Interval Colorectal Cancers in a Fecal Immunochemical Test-Based Screening Program.

IMPORTANCE

Interval colorectal cancers (CRCs) can occur after a negative result on a primary screening test or after a colonoscopy for a positive screening test result without the diagnosis of CRC.

OBJECTIVE

To compare the burden, incidence, mortality, and long-term survival of interval CRC after a fecal immunochemical test (post-FIT interval CRC) and colonoscopy (postcolonoscopy interval CRC) in the FIT-based Taiwan CRC Screening Program.

DESIGN, SETTING, AND PARTICIPANTS: This cohort study analyzed participants (aged 50-74 years) in the Taiwan CRC Screening Program who completed at least 1 FIT between January 1, 2004, and December 31, 2012. Two cohorts were created: one including individuals with a negative FIT result, and the other comprising individuals with a positive FIT result followed by a negative colonoscopy result. Post-FIT interval CRC or postcolonoscopy interval CRC and their survival status were identified and verified via linkage to Taiwan Cancer Registry and Taiwan Death Registry through December 31, 2019. Data analysis was performed between January 2004 and December 2019.

EXPOSURES

Participants were screened using FIT, with follow-up colonoscopy offered to individuals who tested positive. The main exposures of interest were (1) negative FIT vs positive FIT result with a negative colonoscopy result, and (2) negative FIT and positive FIT results with a negative colonoscopy result further stratified by hospital-level adenoma detection rate (ADR). Low was defined as less than 40% ADR, middle was between 40% and less than 65% ADR, and high was 65% or higher ADR.

MAIN OUTCOMES AND MEASURES

The primary outcomes were the incidence, CRC-specific mortality, and survival rates for post-FIT interval CRC and postcolonoscopy interval CRC.

RESULTS

A total of 4018 patients with interval CRC (mean [SD] age at diagnosis, 62.8 [5.9] years; 2118 males [52.7%]) were included in the analysis. Among 15 386 total CRCs in the program, 10 515 (68.3%) were screen-detected CRCs and 4018 (26.2%) were interval CRCs. Of these interval CRC cases, 2782 (18.1%) were post-FIT interval CRCs and 1236 (8.1%) were postcolonoscopy interval CRCs. Compared with post-FIT interval CRC detected after a negative FIT result, postcolonoscopy interval CRC incidence (0.75 [95% CI, 0.71-0.79] per 1000 person-years vs 0.09 [95% CI, 0.09-0.10] per 1000 person-years) and mortality (0.12 [95% CI, 0.11-0.14] per 1000 person-years vs 0.02 [95% CI, 0.02-0.03] per 1000 person-years) after follow-up colonoscopy were higher, with adjusted hazard ratios (AHRs) of 7.06 (95% CI, 6.35-7.57) and 5.04 (95% CI, 4.33-5.85), respectively. Hospital-level ADR was inversely associated with postcolonoscopy interval CRC incidence (high- vs low-ADR group AHR, 0.26; 95% CI, 0.20-0.36) and mortality (AHR, 0.28; 95% CI, 0.19-0.41). However, among patients with postcolonoscopy interval CRC, CRC-specific death was higher in the high-ADR vs low-ADR setting (AHR, 1.89; 95% CI, 1.04-3.43).

CONCLUSIONS AND RELEVANCE

This cohort study found that interval CRC incidence and mortality were higher after follow-up colonoscopy than after negative FIT result, which suggests that individuals with a positive FIT result undergoing a colonoscopy exhibit a much higher risk for CRC. Survival after the diagnosis of CRC in the high-ADR subgroup was worse, suggesting a ceiling effect for current colonoscopic techniques and highlighting a possible interplay between procedural and biological factors.