Frontiers in Medicine · Oct 30, 2026
Evolution of colorectal cancer screening technologies and health economic evaluation: optimization strategies based on an economic gradient and dynamic risk stratification
By Xingyu Yin, Yang Qi, Kejin Zhu et al.
Colorectal cancer (CRC) remains a major contributor to the global cancer burden, with approximately 555,000 new cases and 286,000 deaths annually in China, where most patients are still diagnosed at intermediate or advanced stages. This narrative review synthesizes recent evidence on the evolution of CRC screening technologies and their health economic evaluation, with the aim of informing resource-adapted screening strategies for China and comparable settings. PubMed, Embase, Web of Science, the Cochrane Library, and three major Chinese databases (CNKI, Wanfang Data, and VIP) were searched from database inception to 7 November 2025, with emphasis on studies published within the past five years. The evidence supports three principal findings. First, noninvasive molecular testing has substantially improved diagnostic performance: next-generation multitarget stool DNA testing achieves a sensitivity of 93.9% for CRC (versus 67.3% for the fecal immunochemical test [FIT] in the same head-to-head cohort), and stool syndecan-2 methylation testing outperforms FIT for advanced colorectal neoplasia, although sensitivity for advanced precancerous lesions remains below 50% for all currently available noninvasive assays. Second, health economic analyses consistently identify FIT as the most cost-effective primary screening tool under real-world adherence conditions, whereas structured risk-stratified and opportunistic screening models achieve higher benefit-cost ratios (9.29 for opportunistic screening versus 6.43 for population-based screening in a representative Chinese community program, although these ratios derive from a single community program, are subject to selection bias, and are not incremental comparisons). Third, artificial intelligence (AI)-assisted colonoscopy—specifically computer-aided detection (CADe), as distinct from computer-aided diagnosis (CADx)—reduces adenoma miss rates in randomized tandem studies; in a single-center diagnostic study, combined AI and conventional colonoscopy showed high agreement with histopathology, although this finding derives from an enriched symptomatic cohort rather than an average-risk screening population. Most pivotal trials of emerging molecular assays were industry-sponsored, which may inflate reported performance. Building on this evidence, we propose a conceptual dual-cycle optimization framework with illustrative, author-proposed thresholds: an outer cycle that matches the population coverage of emerging high-cost technologies to regional per-capita gross domestic product and budget constraints, and an inner cycle that recalibrates individual risk annually through a four-dimensional assessment system integrating the Asia-Pacific Colorectal Screening score, serial molecular monitoring, AI-derived imaging features, and genetic susceptibility. Neither a single highly sensitive test nor a low-cost primary screening tool alone can simultaneously maximize early detection and contain health expenditure; tiered, risk-stratified, and economically graded strategies are required to reduce the CRC burden in resource-heterogeneous settings. The framework is hypothesis-generating, and its thresholds require prospective calibration before implementation.