2020) Faecal haemoglobin concentration among subjects with negative FIT results is associated with the detection rate of neoplasia at subsequent rounds: a prospective study in the context of population based screening programmes in Italy. Gut 69:523\u0026ndash;530. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2018-318198\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2018-318198\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDigby J et al (2025) Combining faecal haemoglobin, iron deficiency anaemia status and age can improve colorectal cancer risk prediction in patients attending primary care with bowel symptoms: a retrospective observational study. Gut. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2024-334248\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2024-334248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHong Kong Colorectal Cancer Screening Program \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.colonscreen.gov.hk/en/public/index.html\u003c/span\u003e\u003cspan address=\"https://www.colonscreen.gov.hk/en/public/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCollaborators GBDCC (2022) Global, regional, and national burden of colorectal cancer and its risk factors, 1990\u0026ndash;2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Gastroenterol Hepatol 7:627\u0026ndash;647. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S2468-1253(22)00044-9\u003c/span\u003e\u003cspan address=\"10.1016/S2468-1253(22)00044-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSergeev D, Heisser T, Hoffmeister M, Brenner H (2025) Potential for enhancing efficacy of screening colonoscopy by lowering starting ages and extending screening intervals: A modelling study for Germany. Int J Cancer 156:2303\u0026ndash;2310. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ijc.35322\u003c/span\u003e\u003cspan address=\"10.1002/ijc.35322\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBlom J et al (2019) Gender-specific cut-offs in colorectal cancer screening with FIT: Increased compliance and equal positivity rate. J Med Screen 26:92\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0969141318804843\u003c/span\u003e\u003cspan address=\"10.1177/0969141318804843\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNjor SH, Rasmussen M, Friis-Hansen L, Andersen B (2022) Varying fecal immunochemical test screening cutoffs by age and gender: a way to increase detection rates and reduce the number of colonoscopies. Gastrointest Endosc 95:540\u0026ndash;549. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.gie.2021.09.038\u003c/span\u003e\u003cspan address=\"10.1016/j.gie.2021.09.038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHarlass M et al (2025) Benefits of colorectal cancer screening using FIT with varying positivity thresholds by age and sex. JNCI: J Natl Cancer Inst. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/jnci/djaf149\u003c/span\u003e\u003cspan address=\"10.1093/jnci/djaf149\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKastrinos F, Kupfer SS, Gupta S (2023) Colorectal Cancer Risk Assessment and Precision Approaches to Screening: Brave New World or Worlds Apart? \u003cem\u003eGastroenterology\u003c/em\u003e 164, 812\u0026ndash;827 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2023.02.021\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2023.02.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan den Berg DMN, van den Puttelaar R, de Jonge L, Lansdorp-Vogelaar I, Toes-Zoutendijk E (2025) Fecal Hemoglobin Levels in Prior Negative Screening and Detection of Colorectal Neoplasia: A Dose-Response Meta-Analysis. Gastroenterology 168:587\u0026ndash;597. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2024.10.047\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2024.10.047\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHong Kong Cancer Registry \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www3.ha.org.hk/cancereg/hkcar.html\u003c/span\u003e\u003cspan address=\"https://www3.ha.org.hk/cancereg/hkcar.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWong MC et al (2014) A validated tool to predict colorectal neoplasia and inform screening choice for asymptomatic subjects. Gut 63:1130\u0026ndash;1136\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoss SM et al (2012) Performance measures in three rounds of the English bowel cancer screening pilot. Gut 61:101\u0026ndash;107. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gut.2010.236430\u003c/span\u003e\u003cspan address=\"10.1136/gut.2010.236430\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBozhar H et al (2022) Socio-economic inequality of utilization of cancer testing in Europe: A cross-sectional study. Prev Med Rep 26:101733. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pmedr.2022.101733\u003c/span\u003e\u003cspan address=\"10.1016/j.pmedr.2022.101733\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCarethers JM, Doubeni CA (2020) Causes of Socioeconomic Disparities in Colorectal Cancer and Intervention Framework and Strategies. Gastroenterology 158:354\u0026ndash;367. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2019.10.029\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2019.10.029\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang B, Xie SH, Yu IT (2018) Differential incidence trends of colon and rectal cancers in Hong Kong: an age-period-cohort analysis. Cancer Commun (Lond) 38:42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40880-018-0311-2\u003c/span\u003e\u003cspan address=\"10.1186/s40880-018-0311-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDeYoreo M, Rutter CM, Ozik J, Collier N (2022) Sequentially calibrating a Bayesian microsimulation model to incorporate new information and assumptions. BMC Med Inf Decis Mak 22:12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12911-021-01726-0\u003c/span\u003e\u003cspan address=\"10.1186/s12911-021-01726-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHong Kong Census and Statistics Department \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.censtatd.gov.hk/home/index.jsp\u003c/span\u003e\u003cspan address=\"https://www.censtatd.gov.hk/home/index.jsp\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLokuhetty D, Organization WH, Cancer IA (2019) f. R. o. WHO Classification of Tumours of the Digestive System Tumours. IARC\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEast JE et al (2017) British Society of Gastroenterology position statement on serrated polyps in the colon and rectum. Gut 66:1181\u0026ndash;1196. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2017-314005\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2017-314005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSzylberg L, Janiczek M, Popiel A, Marszalek A (2015) Serrated Polyps and Their Alternative Pathway to the Colorectal Cancer: A Systematic Review. \u003cem\u003eGastroent Res Pract\u003c/em\u003e 2015 https://doi.org/Artn 573814\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e1155/2015/573814\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEast JE, Vieth M, Rex DK (2015) Serrated lesions in colorectal cancer screening: detection, resection, pathology and surveillance. Gut 64:991\u0026ndash;1000. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/gutjnl-2014-309041\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2014-309041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8110710/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8110710/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eColorectal cancer (CRC) epidemiology remains insufficiently characterized in many settings, limiting optimal prevention strategies. Using quantitative fecal immunochemical test (FIT) data from 248,692 first-time participants (aged 49–77; 56% female) in the Hong Kong CRC Screening Programme, we developed a natural history model incorporating adenoma and serrated pathways with stage-specific FIT distributions. Colonoscopy referral was triggered if either of two submitted samples exceeded 100 ng/mL (13% positivity).\u003c/p\u003e\n\u003cp\u003eWe estimated that 37% (95% credible interval = 36–39%) of males and 27% (26–29%) of females had colorectal neoplasms at age 50; ~8% had advanced colorectal neoplasms (advanced adenoma, serrated lesions, or CRC). Prevalence of advanced neoplasms increased ~ 0.5% per year after age 50. Annual progression to CRC was ~ 4% for advanced adenoma and 1–2% for serrated lesions. Preclinical CRC advanced from stages I-II to III-IV within 3–4 years. At the 100 ng/mL threshold, FIT demonstrated 88–97% sensitivity for CRC. The positive predictive value for advanced neoplasms rose from ~ 20% at age 50 by ~ 1% annually, while the negative predictive value remained \u0026gt; 90%. Males with advanced neoplasms had higher FIT values than females.\u003c/p\u003e\n\u003cp\u003eQuantitative FIT data thus enables robust characterization of CRC epidemiology and progression, providing a foundation for evaluating screening strategies and cost-effectiveness.\u003c/p\u003e","manuscriptTitle":"Characterizing the epidemiology and natural history of colorectal cancer using fecal immunochemical test data from screening programs: a modelling study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-03 00:48:52","doi":"10.21203/rs.3.rs-8110710/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"fe28d259-4175-4193-837d-980c93e9e20c","owner":[],"postedDate":"December 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":58407691,"name":"Health sciences/Gastroenterology/Gastrointestinal diseases/Gastrointestinal cancer/Colorectal cancer"},{"id":58407692,"name":"Health sciences/Oncology/Cancer/Cancer screening"},{"id":58407693,"name":"Health sciences/Oncology/Cancer/Cancer prevention"},{"id":58407694,"name":"Health sciences/Health care/Public health/Population screening"},{"id":58407695,"name":"Health sciences/Medical research/Epidemiology"}],"tags":[],"updatedAt":"2026-02-05T12:06:37+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-03 00:48:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8110710","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8110710","identity":"rs-8110710","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}