Left ventricular (LV) size and shape is important for quantifying cardiac remodelling in response to cardiovascular disease. Geometric remodelling indices have been shown to have prognostic value in predicting adverse events in the clinical literature, but these do not independently describe shape changes. We developed a method for deriving orthogonal shape components directly from any set of clinical indices.

Six clinical indices were chosen and ranked based on their variance in the population:

**End-Diastolic Volume Index (EDVI)**, which is end-diastolic volume (EDV) divided by body surface area.**Sphericity**= EDV divided by the volume of a sphere with a diameter corresponding to the major axis at ED in LV long-axis view.**Ejection Fraction (EF)**, which is (EDV - ESV) / EDV, where ESV = End-Systolic Volume.**Relative Wall Thickness (RWT)**= twice the posterior wall thickness divided by the ED diameter.**Conicity**= the ratio of the apical diameter (defined as the diameter of the endocardium one third above the apex) over the basal diameter at ED.**Longitudinal Shortening (LS)**, which is the difference of the distance of the central basal point to the apical point at ED and ES over the distance at ED.

The method used an iterative of partial least square (PLS) regressions followed by Gram-Schmidt orthogonalization procedures. At each step, the contribution of the previous component was removed mathematically from the shape description.

The inputs to the methods are homologous surface sample points and the number of latent (variable) to compute the PLS regression.

### Source codes

#### Download

Source codes are available from GitHub.

There are two versions: Matlab and R. Cloning the source tree:

- R version (master head)

git clone https://github.com/CardiacAtlasProject/lv-ortho-modes.git

git clone -b Matlab https://github.com/CardiacAtlasProject/lv-ortho-modes.git

ZIP download is also available from the GitHub page. Data used in this example were taken from MESA and DETERMINE data sets.