Large deformation diffeomorphic cardiac strain mapping

Beatrice Moscoloni, Patrick Segers, Mathias Peirlinck · 2025

Read Large deformation diffeomorphic cardiac strain mapping by Beatrice Moscoloni, Patrick Segers, Mathias Peirlinck (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03291.txt.

Introduction

Beatrice Moscoloni, Patrick Segers, and Mathias Peirlinck present a 2025 method for mapping heart muscle strain in fine local detail. Strain, the pattern of stretch and squeeze as the heart beats, is a key sign of heart health. Standard tools blur local detail by averaging over segments and missing motion through the imaging plane. The authors extend a shape-matching framework that handles large deformations smoothly, and they add a term that respects a basic fact of heart tissue: it barely changes volume.

The method reads time-resolved three-dimensional meshes and returns displacement and strain throughout the wall. Tested against a detailed computer simulation of a contracting heart, its estimates matched the known truth closely. The authors offer it as a mechanically honest way to study local heart function.

Catalog record

CREATORBeatrice Moscoloni, Patrick Segers, Mathias Peirlinck
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-03291

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Large deformation diffeomorphic cardiac strain mapping Beatrice Moscoloni, Patrick Segers, Mathias Peirlinck Abstract Cardiac deformation is a crucial biomarker for the evaluation of cardiac function. Current methods for estimating cardiac strain might underestimate local deformation due to through-plane motion and segmental averaging. Mesh-based mapping methods are gaining interest for localized analysis of cardiac motion and strain, yet they often do not consider important properties of cardiac tissue. In this work, we propose an extension of the large deformation diffeomorphic metric mapping framework to incorporate near incompressibility into the loss function that guides the mapping. As such, our mechanically regularized mLDDMM allows for accurate and mechanically coherent estimation of volume displacement and strain tensors from time-resolved three-dimensional meshes. We benchmark our method against the results of a finite element simulation of cardiac contraction and find a very good agreement between our estimation and the simulated ground truth. Our method forms a promising technique to extract volume displacement and strain tensors from time-resolved meshes while accounting for the incompressibility of cardiac tissue. 1 Introduction Cardiac strain analysis plays a critical role in diagnosing pathological heart conditions, with early measurements serving as crucial biomarkers for cardiac dysfunction [ 18 ] . Traditional methods, such as speckle-tracking echocardiography and cardiac magnetic resonance imaging feature tracking, are widely used but suffer from limitations. These include segmental averaging, which may obscure localized deformations, manual segmentation, and through-plane motion effects that can introduce variability in strain estimation [ 24 , 35 , 30 , 32 ] . Additionally, these techniques often rely on vendor-specific software, which limits broader applicability [ 23 ] . Mesh-based motion and strain analysis offers a more precise alternative for three-dimensional motion tracking of cardiovascular structures [ 6 , 29 , 15 , 21 , 23 , 9 ] . Recent advances in computer vision enable the automated extraction of detailed three-dimensional meshes from medical images [ 22 ] , although challenges remain to maintain a consistent mesh structure and vertex correspondence during displacement field estimation [ 6 , 16 , 21 ] . Large deformation diffeomorphic metric mapping (LDDMM) has been explored for motion tracking of cardiac structures, but it primarily focuses on surface meshes and does not capture the expected volumetric deformation of myocardial tissue [ 6 , 4 , 22 ] . In this work, we propose a novel method for extracting volumetric displacement and strain tensors from time-resolved meshes over a cardiac cycle. Our enriched LDDMM framework, which we call mLDDMM , incorporates a near-incompressibility constraint into the loss function that drives the mapping, ensuring accurate and mechanically coherent mapping - and reliable estimates - of volumetric deformation and strain tensors. We validate our approach with ground-truth data from a finite element simulation of cardiac contraction. 2 Methods 2.1 Synthetic data generation: finite element analysis of cardiac mechanics We set up a high-fidelity, four-chamber human heart model with proximal vasculature to simulate realistic cardiac mechanics throughout the cardiac cycle. The model, derived from high-resolution magnetic resonance and computed tomography imaging of a healthy, 50th percentile U.S. male [ 2 , 26 , 36 ] , captures detailed anatomical features, including chambers, valves, chordae tendinae, and papillary muscles. We assign the heart muscle’s intrinsic myofiber architecture using rule-based algorithms, informed by histological observations and diffusion tensor magnetic resonance images [ 3 , 17 ] . Specifically, we implement the Laplace-Dirichlet rule-based algorithm and impose the myofibers’ helical angle to vary transmurally from [formula] in the outer wall to [formula] in the inner wall [ 29 ] . We leverage a reconstructed geometry at 70% diastole where we discretize the myocardial wall using 303,870 quadratic tetrahedral elements and 69,910 nodes. Since the geometry of the heart is constructed while it is already hemodynamically loaded, we computationally account for the in vivo prestress at the beginning of the simulation using an inverse elastostatics method [ 27 ] . To model cardiac tissue mechanics, we enforce mass and momentum conservation laws and incorporate Holzapfel and Ogden’s orthotropic invariant-based constitutive model for passive elasticity [ 14 , 19 , 28 ] and a time-varying elastance model for active contraction [ 34 , 29 ] . We impose homogeneous Dirichlet boundary conditions at vascular outlets [ 29 ] and prescribe physiological pre- and after-loading conditions through a closed-loop lumped-parameter network model [ 2 ] . We sequentially calibrate passive and active material properties based on average in vivo pressure volume data for healthy subjects [ 26 ] . Specifically, we separately scale the combined sets of linear and exponential constitutive parameters, identifying optimal scaling factors by minimizing the difference between the simulated and experimentally derived diastolic pressure-volume relationships constructed using the Klotz curve [ 25 ] . Next, we optimize maximum contractility to obtain healthy peak systolic interventricular pressures and stroke volumes. Model validation is performed using clinical metrics such as ejection fraction and left ventricular twist [ 29 , 26 ] . Using the calibrated parameters, we simulate five cardiac cycles, after which the heart reaches cyclic steady-state equilibrium. Finally, we extract the ventricular displacement fields \mathbf{u} , which serve as ground-truth input data for our subsequent analyses. To ensure that the output of our deformation estimation and the ground truth deformation profile of our fifth steady-state cardiac cycle are comparable, we adjust the simulation output by subtracting the initial pre-stretch displacement from the entire displacement field. As such, both the estimation and the ground-truth consistently begin from the same zero-displacement baseline. 2.2 Volume displacement and deformation estimation Using the open-source LDDMM implementation in Deformetrica , we map the time-resolved mesh that geometrically discretizes the reference myocardial ventricular solid domain \Omega_{0} to a subsequent target mesh of the displaced and deformed domain \Omega . The mapping process involves iteratively deforming an ambient space of equally spaced control points embedding the template, while optimizing a loss function \mathcal{L} [ 7 , 12 , 22 ] which - classically - aims to minimize the Varifold distance \mathcal{D} between the reference surface mesh \Gamma_{0}=\partial\Omega_{0} and the target surface mesh \Gamma=\partial\Omega . The deformed collection of control points translates into a volumetric displacement field \mathbf{u}(\mathbf{X})=x which maps material points \mathbf{X}\in\Omega_{0} to their spatial positions \mathbf{x}\in\Omega in the displaced and deformed configuration. Given that the optimization of the surface-distance-based loss function does not intrinsically obey mechanical constraints of deformation within the tissue, we expand the classic LDDMM formulation to a mechanically regularized mLDDMM framework by including mechanical constraints in the loss function. More specifically, we constrain the solution space to only produce displacement fields \mathbf{u} that abide by the natural near-incompressible behavior of myocardial tissue. Towards this goal, during each mLDDMM-iteration i we leverage the proposed Varifold distance-based surface-matching \mathbf{u}_{\Gamma,i} to solve the boundary value problem for Laplace’s equation \displaystyle\nabla^{2}\mathbf{u}_{i}=0\quad \displaystyle\text{in } \displaystyle\Omega (1) \displaystyle\mathbf{u}=\mathbf{u}_{\Gamma,i}\quad \displaystyle\text{on } \displaystyle\partial\Omega=\Gamma. to compute the volumetric deformation gradient for the whole domain as: \mathbf{F}_{i}=\nabla{\mathbf{u}_{i}}+\mathbf{I}. (2) Next, we compute the mechanical incompressibility regularization term \mathcal{I}_{i} which penalizes deviations of the determinant of the deformation gradient \mathbf{F} from 1, weighted by the element volume in a finite element mesh: \mathcal{I}_{i}=\sum_{e}V_{e}\left(\det\left(\mathbf{F}_{i,e}\right)-1\right)^{2} (3) where V_{e} is the element volume, \mathbf{F}_{i,e} is the deformation gradient of iteration i , and \det\left(\mathbf{F}_{i,e}\right) measures local volume change. Our resulting mechanically

Read the complete text