Three-dimensional cardiovascular imaging-genetics: a mass univariate framework

Abstract

Motivation: Left ventricular (LV) hypertrophy is a strong predictor of cardiovascular outcomes, but its genetic regulation remains largely unexplained. Conventional phenotyping relies on manual calculation of LV mass and wall thickness, but advanced cardiac image analysis presents an opportunity for high-throughput mapping of genotype-phenotype associations in three dimensions (3D).

Results: High-resolution cardiac magnetic resonance images were automatically segmented in 1124 healthy volunteers to create a 3D shape model of the heart. Mass univariate regression was used to plot a 3D effect-size map for the association between wall thickness and a set of predictors at each vertex in the mesh. The vertices where a significant effect exists were determined by applying threshold-free cluster enhancement to boost areas of signal with spatial contiguity. Experiments on simulated phenotypic signals and SNP replication show that this approach offers a substantial gain in statistical power for cardiac genotype-phenotype associations while providing good control of the false discovery rate. This framework models the effects of genetic variation throughout the heart and can be automatically applied to large population cohorts.

Availability and implementation: The proposed approach has been coded in an R package freely available at https://doi.org/10.5281/zenodo.834610 together with the clinical data used in this work.

Contact: declan.oregan@imperial.ac.uk.

Supplementary information: Supplementary data are available at Bioinformatics online.

Fig. 1.

Computational image analysis. (A) Short axis cardiac magnetic resonance image demonstrating automated segmentation of the endocardial and epicardial boundaries of the left ventricle. (B) The segmentation is used to construct a three dimensional mesh of the cardiac surfaces (left ventricle shown as a mesh, right ventricle shown as a solid) that is co-registered to a standard coordinate space. Phenotypic parameters, such as wall thickness, are then derived for each vertex in the model

Fig. 2.

Outline of three-dimensional mass univariate framework. A statistical atlas provides point-wise measures of ventricular geometry and function which can be linked to a given predictor through a general linear model. Using mass univariate regression, three-dimensional maps of a test statistic and the degree of association (β) can be derived. Threshold free cluster enhancement (TFCE) coupled with permutation testing produces vertex-wise P-values weighted to the degree of coherent spatial support. Finally, P-values are corrected for multiple testing. Regression coefficients enclosed by significance contours are represented on a model of the left ventricle

Fig. 3.

Applying three-dimensional analysis to single nucleotide polymorphism (SNP) replication. β coefficients are plotted on the surface of the left ventricle for the effect of 4 distinct SNPs on wall thickness (WT) adjusted for age, gender, body surface area and systolic blood pressure. Yellow contours enclose standardized regression coefficients reaching significance after multiple testing

Fig. 4.

Assessment of power using synthetic data. Plots of our framework’s sensitivity at different sample sizes N and signal intensities I to detect a synthetic signal on (A) 10% and (B) 60% of the LV surface. A black line on the plots indicates a threshold of 80% sensitivity