Impact of the serrated pathwayon the simulated comparative effectiveness of colorectal cancer screening tests.

Abstract

Colorectal cancers (CRCs) arise from adenomas, which can produce fecal occult blood and can be detected endoscopically, or sessile serrated lesions (SSLs), which rarely bleed and may be more challenging to detect. Models informing CRC screening policy should reflect both pathways, accounting for uncertainty. Novel decision-analytic model of the adenoma and serrated pathways for CRC (ANSER) to compare current and emerging screening strategies, accounting for differential test sensitivities for adenomas and SSLs, and uncertainty. Strategies included colonoscopy every 10 years, stool-DNA/FIT (sDNA-FIT) every 1-3 years, or fecal immunochemical testing (FIT) every year from age 45 to 75 years. Outcomes included CRC cases and deaths, cost-effectiveness (cost/quality-adjusted life-year [QALY] gained), and burden-benefit (colonoscopies/life-year gained), with 95% uncertainty intervals (UIs). ANSER predicted 62.5 (95% UI = 58.8-66.3) lifetime CRC cases and 24.1 (95% UI = 22.5-25.7) CRC deaths/1000 45-year-olds without screening, and 78%-87% CRC mortality reductions with screening. The tests' outcome distributions overlapped for QALYs gained but separated for required colonoscopies and costs. All strategies cost less than $100 000/QALY gained vs no screening. Colonoscopy was the most effective and cost-effective, costing $9300/life-year gained (95% UI = $500-$21 900) vs FIT. sDNA-FIT cost more than $500 000/QALY gained vs FIT. As more CRCs arose from SSLs, colonoscopy remained preferred based on clinical benefit and cost-effectiveness, but cost-effectiveness improved for a next-generation sDNA-FIT. When the serrated pathway is considered, modeling suggests that colonoscopy is cost-effective vs FIT. In contrast, modeling suggests that sDNA-FIT is not cost-effective vs FIT despite its greater sensitivity for SSLs, even if a substantial minority of CRCs arise from SSLs.