Abstract / Summary
Background: Cardiovascular magnetic resonance (CMR) provides accurate, reproducible quantification of cardiac chamber size and function. Population-specific reference ranges, standardized categories for grading degree of deviation, and evidence for their prognostic relevance are needed to enhance clinical utility. Using the Healthy Hearts Consortium (HHC), we derived centile-based reference ranges and deviation categories for ventricular and atrial CMR measures across multiple demographic groups and two software platforms, and evaluated their prognostic relevance in a large longitudinal cohort. Methods: We analyzed 9,059 CMR studies from the HHC, comprising healthy adults aged 18-83 years from six international cohorts. Images were processed automatically with cvi42 (Circle Cardiovascular Imaging) and suiteHEART (NeoSoft) to derive left and right ventricular volumes, ejection fraction and mass, and atrial volumes and function. Age- and sex-specific centile curves were estimated using generalized additive models for location, scale, and shape. Reference ranges were defined by the 2.5th-97.5th percentiles; values in the 0.5th-2.5th or 97.5th-99.5th percentiles were classified as mild deviation, and those beyond the 0.5th or 99.5th percentiles as at least moderate deviation. Associations of these categories with incident cardiovascular events over a median 4-year follow-up were evaluated in 59,916 UK Biobank participants. Results: The reference cohort included 4,624 women (51.0%) and 4,435 men (49.0%), mean age 61+/-13 years. Cardiac anatomy and function varied systematically with age and sex, enabling categories beyond conventional reference limits. Measurements from cvi42 and suiteHEART were comparable for ventricular parameters. In the UK Biobank Imaging Substudy, progression from within reference range to mildly and at least moderately out of range showed stepwise increases in cardiovascular risk across left and right ventricular metrics. Associations were strongest for heart failure and major cardiovascular disease; an at least moderately out-of-range LVEF conferred higher risk of incident heart failure (HR 8.81, 95% CI 7.08-10.97) and major cardiovascular disease (HR 2.74, 95% CI 2.32-3.23). Conclusions: We present reference ranges and centile-based categories for ventricular and atrial CMR measures from a large multi-ethnic reference population. These categories had demonstrable prognostic significance, identifying individuals at higher risk of major cardiovascular outcomes, and could support more standardized, clinically meaningful CMR reporting.