Pith. sign in

REVIEW 1 major objections 7 minor 12 references

4D-CTA Image and geometry dataset for kinematic analysis of abdominal aortic aneurysms

T0 review · 1 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper releases a public 4D-CTA dataset for abdominal aortic aneurysm kinematics, with ten patients' cardiac-gated CT images, geometries, and finite-element synthetic ground truth.

desk verdict Useful 4D-CTA AAA dataset with a solid real-image component, but the synthetic ground truth is built on a likely sign error in Eq. (1) that must be fixed before the benchmark is trustworthy. read the letter →

arxiv 2505.17647 v2 pith:L44WB5YJ submitted 2025-05-23 cs.CE

Mostafa Jamshidian , Adam Wittek , Saeideh Sekhavat , Farah Alkhatib , Jens Carsten Ritter , Paul M. Parizel , Donatien Le Liepvre , Florian Bernard
show 5 more authors
Ludovic Minvielle Antoine Fondanèche Jane Polce Christopher Wood Karol Miller
This is my paper · ORCID
classification cs.CE
keywords AbdominalaorticaneurysmPatient-specificanalysisWalldisplacementstrainImageregistrationComputedtomographyangiographyBiomechanicsNon-invasivemethod
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 establishes a dataset rather than a new measurement or model: it provides, to the authors' knowledge, the first publicly available time-resolved CT angiography (4D-CTA) resource for abdominal aortic aneurysm (AAA). The data cover ten patients, with ECG-gated 3D-CTA frames sampled across the cardiac cycle, patient-specific external-wall geometries, and a synthetic ground-truth package built from a finite element simulation of one patient's aneurysm. A sympathetic reader would care because the dataset lets independent groups develop and test non-invasive, image-based methods for measuring AAA wall displacement and strain, an area that has received far less attention than wall-stress analysis. The paper also supplies the materials needed to reproduce the kinematic results reported in its companion biomechanics study.

What carries the argument

The mechanism that carries the argument is the pairing of real multi-phase imaging with a synthetic known-deformation target. For each patient, ECG-gated CTA frames are acquired at ten-percent intervals of the R-R interval (with some patients having fewer phases), and the systolic frame is segmented through an automated pipeline to produce a triangulated external-wall surface. For the ground truth, a finite element simulation on a hexahedral mesh supplies a displacement field for Patient 1; a B-spline scattered transform built from the FE nodal coordinates warps the diastolic image into a synthetic systolic image, so the true wall displacement is known by construction.

What would settle it

Download the dataset and run an independent image-registration algorithm on Patient 1's diastolic and synthetic-systolic images; if the recovered displacement field does not match the provided FE nodal displacements within a few voxels, the ground truth's internal consistency is called into question. Separately, a systematic search of public repositories for any earlier public 4D-CTA AAA dataset would directly test the 'first publicly available' claim.

Watch

Extended reading notes

Core claim

The paper's central claim is that this collection—4D-CTA images and AAA geometries for ten patients plus a finite-element-derived synthetic systolic image with known wall displacements—constitutes a reusable benchmark for AAA kinematic analysis. To the authors' knowledge it is the first publicly available 4D-CTA resource for AAA. The ground-truth portion is generated by creating a patient-specific hexahedral finite element model of Patient 1's AAA, applying a uniform 13 kPa internal pressure with rigidly constrained ends and literature-based hyperelastic wall parameters, and using the computed nodal displacement field to warp the diastolic-phase CTA image into a synthetic systolic image. The stated purpose is method verification: image-registration algorithms can be checked against known displacements, and the companion study's results can be reproduced.

Load-bearing premise

The load-bearing premise is that a finite element model using uniform 13 kPa pressure, rigidly constrained ends, and literature-based tissue parameters produces a 'physiologically plausible' displacement field good enough to stand in for real AAA wall motion when verifying new methods; if that simplification fails to represent in-vivo motion, algorithms tuned to this benchmark may not transfer to patients.

Editorial extensions

If this is right

  • Anyone can download the dataset and reproduce the wall-displacement and strain results of the companion study [1].
  • The synthetic ground truth gives registration and strain-measurement methods a quantitative target to be checked against, without requiring invasive markers.
  • Because the images are stored as NRRD and geometries as STL, the data can be used across different open-source and commercial analysis platforms.
  • The pipeline's reliance on standard clinical CT scanners means the approach could be applied wherever ECG-gated CT angiography is available.

Reading between the lines

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

  • This benchmark could make AAA wall strain a practical endpoint for multi-center studies, in the way that public datasets have accelerated other image-analysis fields.
  • The simplified FE ground truth (uniform pressure, rigid ends, literature material constants) is a simplification; a natural extension would be to generate multiple synthetic frames across the cardiac cycle and to vary the FE assumptions, so methods can be stress-tested against realistic modeling uncertainty.
  • The irregular temporal sampling between patients (10, 7, or only 2 phases) is itself a stress test: methods validated on this dataset will need to handle sparse and uneven cardiac-phase coverage, which mirrors real clinical variability.
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

1 major / 7 minor

Summary. This data descriptor presents a publicly available dataset (Zenodo DOI 10.5281/zenodo.15477710) of ECG-gated 4D-CTA image sequences and patient-specific AAA wall geometries for ten patients, together with a synthetic ground-truth package derived from Patient 1's diastolic image: a patient-specific finite element model, nodal displacement field, image-warping transform, and a synthetic systolic 3D-CTA image. The stated purpose is to enable reproduction of the companion kinematics study [1] and to provide a benchmark for non-invasive estimation of AAA wall displacement and strain.

Significance. If the ground truth is physically sound, the dataset fills a clear gap in public resources for AAA kinematic analysis; the open Zenodo deposition, standard NRRD/STL/inp/VTK formats, and explicit file inventory are genuine strengths that should make the resource broadly reusable. The paper also gives a detailed account of the segmentation and geometry extraction pipeline. However, the apparent sign error in the strain-energy function in Eq. (1) currently undermines the central verification claim, because the synthetic ground truth is the only part of the dataset explicitly intended as a benchmark for method verification. This must be resolved before the dataset can be relied upon as a benchmark.

major comments (1)
  1. [Ground Truth, Eq. (1)] Equation (1) prints W = α(I_B−3) − β(I_B−3)^2 with α = 0.174 MPa and β = 1.881 MPa. With the minus sign, dW/dI_B = α − 2β(I_B−3), which becomes negative once I_B−3 exceeds α/(2β) ≈ 0.046, so the strain energy is non-convex and the material model is not a stable hyperelastic solid. The cited Raghavan–Vorp model [10] uses a plus sign: W = α(I_B−3) + β(I_B−3)^2. Thus either the Abaqus simulation was run with the printed nonphysical model, making the synthetic ground truth non-physiological and unsuitable as a benchmark, or it was run with the correct positive sign, making Eq. (1) a misdescription that breaks the reproducibility claim for the ground truth. Please correct the sign, state explicitly which material definition was used in Abaqus, and confirm that the FE displacement field was generated with a convex, stable material model.
minor comments (7)
  1. [Background] The claim 'To the best of our knowledge, this dataset is the first publicly available 4D-CTA resource for AAA' is not supported by a systematic literature search; please either provide the search strategy or soften the statement to avoid an unsupported novelty claim.
  2. [Ground Truth] The abbreviation I_B should be defined as the first invariant of the left Cauchy–Green tensor at first use, since it appears in Eq. (1) before any definition is given.
  3. [Experimental Design, Materials and Methods] There is a typo in the Segmentation subsection: 'MATALB' should be 'MATLAB'.
  4. [Data Description] Per-patient acquisition parameters (voxel spacing, matrix size, reconstruction kernel, contrast protocol, and phase availability) are not tabulated; providing them would materially improve reproducibility and reusability of the image data.
  5. [Ground Truth] The mesh statistics are limited to element type C3D20H; reporting element and node counts and a mesh-sensitivity or convergence check would strengthen confidence in the FE displacement field used as ground truth.
  6. [Ground Truth] The FDA and ASME guidelines invoked to justify plausibility of the synthetic ground truth are not cited; please provide the specific guidance documents.
  7. [Abstract and Data Description] The statement that each 4D-CTA dataset 'typically' contains ten frames is followed by the information that patients 1, 2, 4, and 8 have fewer frames; please clarify whether missing phases are absent from the released dataset or simply not used in the companion study.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the dataset and synthetic ground truth are self-contained inputs rather than derived predictions.

full rationale

This is a data descriptor, not a derivation. The central claim is that a 4D-CTA AAA dataset is publicly available and is, to the authors' knowledge, the first such resource; that claim is a factual availability statement independent of any fitted model or companion derivation. The synthetic ground truth is constructed from a patient-specific finite element model with literature material parameters (alpha = 0.174 MPa, beta = 1.881 MPa), rigid end constraints, and a prescribed 13 kPa internal pressure; the resulting nodal displacement field is applied to the diastolic CTA image through a B-spline scattered transform in 3D Slicer. The verification target is therefore not defined by the image-registration method being verified, and no parameter is fitted to the method's output. Citations to the companion paper [1] are descriptive references to the algorithm used in the related study, not load-bearing evidence for the dataset's existence or for the FE-based ground-truth construction. The apparent sign error in Eq. (1), if real, is a correctness and reproducibility concern about the FE model's physical stability and documentation, but it is not a circular dependence: the ground-truth displacement field is not equivalent to the method's inputs by construction. The paper's own limitation statement about single-site data further confirms the dataset is presented as an input resource rather than as a derived result.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim is the existence and structure of a dataset, so the main burden falls on the assumptions about data acquisition and processing fidelity. The listed axioms are the domain assumptions under which the dataset is fit for its stated purpose (benchmarking AAA kinematics). No free parameters are fitted to the verification target; all chosen numerical inputs are for generating the synthetic ground truth.

free parameters (2)
  • Uniform internal pressure for FE ground truth model = 13 kPa
    Chosen by hand as an assumed physiological loading condition for the FE model used to generate the synthetic systolic image; not fitted to any measured data.
  • Assumed AAA wall thickness (BioPARR) = not stated
    BioPARR requires an assumed wall thickness to generate the internal wall surface and FE mesh; this value is not given in the paper and directly affects the computed displacement field.
assumptions (4)
  • domain assumption The hyperelastic strain energy function W = alpha(I_B - 3) - beta(I_B - 3)^2 with alpha = 0.174 MPa and beta = 1.881 MPa (Eq. 1) describes AAA wall tissue behavior.
    These parameters are adopted from the cited literature (Raghavan-Vorp model), but the sign of the quadratic term appears inconsistent with the original model; the FE simulation in Abaqus uses whatever is in the input file, so the printed equation may be a typo.
  • domain assumption ECG-gated 4D-CTA images at 10% intervals of the R-R interval represent the cardiac cycle phases as labeled.
    The dataset relies on the scanner's ECG gating to attribute each 3D volume to a specific cardiac phase; no independent verification of phase labeling is provided.
  • domain assumption The FE model assumptions (rigid ends, uniform 13 kPa pressure, homogeneous wall, assumed thickness) produce a displacement field that is a plausible proxy for in-vivo AAA wall motion.
    The paper states that FDA/ASME guidelines require plausibility rather than exact fidelity, but no quantitative validation of this proxy is provided.
  • domain assumption The AI-based segmentation pipeline (PRAEVAorta, in-house post-processing, BioPARR) yields accurate patient-specific AAA geometries.
    The segmentation output is used to create the wall surface models, but its accuracy is not independently validated in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of 4D-CTA Image and geometry dataset for kinematic analysis of abdominal aortic aneurysms." pith.science (2026). https://pith.science/paper/L44WB5YJ

@misc{pith2026250517647,
  author       = {Pith},
  title        = {Pith review of: 4D-CTA Image and geometry dataset for kinematic analysis of abdominal aortic aneurysms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L44WB5YJ}},
  note         = {Machine review of arXiv:2505.17647}
}
read the original abstract

This article presents a dataset used in the article "Kinematics of Abdominal Aortic Aneurysms", published in the Journal of Biomechanics. The dataset is publicly available for download from the Zenodo data repository (https://doi.org/10.5281/zenodo.15477710). The dataset includes time-resolved 3D computed tomography angiography (4D-CTA) images of abdominal aortic aneurysm (AAA) captured throughout the cardiac cycle from ten patients diagnosed with AAA, along with ten patient-specific AAA geometries extracted from these images. Typically, the 4D-CTA dataset for each patient contains ten electrocardiogram (ECG)-gated 3D-CTA image frames acquired over a cardiac cycle, capturing both the systolic and diastolic phases of the AAA configuration. For method verification, the dataset also includes synthetic ground truth data generated from Patient 1's 3D-CTA AAA image in the diastolic phase. The ground truth data includes the patient-specific finite element (FE) biomechanical model and a synthetic systolic 3D-CTA image. The synthetic systolic image was generated by warping Patient 1's diastolic 3D-CTA image using the realistic displacement field obtained from the AAA biomechanical FE model. The images were acquired at Fiona Stanley Hospital in Western Australia and provided to the researchers at the Intelligent Systems for Medicine Laboratory at The University of Western Australia (ISML-UWA), where image-based AAA kinematic analysis was performed. Our dataset enabled the analysis of AAA wall displacement and strain throughout the cardiac cycle using a non-invasive, in vivo, image registration-based approach. The use of widely adopted, open-source file formats (NRRD for images and STL for geometries) facilitates broad applicability and reusability in AAA biomechanics studies that require patient-specific geometry and information about AAA kinematics during cardiac cycle.

Figures

Figures reproduced from arXiv: 2505.17647 by the authors.

Figure 1
Figure 1. Overview of the dataset directory structure and contents, consisting of ten patient folders (P1 to P10) and one “Ground Truth” folder. Each patient folder contains cropped 3D-CTA images of the patient’s abdominal aortic aneurysm (AAA) at multiple cardiac phases (e.g., 30%.nrrd, representing 30% of the cardiac cycle) and a triangulated surface model (Wall.stl) of the AAA external wall in the systolic phase. The “Grou… view at source ↗
Figure 3
Figure 3. Visualization of the synthetic ground truth data used for verifying AAA kinematic analysis methods: (a) The triangulated surface model of Patient 1’s AAA with an assumed wall thickness. (b) The hexahedral finite element (FE) mesh created in HyperMesh (https://altair.com/hypermesh) and imported into the Abaqus FE software (https://www.3ds.com/products/simulia/abaqus) for patient￾specific biomechanical modelling. (c) … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 8 canonical work pages

  1. [1]

    Jamshidian, A

    M. Jamshidian, A. Wittek, S. Sekhavat, K. Miller, Kinematics of abdominal aortic Aneurysms, J Biomech 179 (2025) 112484. https://doi.org/https://doi.org/10.1016/j.jbiomech.2024.112484

    arXiv 2025

  2. [10]

    Raghavan, D.A

    M.L. Raghavan, D.A. Vorp, Toward a biomechanical tool to evaluate rupture potential of abdominal aortic aneurysm: identification of a finite strain constitutive model and evaluation of its applicability, J Biomech 33(4) (2000) 475-482. https://doi.org/10.1016/S0021-9290(99)00201-8

    work page doi:10.1016/s0021-9290(99)00201-8 2000

  3. [2]

    Joldes, K

    G.R. Joldes, K. Miller, A. Wittek, R.O. Forsythe, D.E. Newby, B.J. Doyle, BioPARR: A software system for estimating the rupture potential index for abdominal aortic aneurysms, Scientific Reports 7(1) (2017) 4641. https://doi.org/10.1038/s41598-017-04699-1

    work page doi:10.1038/s41598-017-04699-1 2017

  4. [3]

    Fedorov, R

    A. Fedorov, R. Beichel, J. Kalpathy -Cramer, J. Finet, J.-C. Fillion-Robin, S. Pujol, C. Bauer, D. Jennings, F. Fennessy, M. Sonka, J. Buatti, S. Aylward, J.V. Miller, S. Pieper, R. Kikinis, 3D Slicer as an image computing platform for the Quantitative Imaging Network, Magn. Reson. Imaging 30(9) (2012) 1323 -1341. https://doi.org/http://dx.doi.org/10.1016...

    doi:10.1016/j.mri.2012.05.001 2012

  5. [4]

    Joldes, A

    G.R. Joldes, A. Wittek, S.K. Warfield, K. Miller, Performing Brain Image Warping Using the Deformation Field Predicted by a Biomechanical Model, Computational Biomechanics for Medicine: Deformation and Flow (2012) 89-96. https://doi.org/10.1007/978-1-4614-3172-5_10

    work page doi:10.1007/978-1-4614-3172-5_10 2012

  6. [5]

    Wanhainen, F

    A. Wanhainen, F. Verzini, I. Van Herzeele, E. Allaire, M. Bown, T. Cohnert, F. Dick, J. van Herwaarden, C. Karkos, M. Koelemay, T. Kölbel, I. Loftus, K. Mani, G. Melissano, J. Powell, Z. Szeberin, E.G. Committee, G.J. de Borst, N. Chakfe, S. Debus, R. Hinchliffe, S. Kakkos, I. Koncar, P . Kolh, J.S. Lindholt, M. de Vega, F. Vermassen, r. Document, M. Björ...

    work page 2019

  7. [6]

    Vorp, Biomechanics of abdominal aortic aneurysm, J Biomech 40(9) (2007) 1887 -1902

    D.A. Vorp, Biomechanics of abdominal aortic aneurysm, J Biomech 40(9) (2007) 1887 -1902. https://doi.org/10.1016/j.jbiomech.2006.09.003

    doi:10.1016/j.jbiomech.2006.09.003 2007

  8. [7]

    Generation of Patient-specific Structured Hexahedral Mesh of Aortic Aneurysm Wall

    F. Alkhatib, G.C. Bourantas, A. Wittek, K. Miller, Generation of Patient -specific Structured Hexahedral Mesh of Aortic Aneurysm Wall, in: M.P . Nash, A. Wittek, P .M.F. Nielsen, M. Kobielarz, A.R. Babu, K. Miller (Eds.), Computational Biomechanics for Medicine - Towards Automation and Robustness of Computations in the Clinic, Springer International Publi...

    work page Pith review arXiv doi:10.48550/arxiv.2206.06175 2023

Show all 12 references
  1. [8]

    Desjardins, E.A

    B. Desjardins, E.A. Kazerooni, ECG -gated cardiac CT, American Journal of Roentgenology 182(4) (2004) 993-1010. https://doi.org/https://doi.org/10.2214/ajr.182.4.1820993

    2004 doi

  2. [9]

    Valette, J.M

    S. Valette, J.M. Chassery, R. Prost, Generic remeshing of 3D triangular meshes with metric -dependent discrete Voronoi diagrams, Ieee T Vis Comput Gr 14(2) (2008) 369 -381. https://doi.org/10.1109/TVCG.2007.70430

    2008

  3. [11]

    Raghavan, D.A

    M.L. Raghavan, D.A. Vorp, M.P . Federle, M.S. Makaroun, M.W. Webster, Wall stress distribution on three- dimensionally reconstructed models of human abdominal aortic aneurysm, Journal of Vascular Surgery 31(4) (2000) 760-769. https://doi.org/10.1067/mva.2000.103971

    2000

  4. [12]

    Fillinger, M.L

    M.F. Fillinger, M.L. Raghavan, S.P . Marra, J.L. Cronenwett, F.E. Kennedy, In vivo analysis of mechanical wall stress and abdominal aortic aneurysm rupture risk, Journal of Vascular Surgery 36(3) (2002) 589 -597. https://doi.org/10.1067/mva.2002.125478

    2002 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.