Fecal immunochemical test surveillance in colorectal cancer following adenoma resection: A longitudinal, population-based linked cohort study in China

Abstract

Background: Although the fecal immunochemical test (FIT) is widely utilized in colorectal cancer (CRC) screening because of its noninvasive, rapid, and cost-effective characteristics, its effectiveness in post-adenoma resection surveillance remains unclear. This study aims to evaluate the benefits of follow-up FIT surveillance in individuals with adenoma resection and to identify risk factors associated with adenoma recurrence.

Methods and findings: As part of China's National Screening Project, we identified a total of 5,911 individuals who underwent adenoma removal during the first round of CRC screening in Jiashan and Haining between 2006 and 2021. All individuals with adenoma removal were invited to participate in a second CRC screening; 2,448 accepted and chose either direct colonoscopy surveillance (n = 989) or FIT surveillance (n = 1,459), while 3,463 declined. The Clone-Censor-Weight method was applied to mitigate time-related biases. Cox proportional hazards and Poisson regression models were used to evaluate the benefits of follow-up surveillance strategies after adenoma resection, adjusting for age, sex, baseline adenoma grade, family history of CRC in first-degree relatives, symptoms, chronic appendicitis or cholecystitis, and stressful life events. Additionally, we examined the risk factors associated with adenoma recurrence using logistic regression. The outcomes were the long-term incidence of CRC and the recurrence of adenomas following adenoma resection. Over average follow-up of 7.79 and 7.46 years, participants who underwent protocol-adherent follow-up FIT surveillance had a 44% lower CRC risk (hazard ratio [HR] = 0.56, 95% confidence interval [CI]: 0.31, 0.98; p = 0.044), and those who underwent direct colonoscopy had a 51% lower risk (HR = 0.49, 95% CI [0.27, 0.89]; p = 0.019), compared to individuals who refused follow-up surveillance. Compared with the direct follow-up colonoscopy group (53.56 per 100,000 person-years), the long-term CRC incidence rates were 70.38 for the follow-up negative FIT group and 80.14 for the positive FIT with adherence to colonoscopy group, with no statistically significant differences (p = 0.852; p = 0.834). Notably, participants who did not undergo colonoscopy following a positive FIT had a significantly increased CRC risk compared to those in the direct follow-up colonoscopy group, with an adjusted incidence rate ratio (aIRR) of 6.64 (95% CI [1.11, 39.83]; p = 0.038). Alcohol consumption (nondrinkers versus >3 times per week: adjusted odds ratio [aOR] = 0.43, 95% CI [0.27, 0.69]; p < 0.001) was associated with adenoma recurrence. Moreover, smoking (current smokers versus nonsmokers: aOR = 3.72, 95% CI [1.19, 11.60]; p = 0.024), obesity (obese versus normal: aOR = 3.21, 95% CI [1.17, 8.80]; p = 0.023), and having advanced adenomas at baseline (aOR = 3.30, 95% CI [1.41, 7.69]; p = 0.006) were associated with recurrence of advanced adenomas. Given the limited number of incident CRC cases and the observational study design, conclusions regarding the impact of follow-up FIT surveillance after adenoma removal should be interpreted with caution.

Conclusion: Protocol-adherent follow-up FIT surveillance after adenoma removal was associated with reduced long-term CRC risk, comparable to that observed with direct colonoscopy. However, improving adherence to colonoscopy after a positive FIT surveillance is crucial.

Fig 1. Overview of the colorectal cancer screening program in Jiashan and Haining, Zhejiang province.

The red arrow represents the adenoma removal population included in this study. ^a, all individuals who had adenomas removed will be invited for the second-round CRC screening, but they have the right to decline; ^b, all individuals undergoing second-round CRC screening can choose either direct colonoscopy or sequential screening based on FIT surveillance; ^c, these individuals underwent a colonoscopy without FIT surveillance. Abbreviation: CRC, Colorectal Cancer; FIT, fecal immunochemical test.

Fig 2. Flow diagram of included participants in the study.

Abnormal colonoscopy findings included intestinal inflammation, bleeding, ulceration, adenomas, nonadenomatous polyps (such as hyperplastic, inflammatory, and serrated polyps), CRC, etc. ^#, five CRC cases detected during colonoscopy were excluded. Abbreviations: FIT, fecal immunochemical test; CRC, colorectal cancer; PY, person-years.

Fig 3. Comparison of CRC incidence between adherence to the complete process of follow-up FIT surveillance (A) and direct colonoscopy (B) vs. nonfollow-up surveillance groups among adenoma removal population based on the Clone-Censor-Weight method.

HRs were calculated based on Cox regression. The covariates included in the Clone-Censor-Weight process were age, sex, adenoma grade at baseline colonoscopy, family history of CRC in first-degree relatives, symptoms, chronic appendicitis or cholecystitis, and stressful life events. Abbreviations: IPW, inverse probability weighting; FIT, fecal immunochemical test; CRC, colorectal cancer; HR, hazard ratio; CI, confidence interval.

Fig 4. Comparison of CRC incidence among different follow-up surveillance groups following adenoma removal, utilizing Kaplan–Meier analysis (A) and Poisson regression (B).

The Kaplan–Meier cumulative hazard estimates for CRC were unadjusted for covariates. The aIRR in Model 1 was estimated using multivariable Poisson regression, with adjustments for age, sex, family history of CRC in first-degree relatives, and accounting for follow-up time. Model 2 extended this adjustment by additionally controlling for clinical symptom, chronic appendicitis or cholecystitis, and stressful life events. The percentages above Fig 4A indicate the cumulative proportion of incident CRC cases at each time point. Abbreviations: FIT, fecal immunochemical test; CRC, colorectal cancer; IRR, incidence rate ratio; aIRR, adjusted incidence rate ratio; CI, confidence interval.