Pith. sign in

REVIEW 4 major objections 5 minor 67 references

WarpPINN-fibers: improved cardiac strain estimation from cine-MR with physics-informed neural networks

T0 review · 4 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Fiber-stretch penalty lifts cine-MRI strain accuracy by 44%

desk verdict Honest incremental extension of WarpPINN with a one-sided synthetic-fiber stretch penalty; the phantom evidence is real, but the in-vivo superiority claim rests on a 0.15 mm landmark gain, no significance test, and hyperparameters tuned on a volunteer who sits inside the aggregate. read the letter →

arxiv 2509.08872 v1 pith:YCZTFZLM submitted 2025-09-10 eess.IV cs.LGphysics.med-ph

Felipe Álvarez Barrientos , Tomás Banduc , Isabeau Sirven , Francisco Sahli Costabal This is my paper
classification eess.IVcs.LGphysics.med-ph
keywords cardiacstrainphysics-informedneuralnetworksimageregistrationcine-MRIfibersfiberstretchlandmarktrackinghyperelasticity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper seeks to show that a single physics constraint — cardiac muscle may shorten along its fiber direction but never stretch — substantially improves the accuracy and physiological realism of cardiac strain estimates computed from ordinary cine-MRI. The authors extend the WarpPINN image-registration network by adding a fiber-stretch term to the loss, with fiber directions generated by recipe rather than measured, so no diffusion-tensor or tagged-MRI acquisition is needed. They report lower landmark-tracking error (median 2.72 mm vs 2.87 mm for the base model, and 3.17, 4.88, and 3.81 mm for three published alternatives) and strain curves with clinically expected signs. They also report a 44% reduction in displacement error on a synthetic phantom. If the claim holds, fiber-informed strain analysis from routine cardiac MR becomes accessible without prolonged scan protocols.

What carries the argument

The load-bearing object is the fiber-stretch regularizer L_f(θ) = µ_f (1/N_f) Σ (max{1, λ²_f} − 1)², where λ²_f = C(φ): f⊗f is the projection of the right Cauchy-Green deformation tensor onto the end-diastolic fiber direction f. Because the max function activates only when λ²_f > 1, the term penalizes stretch while leaving contraction free. Fiber directions come from a Laplace-Dirichlet rule-based algorithm parameterized by a single helix angle α (with sheet angle fixed at zero), so the mechanism injects anisotropic fiber-level mechanics into an otherwise isotropic hyper-elastic regularizer without new imaging data.

What would settle it

Acquire diffusion-tensor MR fiber orientations for a few benchmark subjects, retrain with those measured fibers instead of the rule-based ones, and compare landmark errors and strain curves: if accuracy does not improve over the synthetic-fiber version — or degrades — the synthetic-fiber assumption is not the cause of the gains. Alternatively, run the synthetic phantom with deliberately wrong helix angles (e.g., 0° or 90°) and check whether the one-sided penalty introduces structured errors.

Watch

Extended reading notes

Core claim

The paper proposes WarpPINN-fibers, a physics-informed neural network that estimates heart deformation from cine-MRI by learning a displacement field φ(t,X) = X + û(t,X;θ), trained with a loss that adds a fiber-stretch penalization L_f to the data-similarity and quasi-incompressibility terms of the earlier WarpPINN model. The core claim is that imposing λ²_f ≤ 1 — allowing tissue to shorten along its fiber direction but not stretch — yields deformation fields that match known motion more closely and produce strains consistent with cardiac physiology. On a synthetic phantom, the fiber-aware model lowers displacement mean-square error to 6.47×10⁻² from 1.16×10⁻¹ and removes anomalous fiber str

Load-bearing premise

The central fragile premise is that rule-generated fiber directions — fixed by one helix angle and zero sheet angle — are close enough to each volunteer's true heart-muscle fiber directions that forbidding stretch along them produces the right deformation; the paper itself notes the true orientation is unknown, so this cannot be verified, and it already must exclude the base because synthetic fibers there are disorganized.

Editorial extensions

If this is right

  • Deformation fields from cine-MRI can be constrained by fiber physiology without measuring fiber orientation, since rule-based fibers with helix angles in the 50°–70° range proved sufficient on the benchmark.
  • Fiber stretch stays at or below unity across most of the myocardium (outside the excluded base), eliminating the stretch peaks above 1 that the base model produced.
  • Median landmark tracking on the 15-volunteer benchmark improves to 2.72 mm, the best of the methods compared, with less variance in errors.
  • Strain curves take on clinically expected behavior: negative longitudinal strain at the base (mitral annular motion toward the apex), negative circumferential strain, and positive radial strain from wall thickening.
  • Fiber mechanics can be incorporated into strain estimation from standard cine-MRI at no acquisition cost, potentially improving assessment of conditions linked to impaired myocardial mechanics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the fiber constraint is one-sided and repels stretch anomalies, a similar 'shortening-only along a structural direction' penalty might transfer to other soft-tissue registration problems, such as skeletal muscle or arterial wall motion, where a dominant structural orientation exists.
  • The method's reliance on synthetic fibers is its main empirical risk: if actual fibers deviate from the assumed helix angle, the constraint could bias the deformation. A head-to-head test using patient-specific diffusion-tensor fiber orientations on the same landmarks would be the direct check.
  • The base region is excluded from the penalty because rule-based fibers misalign there; clinical deployment would likely need those segments handled separately or with measured fibers.
  • The result suggests image-only fidelity and physics-based priors can combine so strain estimates gain physiological plausibility, but the magnitude of strain remains hard to validate without ground-truth deformation tags in the same subjects.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes WarpPINN-fibers, an extension of the physics-informed WarpPINN registration framework that adds a fiber-stretch penalty to the loss. Fiber orientations are generated synthetically with the LDRB rule-based method, parameterized by a helix angle α, and the penalty enforces λ²_f ≤ 1 along those fiber directions. The method is validated on a synthetic cylindrical phantom with known deformation and on the STACOM-2011 cine-MRI benchmark with 15 healthy volunteers. The reported results are a synthetic displacement MSE of 6.47e-2 versus 1.16e-1 for WarpPINN, a median landmark error of 2.72 mm versus 2.87 mm for WarpPINN, and qualitative strain-curve comparisons against UPF, INRIA, and CarMEN.

Significance. If the real-data claims were quantitatively established, the contribution would be valuable: adding fiber-aware mechanics to image-registration PINNs without requiring DTI or other extra acquisitions is an appealing and practical direction. The synthetic phantom experiment gives a concrete, reproducible testbed and shows a clear MSE improvement, and the landmark evaluation on a public benchmark is a useful point of comparison. However, the paper's central claim that WarpPINN-fibers 'outperforms alternative methodologies in landmark-tracking and strain curve prediction' is not supported by the evidence as presented: the landmark gain over WarpPINN is small and untested, the strain-curve comparisons have no ground truth or quantitative metric, and hyperparameters were selected on a volunteer included in the aggregate results. With additional statistical analysis and more careful claims, the contribution could become solid; in its current form, the headline superiority is not demonstrated.

major comments (4)
  1. [§3.2, Figure 5] The only quantitative real-data advantage over WarpPINN is a median landmark error of 2.72 mm versus 2.87 mm across 15 volunteers, with no significance test, confidence interval, or per-volunteer breakdown. This 0.15 mm difference is within the noise level of manual landmark tracking and could easily arise from chance. Moreover, §2.7 states that the fiber hyperparameters (α ∈ {50°,60°,70°} and μ_f ∈ {10^-3,...,10^3}) were tuned on volunteer v1, and Figure 6 shows that v1 was used for this selection; v1 is also included in the aggregate boxplot of Figure 5. The reported gain is therefore partly a selection artifact. Please provide paired per-volunteer differences with a Wilcoxon signed-rank test or bootstrap confidence intervals, and report the aggregate results both with and without v1.
  2. [§3.2, Figure 8] The claim in the abstract and conclusion that WarpPINN-fibers 'outperforms alternative methodologies in ... strain curve prediction' is not supported by the presented evidence. §3.2 and §4 explicitly state that there is no ground-truth strain in the STACOM-2011 data, and Figure 8 shows curves for only two volunteers (v4 and v9). No quantitative strain error, correlation, or agreement metric is reported, and the selection criterion for those two volunteers is not stated. Either provide a quantitative strain assessment (e.g., comparison against a reference modality, or strain errors on a synthetic phantom with known ground truth) or revise the claim to state that strain curves are qualitatively plausible and consistent with clinical expectations.
  3. [§2.2.2 and §3.1] The fiber-stretch improvement reported in the synthetic experiment is partly a restatement of the loss. The regularizer L_f(θ) = μ_f (1/N_f) Σ (max{1, λ²_f} - 1)² directly penalizes values above 1, so observing that WarpPINN-fibers produces λ²_f ≤ 1 and upper-bounded stretch is expected by construction, not independent evidence that the deformation is more physiological. The MSE improvement (6.47e-2 vs 1.16e-1) is a valid quantitative result and should be emphasized, but the qualitative 'fixes the extrusion' claim should be backed by displacement error maps in the affected regions. In addition, the choice μ_f=100 for the synthetic experiment is not described as part of a sensitivity analysis; please clarify whether this value was selected using the same phantom and, if so, report robustness.
  4. [§2.7] The physiological correctness of the method rests on the assumption that LDRB-generated fibers with a single helix angle α and β=0 approximate each volunteer's true end-diastolic fiber architecture. The authors acknowledge this explicitly: 'Since real fiber orientation is unknown, we cannot verify if a chosen angle α accurately reflects the anatomical fiber distribution of the patient.' They also exclude the base because synthetic fibers there are 'disorganized.' If this assumption is wrong, the constraint will bias the deformation and strain estimates. The paper should stress-test this assumption, for example by showing that landmark error and strain curves degrade gracefully for deliberately wrong α values, or by comparing against a dataset with DTI-derived fibers. As written, the method's principal real-data gain could be an artifact of the synthetic fiber field rather than of improve
minor comments (5)
  1. [Throughout] The method name is written inconsistently as 'WarpPINN' and 'warpPINN' (e.g., Figure 5 text and §2.7). Please standardize capitalization.
  2. [Figure 7 caption] The caption lists v1 through v16, but the dataset has 15 volunteers and the main text refers to v1, v2, v4, etc. Confirm the numbering and ensure no volunteer is mislabeled.
  3. [§2.2.2] The notation for the fiber stretch is inconsistent: the text says 'max{1,λ}−1' but the loss uses max{1, λ²_f}−1. Clarify whether λ denotes stretch or stretch squared.
  4. [§2.7] The text says landmarks are evaluated at end-systole and end-diastole, but Figure 5 appears to show errors at a single time point. State explicitly which time points are aggregated in the boxplot.
  5. [§3.1] The vibration plot in Figure 3 is useful, but the y-axis range and the density estimator bandwidth should be specified for reproducibility.

Circularity Check

2 steps flagged · score 6.0 of 10

Fiber-stretch 'control' is a restatement of the loss; in-vivo benchmark is partly a v1-tuned result rather than an independent prediction.

  1. self definitional [Section 2.2.2 (L_f definition) and Section 4]
    "We demonstrate that the inclusion of a fiber penalization term favors realistic contraction by making the right Cauchy-Green deformation tensor projected onto the fiber direction at most 1."

    The fiber regularizer is defined in §2.2.2 as L_f(θ)=µ_f(1/N_f)Σ(max{1,λ_f²}-1)², whose minimum is attained when λ_f²≤1. Thus the claimed 'demonstration' that WarpPINN-fibers keeps fiber stretch at most 1 is a restatement of the loss being optimized, not an independent validation of physiological realism. The same is true for the §3.1 violin-plot observation that WarpPINN-fibers is 'upper-bounded by 1'. The independent part of the synthetic experiment is the MSE against the known displacement field, which is not circular, but the paper also presents the constraint-satisfaction as evidence of improved mechanics.

  2. fitted input called prediction [Section 2.7 / Section 3.2, Figures 5-6]
    "To evidence the proficiency of our method over different hyper-parameter configurations, we show in Figure 6 a boxplot that measures the accuracy of our network for v1 over multiple combinations of α and µ_f. ... The dotted gray line indicates the lowest median achieved, which is obtained with WarpPINN-fibers for the parametersµ=10 −5,σ=1, andµ f =10."

    The hyperparameter grid over α∈{50°,60°,70°} and µ_f∈{10^-3,...,10^3} is evaluated on volunteer v1's landmark error (Fig. 6). The headline aggregate median of 2.72 mm (Fig. 5) is computed on the 15-volunteer STACOM cohort that includes v1, using µ_f=10 taken from that v1 grid. No held-out split or significance test is reported, so v1's contribution to the reported improvement is a selected, not predicted, quantity. This is benchmark leakage rather than a full by-construction equivalence, but it makes the in-vivo 'outperforms' claim partly dependent on the fitting choice.

full rationale

The paper contains two specific circular/leakage steps. First, the fiber-stretch control result is definitional: L_f is exactly a penalty on λ²_f>1, so reporting that WarpPINN-fibers keeps λ²_f≤1 is a restatement of the objective, not an independent finding. The synthetic MSE comparison against the known phantom displacement is genuinely independent and does support the mechanism. Second, the in-vivo landmark improvement (2.72 vs 2.87 mm) is weakened by the fact that the fiber-angle and fiber-weight hyperparameters were selected by landmark accuracy on volunteer v1, and v1 is included in the aggregate used for the headline comparison; no hold-out or significance test is reported. The strain-curve component is explicitly acknowledged by the authors to lack ground truth, so it cannot independently substantiate the superiority claim. The self-citation to WarpPINN [1] is a normal baseline citation and is not load-bearing circularity. Overall, the central claim is not fully forced by construction, but it contains a definitional 'prediction' and a partially fitted benchmark, so the circularity score is 6.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

The central result rests on a chain of non-derived modeling assumptions: diffeomorphic registration, neo-Hookean incompressibility, one-sided fiber shortening, and LDRB synthetic fiber fields with a hand-set helix angle. The paper does not validate the fiber fields against patient anatomy and excludes the base where the fields are unreliable.

free parameters (6)
  • Helix angle α = 50°-70° (best unspecified; 60° used in figures)
    User-specified LDRB parameter governing synthetic fiber orientation; cannot be verified against patient anatomy (Sections 2.3, 2.7).
  • Fiber penalty weight µ_f = 100 (synthetic), 10 (clinical)
    Controls strength of fiber-stretch penalization; selected on volunteer v1 in the clinical setting, contributing to the reported median.
  • Neo-Hookean weight µ = 10^-2 (synthetic), 10^-5 (clinical)
    Weights the quasi-incompressibility regularizer; chosen by hand from prior WarpPINN practice.
  • Myocardial bulk modulus λ_myo = 10^4 (synthetic), 10^5 (clinical)
    Enforces quasi-incompressibility in the myocardium; hand-chosen.
  • Background bulk modulus λ_bg = 50 (synthetic), 1 (clinical)
    Keeps background weakly constrained; hand-chosen.
  • Fourier feature parameters (m, σ) = m=8, σ=1 (synthetic); m=32, σ=1 (clinical)
    Controls frequency spectrum of input encoding; adopted from WarpPINN with modest tuning.
assumptions (7)
  • domain assumption A diffeomorphic map φ exists such that R = T∘φ across the cardiac cycle
    Section 2.1 states this as the registration premise; if the true motion is not diffeomorphic at image resolution, the formulation is misspecified.
  • domain assumption Neo-Hookean hyperelastic energy W with quasi-incompressibility is an appropriate physical regularizer for cardiac tissue
    Section 2.2.1 adopts W(φ;λ)=Tr(C)-3-2log(J)+λ(J-1)² for myocardium and background.
  • domain assumption Healthy cardiac fibers only shorten during systole (λ²_f ≤ 1)
    Section 2.2.2 postulates the one-sided constraint used in L_f; unphysiological stretching is penalized everywhere.
  • ad hoc to paper LDRB synthetic fibers with helix angle α approximate true fiber orientation
    Section 2.3 generates fiber fields purely from geometry and α; Section 2.7 admits the orientation cannot be verified.
  • ad hoc to paper Excluding basal points from the fiber penalty is valid because synthetic fibers there are disorganized
    Section 2.7 excludes base points post hoc; the excluded region is still evaluated in landmark metrics.
  • domain assumption Strain decomposition using global longitudinal direction l=(0,0,1) and mesh-derived radial/circumferential axes is valid
    Section 2.7 defines E components for AHA-segment strain curves; this is a post-processing assumption.
  • domain assumption Manually tracked landmarks from tagged MRI, converted to SSFP coordinates, are reliable ground truth at end-systole and end-diastole
    Section 2.7 uses these limits for landmark error; the conversion limits evaluation to two time points.

how reviews work

0 comments
Cite this review

Pith. "Pith review of WarpPINN-fibers: improved cardiac strain estimation from cine-MR with physics-informed neural networks." pith.science (2026). https://pith.science/paper/YCZTFZLM

@misc{pith2026250908872,
  author       = {Pith},
  title        = {Pith review of: WarpPINN-fibers: improved cardiac strain estimation from cine-MR with physics-informed neural networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YCZTFZLM}},
  note         = {Machine review of arXiv:2509.08872}
}
read the original abstract

The contractile motion of the heart is strongly determined by the distribution of the fibers that constitute cardiac tissue. Strain analysis informed with the orientation of fibers allows to describe several pathologies that are typically associated with impaired mechanics of the myocardium, such as cardiovascular disease. Several methods have been developed to estimate strain-derived metrics from traditional imaging techniques. However, the physical models underlying these methods do not include fiber mechanics, restricting their capacity to accurately explain cardiac function. In this work, we introduce WarpPINN-fibers, a physics-informed neural network framework to accurately obtain cardiac motion and strains enhanced by fiber information. We train our neural network to satisfy a hyper-elastic model and promote fiber contraction with the goal to predict the deformation field of the heart from cine magnetic resonance images. For this purpose, we build a loss function composed of three terms: a data-similarity loss between the reference and the warped template images, a regularizer enforcing near-incompressibility of cardiac tissue and a fiber-stretch penalization that controls strain in the direction of synthetically produced fibers. We show that our neural network improves the former WarpPINN model and effectively controls fiber stretch in a synthetic phantom experiment. Then, we demonstrate that WarpPINN-fibers outperforms alternative methodologies in landmark-tracking and strain curve prediction for a cine-MRI benchmark with a cohort of 15 healthy volunteers. We expect that our method will enable a more precise quantification of cardiac strains through accurate deformation fields that are consistent with fiber physiology, without requiring imaging techniques more sophisticated than MRI.

Figures

Figures reproduced from arXiv: 2509.08872 by the authors.

Figure 1
Figure 1. In the neural network, the inputs are the time [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Fiber stretch over the cylindrical phantom for the ground truth (left), WarpPINN (center) and WarpPINN-fibers (right). We show the [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Displacement and fiber stretch of synthetic example for the ground truth (black), WarpPINN (red) and WarpPINN-fibers (blue). On the [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Influence of fiber penalization on deformation prediction and fiber stretch. (a) Fiber stretch in the deformed configuration of v1 at end [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Predictions for WarpPINN (PINN), WarpPINN-fibers (PINN-fibers), UPF, INRIA and Carmen for manually-tracked landmarks. On the [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Box plots for landmark tracking errors on v1 using WarpPINN-fibers with di [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Deformed configurations at end-systole for WarpPINN and WarpPINN-fibers (WarpPINN-fib) with corresponding fiber stretch for each [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: Strain curves predicted by WarpPINN (red; [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

67 extracted references · 29 canonical work pages

  1. [1]

    Warppinn: Cine- mr image registration with physics-informed neural networks

    Arratia López, P., Mella, H., Uribe, S., Hurtado, D.E., Sahli Costabal, F., 2023. Warppinn: Cine- mr image registration with physics-informed neural networks. Medical Image Analysis 89, 102925. URL:https://www.sciencedirect.com/science/article/pii/S1361841523001858, doi:10.1016/ j.media.2023.102925

    arXiv 2023

  2. [2]

    The sequence of regional ventricular motion

    Ballester-Rodés, M., Flotats, A., Torrent-Guasp, F., Carrió-Gasset, I., Ballester-Alomar, M., Carreras, F., Ferreira, A., Narula, J., 2006. The sequence of regional ventricular motion. European Journal of Cardio- Thoracic Surgery 29, S139–S144. URL:https://doi.org/10.1016/j.ejcts.2006.02.058, doi:10. 1016/j.ejcts.2006.02.058

    work page doi:10.1016/j.ejcts.2006.02.058 2006

  3. [3]

    Modeling of cardiac fibers as oriented liquid crystals

    Barnafi, N.A., Osses, A., 2025. Modeling of cardiac fibers as oriented liquid crystals. Computer Methods in Applied Mechanics and Engineering 436, 117710. URL:https://www.sciencedirect.com/science/ article/pii/S0045782524009642, doi:10.1016/j.cma.2024.117710. 16

    arXiv 2025

  4. [4]

    A novel rule-based algorithm for assigning myocar- dial fiber orientation to computational heart models

    Bayer, J.D., Blake, R.C., Plank, G., Trayanova, N.A., 2012. A novel rule-based algorithm for assigning myocar- dial fiber orientation to computational heart models. Annals of Biomedical Engineering 40, 2243–2254. URL: https://doi.org/10.1007/s10439-012-0593-5, doi:10.1007/s10439-012-0593-5

    doi:10.1007/s10439-012-0593-5 2012

  5. [5]

    A comparison of both dense and feature tracking techniques with tagging for the cardiovascular magnetic resonance assessment of myocardial strain

    Cao, J., Ngai, N., Duncanson, L., Cheng, J., Gliganic, K., Chen, Q., 2018. A comparison of both dense and feature tracking techniques with tagging for the cardiovascular magnetic resonance assessment of myocardial strain. Journal of Cardiovascular Magnetic Resonance 20. doi:10.1186/s12968-018-0448-9

    work page doi:10.1186/s12968-018-0448-9 2018

  6. [6]

    Longitudinally and circumferentially directed movements of the left ventricle studied by cardiovascular magnetic resonance phase contrast velocity mapping

    Codreanu, I., Robson, M.D., Golding, S.J., Jung, B.A., Clarke, K., Holloway, C.J., 2010. Longitudinally and circumferentially directed movements of the left ventricle studied by cardiovascular magnetic resonance phase contrast velocity mapping. Journal of Cardiovascular Magnetic Resonance 12, 48. URL:https://www. sciencedirect.com/science/article/pii/S109...

    work page doi:10.1186/1532-429x-12-48 2010

  7. [7]

    De Craene, M., Tobon-Gomez, C., Butakoff, C., Duchateau, N., Piella, G., Rhode, K.S., Frangi, A.F., 2012. Tem- poral diffeomorphic free form deformation (tdffd) applied to motion and deformation quantification of tagged mri sequences, in: Camara, O., Konukoglu, E., Pop, M., Rhode, K., Sermesant, M., Young, A. (Eds.), Sta- tistical Atlases and Computationa...

    2012

  8. [8]

    Studying dynamic myofiber aggregate reorientation in dilated cardiomyopathy using in vivo mag- netic resonance diffusion tensor imaging

    von Deuster, C., Sammut, E., Asner, L., Nordsletten, D., Lamata, P., Stoeck, C.T., Kozerke, S., Razavi, R., 2016. Studying dynamic myofiber aggregate reorientation in dilated cardiomyopathy using in vivo mag- netic resonance diffusion tensor imaging. Circulation: Cardiovascular Imaging 9, e005018. URL:https: //www.ahajournals.org/doi/abs/10.1161/CIRCIMAGI...

    work page doi:10.1161/circimaging.116.005018 2016

Show all 67 references
  1. [9]

    A rule-based method to model myocardial fiber orien- tation in cardiac biventricular geometries with outflow tracts

    Doste, R., Soto-Iglesias, D., Bernardino, G., Alcaine, A., Sebastian, R., Giffard-Roisin, S., Sermesant, M., Berruezo, A., Sanchez-Quintana, D., Camara, O., 2019. A rule-based method to model myocardial fiber orien- tation in cardiac biventricular geometries with outflow tract...

    2019 doi

  2. [10]

    Myofiber angle distributions in the ovine left ventricle do not conform to computationally optimized predictions

    Ennis, D.B., Nguyen, T.C., Riboh, J.C., Wigström, L., Harrington, K.B., Daughters, G.T., Ingels, N.B., Miller, D.C., 2008. Myofiber angle distributions in the ovine left ventricle do not conform to computationally optimized predictions. Journal of biomechanics 41, 3219–3224

    2008

  3. [11]

    Automatic quantitation of re- gional myocardial wall motion and thickening from gated technetium-99m sestamibi myocardial perfusion single-photon emission computed tomography

    Germano, G., Erel, J., Lewin, H., Kavanagh, P.B., Berman, D.S., 1997. Automatic quantitation of re- gional myocardial wall motion and thickening from gated technetium-99m sestamibi myocardial perfusion single-photon emission computed tomography. Journal of the American College...

    1997

  4. [12]

    Novel therapeutic concepts the epidemic of cardiovascular disease in the developing world: global implications

    Gersh, B.J., Sliwa, K., Mayosi, B.M., Yusuf, S., 2010. Novel therapeutic concepts the epidemic of cardiovascular disease in the developing world: global implications. European Heart Journal 31, 642–648. URL:https://doi.org/10.1093/eurheartj/ehq030, doi:10.1093/eurheartj/ehq030...

    2010 doi

  5. [13]

    Characterizing cardiac involvement in amyloidosis using cardiovascular magnetic resonance diffusion tensor imaging

    Gotschy, A., Deuster, C., Gorkum, R., Gastl, M., Vintschger, E., Schwotzer, R., Flammer, A., Manka, R., Stoeck, C., Kozerke, S., 2019. Characterizing cardiac involvement in amyloidosis using cardiovascular magnetic resonance diffusion tensor imaging. Journal of Cardiovascular ...

    2019

  6. [14]

    Cmr diffusion tensor imaging provides novel imaging markers of adverse myocardial remodeling in aortic stenosis

    Gotschy, A., von Deuster, C., Weber, L., Gastl, M., Schmiady, M.O., van Gorkum, R.J., Stimm, J., von Spiczak, J., Manka, R., Kozerke, S., Stoeck, C.T., 2021. Cmr diffusion tensor imaging provides novel imaging markers of adverse myocardial remodeling in aortic stenosis. JACC: ...

    2021 doi

  7. [15]

    Physics-informed neural networks to learn cardiac fiber orientation from multiple electroanatomical maps

    Herrera, C.R., Grandits, T., Plank, G., Perdikaris, P., Costabal, F.S., Pezzuto, S., 2022. Physics-informed neural networks to learn cardiac fiber orientation from multiple electroanatomical maps. Engineering with Computers 38, 3957–3973. URL:https://doi.org/10.1007/s00366-022...

    2022 doi

  8. [16]

    Structural changes in muscle during contraction: Interference microscopy of living muscle fibres

    Huxley, A.F., Niedergerke, R., 1954. Structural changes in muscle during contraction: Interference microscopy of living muscle fibres. Nature 173, 971–973. URL:https://doi.org/10.1038/173971a0, doi:10.1038/ 173971a0

    1954 doi

  9. [17]

    Changes in the cross-striations of muscle during contraction and stretch and their structural interpretation

    Huxley, H., Hanson, J., 1954. Changes in the cross-striations of muscle during contraction and stretch and their structural interpretation. Nature 173, 973–976. URL:https://doi.org/10.1038/173973a0, doi:10.1038/ 173973a0

    1954 doi

  10. [18]

    Diffusion tensor cardiovascular magnetic reso- nance in cardiac amyloidosis

    Khalique, Z., Ferreira, P.F., Scott, A.D., Nielles-Vallespin, S., Martinez-Naharro, A., Fontana, M., Hawkins, P., Firmin, D.N., Pennell, D.J., 2020. Diffusion tensor cardiovascular magnetic reso- nance in cardiac amyloidosis. Circulation: Cardiovascular Imaging 13, e009901. UR...

    2020 doi

  11. [19]

    Adam: A method for stochastic optimization

    Kingma, D.P., Ba, J., 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

    2014 arXiv

  12. [20]

    Systematic review and meta-analysis for the value of cardiac magnetic resonance strain to predict cardiac outcomes

    Korosoglou, G., Sagris, M., André, F., et al., 2024. Systematic review and meta-analysis for the value of cardiac magnetic resonance strain to predict cardiac outcomes. Scientific Reports 14, 1094. URL:https: //doi.org/10.1038/s41598-023-50835-5, doi:10.1038/s41598-023-50835-5

    2024 doi

  13. [21]

    Statistically-driven 3d fiber recon- struction and denoising from multi-slice cardiac dti using a markov random field model

    Lekadir, K., Lange, M., Zimmer, V .A., Hoogendoorn, C., Frangi, A.F., 2016. Statistically-driven 3d fiber recon- struction and denoising from multi-slice cardiac dti using a markov random field model. Medical Image Analy- sis 27, 105–116. URL:https://www.sciencedirect.com/scie...

    2016 doi

  14. [22]

    Effects of orientation of myocardial fibers on the contractility of left ventricle

    Li, P., Sun, A., Guo, C., Peng, Z., Wang, C., 2025. Effects of orientation of myocardial fibers on the contractility of left ventricle. Journal of the Mechanical Behavior of Biomedical Materials 168, 107025. URL:https://www.sciencedirect.com/science/article/pii/S17516161250014...

    2025

  15. [23]

    Lin, K., Collins, J.D., Chowdhary, V ., Markl, M., Carr, J.C., 2016. Heart deformation analysis for automated quantification of cardiac function and regional myocardial motion patterns: A proof of concept study in patients with cardiomyopathy and healthy subjects. European Jou...

    2016 doi

  16. [24]

    Human atlas of the cardiac fiber architecture: Study on a healthy population

    Lombaert, H., Peyrat, J.M., Croisille, P., Rapacchi, S., Fanton, L., Cheriet, F., Clarysse, P., Magnin, I., Delingette, H., Ayache, N., 2012. Human atlas of the cardiac fiber architecture: Study on a healthy population. IEEE Transactions on Medical Imaging 31, 1436–1447. doi:1...

    2012

  17. [25]

    ilogdemons: A demons-based registra- tion algorithm for tracking incompressible elastic biological tissues

    Mansi, T., Pennec, X., Sermesant, M., Delingette, H., Ayache, N., 2011. ilogdemons: A demons-based registra- tion algorithm for tracking incompressible elastic biological tissues. International Journal of Computer Vision 92, 92–111. doi:10.1007/s11263-010-0405-z

    2011 doi

  18. [26]

    2025 heart disease and stroke statistics: A report of us and global data from the american heart association

    Martin, S.S., Aday, A.W., Allen, N.B., Almarzooq, Z.I., Anderson, C.A., Arora, P., Avery, C.L., Baker-Smith, C.M., Bansal, N., Beaton, A.Z., ..., Palaniappan, L.P., on behalf of the American Heart Association Council on Epidemiology, Committee, P.S., Committee, S.S., 2025. 202...

    2025 doi

  19. [27]

    Mitral annular plane systolic excursion as a surrogate for left ventricular ejection fraction

    Matos, J., Kronzon, I., Panagopoulos, G., Perk, G., 2012. Mitral annular plane systolic excursion as a surrogate for left ventricular ejection fraction. Journal of the American Society of Echocardiography 25, 969–974. URL:https://www.sciencedirect.com/science/article/pii/S0894...

    2012 doi

  20. [28]

    Mauger, C., Gilbert, K., Suinesiaputra, A., Pontré, B., Omens, J., McCulloch, A., Young, A., 2018. An iterative diffeomorphic algorithm for registration of subdivision surfaces: Application to congenital heart disease, in: 2018 40th Annual International Conference of the IEEE ...

    2018

  21. [29]

    Global burden of cardiovascular diseases and risks, 1990-

    Mensah, G.A., Fuster, V ., Murray, C.J., Roth, G.A., Mensah, G.A., Abate, Y .H., Abbasian, M., Abd-Allah, F., Abdollahi, A., Abdollahi, M., ..., Roth, G.A., 2023. Global burden of cardiovascular diseases and risks, 1990-

    2023

  22. [30]

    Three-dimensional systolic strain patterns in the normal human left ventricle: characterization with tagged mr imaging

    Moore, C.C., Lugo-Olivieri, C.H., McVeigh, E.R., Zerhouni, E.A., 2000. Three-dimensional systolic strain patterns in the normal human left ventricle: characterization with tagged mr imaging. Radiology 214, 453–466

    2000

  23. [31]

    Deepstrain: A deep learning workflow for the auto- mated characterization of cardiac mechanics

    Morales, M.A., van den Boomen, M., Nguyen, C., Kalpathy-Cramer, J., Rosen, B.R., Stultz, C.M., Izquierdo-Garcia, D., Catana, C., 2021. Deepstrain: A deep learning workflow for the auto- mated characterization of cardiac mechanics. Frontiers in Cardiovascular Medicine V olume 8 -

    2021

  24. [32]

    Myofiber strain in healthy humans using dense and cdti

    Moulin, K., Croisille, P., Viallon, M., Verzhbinsky, I.A., Perotti, L.E., Ennis, D.B., 2021a. Myofiber strain in healthy humans using dense and cdti. Magnetic Resonance in Medicine 86, 277–292. URL:https://doi. org/10.1002/mrm.28724, doi:10.1002/mrm.28724

    doi

  25. [33]

    Myofiber strain in healthy humans using dense and cdti

    Moulin, K., Croisille, P., Viallon, M., Verzhbinsky, I.A., Perotti, L.E., Ennis, D.B., 2021b. Myofiber strain in healthy humans using dense and cdti. Magnetic resonance in medicine 86, 277–292

  26. [34]

    Ordas, S., Frangi, A., 2005. Automatic quantitative analysis of myocardial wall motion and thickening from long-and short-axis cine mri studies, in: 2005 IEEE Engineering in Medicine and Biology 27th Annual Confer- ence, pp. 7028–7031. doi:10.1109/IEMBS.2005.1616124

    2005

  27. [35]

    Estimat- ing cardiomyofiber strain in vivo by solving a computational model

    Perotti, L.E., Verzhbinsky, I.A., Moulin, K., Cork, T.E., Loecher, M., Balzani, D., Ennis, D.B., 2021. Estimat- ing cardiomyofiber strain in vivo by solving a computational model. Medical Image Analysis 68, 101932. URL:https://www.sciencedirect.com/science/article/pii/S1361841...

    2021

  28. [36]

    Quantification of 3d regional myocardial wall thickening from gated magnetic resonance images

    Prasad, M., Ramesh, A., Kavanagh, P., Tamarappoo, B., Nakazato, R., Gerlach, J., Cheng, V ., Thomson, L., Berman, D., Germano, G., Slomka, P., 2010. Quantification of 3d regional myocardial wall thickening from gated magnetic resonance images. Journal of magnetic resonance ima...

    2010

  29. [37]

    On the spectral bias of neural networks, in: Chaudhuri, K., Salakhutdinov, R

    Rahaman, N., Baratin, A., Arpit, D., Draxler, F., Lin, M., Hamprecht, F., Bengio, Y ., Courville, A., 2019. On the spectral bias of neural networks, in: Chaudhuri, K., Salakhutdinov, R. (Eds.), Proceedings of the 36th International Conference on Machine Learning, PMLR. pp. 530...

    2019

  30. [38]

    Optimization of cardiac fiber orientation for homogeneous fiber strain at beginning of ejection

    Rijcken, J., Bovendeerd, P.H., Schoofs, A.J., van Campen, D.H., Arts, T., 1997. Optimization of cardiac fiber orientation for homogeneous fiber strain at beginning of ejection. Journal of Biomechanics 30, 1041–1049. doi:10.1016/s0021-9290(97)00064-x. 19

    1997 doi

  31. [39]

    Impact of intraventricular septal fiber orientation on cardiac electromechanical function

    Rodríguez-Padilla, J., Petras, A., Magat, J., Bayer, J., Bihan-Poudec, Y ., El Hamrani, D., Ramlugun, G., Neic, A., Augustin, C.M., Vaillant, F., Constantin, M., Benoist, D., Pourtau, L., Dubes, V ., Rogier, J., Labrousse, L., Bernus, O., Quesson, B., Haïssaguerre, M., Gsell, ...

    2022

  32. [40]

    A technique for measuring anisotropy in atrial conduction to estimate conduction velocity and atrial fibre direction

    Roney, C.H., Whitaker, J., Sim, I., O’Neill, L., Mukherjee, R.K., Razeghi, O., Vigmond, E.J., Wright, M., O’Neill, M.D., Williams, S.E., Niederer, S.A., 2019. A technique for measuring anisotropy in atrial conduction to estimate conduction velocity and atrial fibre direction. ...

    2019 doi

  33. [41]

    Fiber orientation and ejection fraction in the human left ventricle

    Sallin, E.A., 1969. Fiber orientation and ejection fraction in the human left ventricle. Biophysical Journal 9, 954–964. doi:10.1016/S0006-3495(69)86429-5

    1969 doi

  34. [42]

    Histological validation of myocardial mi- crostructure obtained from diffusion tensor magnetic resonance imaging

    Scollan, D.F., Holmes, A., Winslow, R., Forder, J., 1998. Histological validation of myocardial mi- crostructure obtained from diffusion tensor magnetic resonance imaging. American Journal of Physiology 275, H2308–H2318. URL:https://doi.org/10.1152/ajpheart.1998.275.6.H2308, d...

    1998 doi

  35. [43]

    Reconstruction of cardiac ventricular geometry and fiber orientation using magnetic resonance imaging

    Scollan, D.F., Holmes, A., Zhang, J., Winslow, R.L., 2000. Reconstruction of cardiac ventricular geometry and fiber orientation using magnetic resonance imaging. Annals of Biomedical Engineering 28, 934–944. URL: https://doi.org/10.1114/1.1312188, doi:10.1114/1.1312188

    2000 doi

  36. [44]

    Defini- tion of left ventricular segments for cardiac magnetic resonance imaging

    Selvadurai, B.S., Puntmann, V .O., Bluemke, D.A., Ferrari, V .A., Friedrich, M.G., Kramer, C.M., Kwong, R.Y ., Lombardi, M., Prasad, S.K., Rademakers, F.E., Young, A.A., Kim, R.J., Nagel, E., 2018. Defini- tion of left ventricular segments for cardiac magnetic resonance imagin...

    2018 doi

  37. [45]

    Pulmonary hypertension and indicators of right ventricular function

    von Siebenthal, C., Aubert, J.D., Mitsakis, P., Yerly, P., Prior, J.O., Nicod, L.P., 2016. Pulmonary hypertension and indicators of right ventricular function. Frontiers in Medicine 3, 23. URL:http://dx.doi.org/10. 3389/fmed.2016.00023, doi:10.3389/fmed.2016.00023

    2016

  38. [46]

    Soloveva, A., Gale, C.P., Han, N.T., Hurdus, B., Aktaa, S., Palin, V ., Mebrahtu, T.F., Van Spall, H., Batra, G., Dondo, T.B., Bäck, M., Munyombwe, T., 2023. Associations of health-related quality of life with major adverse cardiovascular and cerebrovascular events for individ...

    2023 doi

  39. [47]

    Steen, H., Giusca, S., Montenbruck, M., Patel, A.R., Pieske, B., Florian, A., Erley, J., Kelle, S., Korosoglou, G., 2021. Left and right ventricular strain using fast strain-encoded cardiovascular magnetic resonance for the diagnostic classification of patients with chronic no...

    2021 doi

  40. [48]

    Chapter 1.2 - cardiovascular mechanics and disease, in: Niebur, G.L

    Sucosky, P., Shar, J.A., Barrientos, J., 2020. Chapter 1.2 - cardiovascular mechanics and disease, in: Niebur, G.L. (Ed.), Mechanobiology. Elsevier, pp. 23–45. URL:https://www.sciencedirect.com/science/article/ pii/B9780128179314000029, doi:10.1016/B978-0-12-817931-4.00002-9

    2020 doi

  41. [49]

    Global bur- den of cardiovascular diseases and its risk factors, 1990–2021: A systematic analysis for the global burden of disease study 2021

    Tan, S.C.W., Zheng, B.B., Tang, M.L., Chu, H., Zhao, Y .T., Weng, C., 2025. Global bur- den of cardiovascular diseases and its risk factors, 1990–2021: A systematic analysis for the global burden of disease study 2021. QJM: An International Journal of Medicine , hcaf022URL:htt...

    2025 doi

  42. [50]

    Fourier features let networks learn high frequency functions in low dimensional domains

    Tancik, M., Srinivasan, P., Mildenhall, B., Fridovich-Keil, S., Raghavan, N., Singhal, U., Ramamoorthi, R., Barron, J., Ng, R., 2020. Fourier features let networks learn high frequency functions in low dimensional domains. Advances in neural information processing systems 33, ...

    2020 doi

  43. [51]

    Tobon-Gomez, C., De Craene, M., McLeod, K., Tautz, L., Shi, W., Hennemuth, A., Prakosa, A., Wang, H., Carr- White, G., Kapetanakis, S., Lutz, A., Rasche, V ., Schaeffter, T., Butakoff, C., Friman, O., Mansi, T., Sermesant, M., Zhuang, X., Ourselin, S., Peitgen, H.O., Pennec, X...

  44. [52]

    Measurement of left ventricular wall thick- ness and mass by echocardiography

    Troy L., B.L., Pombo, J., Rackley, C.E., 1972. Measurement of left ventricular wall thick- ness and mass by echocardiography. Circulation 45, 602–611. URL:https://www. ahajournals.org/doi/abs/10.1161/01.CIR.45.3.602, doi:10.1161/01.CIR.45.3.602, arXiv:https://www.ahajournals.o...

    1972 doi

  45. [53]

    A model to determine the effect of collagen fiber alignment on heart function post myocardial infarction

    V oorhees, A.P., Han, H.C., 2014. A model to determine the effect of collagen fiber alignment on heart function post myocardial infarction. Theoretical Biology and Medical Modelling 11, 6. URL:https://doi.org/10. 1186/1742-4682-11-6, doi:10.1186/1742-4682-11-6

    2014 doi

  46. [54]

    On the eigenvector bias of fourier feature networks: From regression to solving multi-scale pdes with physics-informed neural networks

    Wang, S., Wang, H., Perdikaris, P., 2021. On the eigenvector bias of fourier feature networks: From regression to solving multi-scale pdes with physics-informed neural networks. Computer Methods in Applied Mechanics and Engineering 384, 113938. doi:10.48550/arXiv.2012.10047

    2021 doi arXiv

  47. [55]

    Normal reference values for mitral annular plane systolic excursion by motion-mode and speckle tracking echocardiography: a prospective, multicentre, population-based study

    Wang, Y .H., Sun, L., Li, S.W., Wang, C.F., Pan, X.F., Liu, Y ., Wu, J., Guan, X.P., Zhang, S.L., Dun, G.L., Liu, Y .L., Wang, L.Y ., Cui, L., Liu, Y ., Lai, Y .Q., Ding, M.Y ., Lu, G.L., Tan, J., Yang, X.J., Li, Y .H., Zhang, X.T., Fan, M., Yu, J.H., Zheng, Q.J., Ma, C.Y ., R...

  48. [56]

    Metabolic disorders and risk of cardiovascular diseases: a two-sample mendelian randomization study

    Wang, Z., Chen, J., Zhu, L., Jiao, S., Chen, Y ., Sun, Y ., 2023b. Metabolic disorders and risk of cardiovascular diseases: a two-sample mendelian randomization study. BMC Cardiovascular Disorders 23, 529. doi:10.1186/ s12872-023-03567-3

  49. [57]

    Generating fibre orientation maps in human heart models using poisson interpolation

    Wong, J., Kuhl, E., 2012. Generating fibre orientation maps in human heart models using poisson interpolation. Computer methods in biomechanics and biomedical engineering 17. doi:10.1080/10255842.2012.739167

    2012

  50. [58]

    Cardiovascular diseases (cvds) fact sheet.http://www.who.int/en/ news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)

    World Health Organization, 2025. Cardiovascular diseases (cvds) fact sheet.http://www.who.int/en/ news-room/fact-sheets/detail/cardiovascular-diseases-(cvds). Last accessed on 2025-05-03

    2025

  51. [59]

    Global burden of metabolic diseases, 1990–2021

    Zhang, H., Zhou, X.D., Shapiro, M.D., Lip, G.Y ., Tilg, H., Valenti, L., Somers, V .K., Byrne, C.D., Targher, G., Yang, W., ..., Zheng, M.H., 2024. Global burden of metabolic diseases, 1990–2021. Metabolism 160, 155999. URL:https://www.sciencedirect.com/science/article/pii/S00...

    2024

  52. [60]

    Numerical evaluation of myofiber orientation and transmural contractile strength on left ventricular function

    Zhang, X., Haynes, P., Campbell, K.S., Wenk, J.F., 2015. Numerical evaluation of myofiber orientation and transmural contractile strength on left ventricular function. Journal of Biomechanical Engineering 137, 044502. doi:10.1115/1.4028990. epub 2015 Feb 5

    2015 doi

  53. [61]

    Zhang, Y ., Wang, V .Y ., Morgan, A.E., Kim, J., Tafreshi, R., Wallace, A.W., Guccione, J.M., Weinsaft, J.W., Ge, L., Ratcliffe, M.B., 2021. Finite-element based optimization of left ventricular passive stiff- ness in normal volunteers and patients after myocardial infarction:...

    2021

  54. [62]

    Heart-muscle fiber reconstruction from diffusion tensor mri, in: IEEE Visualization,

    Zhukov, L., Barr, A., 2003. Heart-muscle fiber reconstruction from diffusion tensor mri, in: IEEE Visualization,

    2003

  55. [1367]

    1016/S0735-1097(97)00276-3

    URL:https://www.sciencedirect.com/science/article/pii/S0735109797002763, doi:10. 1016/S0735-1097(97)00276-3

  56. [2003]

    VIS 2003., pp. 597–602. doi:10.1109/VISUAL.2003.1250425. Appendix A. Synthetic images For the synthetic example, we generate a reference image of the cylinder at rest and a template image of the cylinder under compression and torsion. The top row of Figure A.9 shows a cut in t...

    2003 arXiv

  57. [2013]

    Medical Image Analysis 17, 632–648

    Benchmarking framework for myocardial tracking and deformation algorithms: An open access database. Medical Image Analysis 17, 632–648. URL:https://www.sciencedirect.com/science/article/pii/ S1361841513000388, doi:10.1016/j.media.2013.03.008

    2013 doi

  58. [2021]

    3389/fcvm.2021.730316, doi:10.3389/fcvm.2021.730316

    URL:https://www.frontiersin.org/journals/cardiovascular-medicine/articles/10. 3389/fcvm.2021.730316, doi:10.3389/fcvm.2021.730316

    2021

  59. [2022]

    URL:https://www.sciencedirect

    Journal of the American College of Cardiology 82, 2350–2473. URL:https://www.sciencedirect. com/science/article/pii/S0735109723080233, doi:10.1016/j.jacc.2023.11.007. global Burden of Cardiovascular Diseases and Risk Factors, 1990–2022

    2023 doi

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.