Population imaging studies generate data for developing and implementing personalised health strategies to prevent, or more effectively treat disease. Large prospective epidemiological studies acquire imaging for pre-symptomatic populations. These studies enable the early discovery of alterations due to impending disease, and enable early identification of individuals at risk. Such studies pose new challenges requiring automatic image analysis. To date, few large-scale population-level cardiac imaging studies have been conducted. One such study stands out for its sheer size, careful implementation, and availability of top quality expert annotation; the UK Biobank (UKB). The resulting massive imaging datasets (targeting ca. 100,000 subjects) has put published approaches for cardiac image quantification to the test. In this paper, we present and evaluate a cardiac magnetic resonance (CMR) image analysis pipeline that properly scales up and can provide a fully automatic analysis of the UKB CMR study. Without manual user interactions, our pipeline performs end-to-end image analytics from multi-view cine CMR images all the way to anatomical and functional bi-ventricular quantification. All this, while maintaining relevant quality controls of the CMR input images, and resulting image segmentations. To the best of our knowledge, this is the first published attempt to fully automate the extraction of global and regional reference ranges of all key functional cardiovascular indexes, from both left and right cardiac ventricles, for a population of 20,000 subjects imaged at 50 time frames per subject, for a total of one million CMR volumes. In addition, our pipeline provides 3D anatomical bi-ventricular models of the heart. These models enable the extraction of detailed information of the morphodynamics of the two ventricles for subsequent association to genetic, omics, lifestyle habits, exposure information, and other information provided in population imaging studies. We validated our proposed CMR analytics pipeline against manual expert readings on a reference cohort of 4620 subjects with contour delineations and corresponding clinical indexes. Our results show broad significant agreement between the manually obtained reference indexes, and those automatically computed via our framework. 80.67% of subjects were processed with mean contour distance of less than 1 pixel, and 17.50% with mean contour distance between 1 and 2 pixels. Finally, we compare our pipeline with a recently published approach reporting on UKB data, and based on deep learning. Our comparison shows similar performance in terms of segmentation accuracy with respect to human experts.
CARDIOMARKER- Computational imaging phenomics in population cardiac MRI with automatic image quality assessment: benchmarking, scalability and inference with state-of-the-art algorithms.
Several cardiovascular conditions like heart failure, coronary artery disease, diabetes, and structural heart disease manifest in alterations of the anatomy or deformation of the myocardium. The hypothesis of this study is that existing tools for cardiac image analysis providing information on 3D cardiac morphology and deformation, developed for small patient cohorts, scale up to handle datasets in the order of hundreds and thousands of subjects. We will simultaneously undertake benchmarking of competing algorithms as demonstrate the impact of image analysis errors on relevant associative and causal inference tasks.
CARDIOMARKER will carry out a large-scale scalability testing of current image analysis tools ultimately helping UK Biobank researchers in extracting objective imaging phenotypic biomarkers of cardiac morphology and deformation correlated to disease presence, severity or progression. We will manually assess a subset of the datasets by two operators in two independent sessions. We will compare the manually segmented cardiac structures (MR cine) and tag intersections (MR tagging) from manual analysis against those of our automatic techniques.
CARDIOMARKER will elucidate how errors in CMR biomarkers influence the strength of associative and causal models. In collaboration with our clinical experts, we will formulate illustrative hypothesis re the association between cardiac morphology/deformation and genetic, lifestyle (activity, body mass composition), metabolism-related (bone ageing, liver function), environmental (exposures), and physiological (HR, BP) variables. We will generate associative/causative models, and will study the influence on those models of errors in CMR biomarkers derived from automatic analysis. This will shed light on the strength of the associations/causal relationships as a function of the size of the population and the noise level in the markers themselves.
We want to answer these questions in relationship to manual delineation and previous performance of the techniques:
a) What is the accuracy in cardiac anatomy delineation in population imaging studies?
b) What is the accuracy of extracted cardiac deformation fields?
c) What is the failure rate of automated methods operating on large-scale population imaging?
d) What is the impact of automated segmentation/registration errors on associative/causative models of phenotype-genotype relationship?
We will deliver the UKBB automatic and objective quantitative imaging information on the full cohort of patients imaged with CMR helping. These will establish population ranges of normality and thresholds of abnormality useful in cardiology. We will deliver knowledge on how to interpret and how errors compound in statistical modelling when attempting to unravel associative/causal relationships involving not only CMR but also other biomarkers.
|Lead investigator:||Professor Alejandro Frangi|
|Lead institution:||University of Leeds|
|3127||Quantitative CMR population imaging on 20,000 subjects of the UK Biobank imaging study: LV/RV quantification pipeline and its evaluation||Attar, R. et al||2019||Medical Image Analysis. 2019|