Leveraging biomarkers and primary care embedding for scalable precision cancer prevention in China: Insights from the FuSion study
From population to individual cancer prevention
Cancer is a leading cause of death in China, with an estimated 2.57 million deaths in 2022. Despite national screening efforts, most strategies remain age-based or rely on non-specific indicators such as lifestyle or regional incidence, offering limited biological specificity and insufficiently accounting for population heterogeneity. Precision prevention integrates clinical, behavioral, molecular, and environmental data to enable biologically informed risk stratification and targeted interventions,2 but scalable deployment requires platforms that link discovery, validation, and primary-care implementation. Despite progress from UK Biobank and Biobank Japan, few systems—globally or in China—combine long-term follow-up, deep molecular profiling, and clinical integration for prevention at scale.
To address these gaps and support the practical deployment of precision cancer prevention in China, we initiated the FuSion (Fudan-Singlera for Cancer Early Detection) study in 2021 in Taizhou, Jiangsu Province. Building upon the Taizhou Longitudinal Study (TZL), FuSion has enrolled 42,666 adults aged 40–74 with more than a decade of follow-up. It integrates deep phenotyping, clinical screening, and multi-level analytics to support a three-stage implementation framework: (1) biomarker discovery and validation, (2) risk model deployment in community screening, and (3) translation into public health interventions. FuSion serves not only as a research cohort but also as a delivery-oriented platform designed to meet the practical demands of precision, affordability, and equity in cancer prevention.
Routine biomarkers for scalable risk stratification
Multi-omics and liquid biopsy offer promise for early detection but are costly and largely inaccessible for large-scale use, especially in resource-limited settings. Routine biomarkers are already embedded in primary care and could enable low-cost population stratification, yet their panel-level predictive utility remains underexplored.
FuSion evaluated 54 routinely measured biomarkers, including hepatic, metabolic, endocrine, inflammatory, and tumor-related markers, across the cohort. By 2022, over 1,900 newly diagnosed cancer cases enabled robust site-specific signals, informing a stratification model now under validation. In the first-round screening, the highest-risk tier showed three- to seventeen-fold higher detection of precancerous lesions and early-stage cancers, underscoring translational potential.
To mitigate false positives, overdiagnosis, and limited actionability, FuSion uses decision-curve analysis to set thresholds under capacity and equity constraints, and microsimulation/budget-impact modeling to compare screening bundles (questionnaires + biomarkers, with selective imaging as confirmatory testing) over a 5-year horizon. This framework informs protocol design and recalibration, while establishing a pragmatic pathway to selectively incorporate multi-omics and liquid biopsy–based multi-cancer early detection (MCED) tools in the future.
