Predicting arrhythmia recurrence post-ablation in atrial fibrillation using explainable machine learning: Atrial meshes
Data files
Jul 22, 2025 version files 6.88 GB
-
holdout_001.tgz
85.12 MB
-
holdout_002.tgz
72.48 MB
-
holdout_003.tgz
70.57 MB
-
holdout_004.tgz
109.24 MB
-
holdout_005.tgz
113.14 MB
-
holdout_006.tgz
112.19 MB
-
holdout_007.tgz
71.70 MB
-
holdout_008.tgz
68.63 MB
-
holdout_009.tgz
67.28 MB
-
holdout_010.tgz
102.76 MB
-
holdout_011.tgz
110.60 MB
-
holdout_012.tgz
94.15 MB
-
holdout_013.tgz
73.61 MB
-
holdout_014.tgz
75.13 MB
-
holdout_015.tgz
67.94 MB
-
original_021.tgz
69.15 MB
-
original_022.tgz
41.75 MB
-
original_023.tgz
37.10 MB
-
original_024.tgz
61.34 MB
-
original_025.tgz
96.39 MB
-
original_026.tgz
77.54 MB
-
original_027.tgz
45.40 MB
-
original_028.tgz
111.96 MB
-
original_029.tgz
80.43 MB
-
original_030.tgz
55.30 MB
-
original_031.tgz
186.35 MB
-
original_032.tgz
49.83 MB
-
original_033.tgz
75.32 MB
-
original_034.tgz
83.07 MB
-
original_035.tgz
101.74 MB
-
original_036.tgz
110.37 MB
-
original_037.tgz
93.62 MB
-
original_038.tgz
80.03 MB
-
original_039.tgz
51.95 MB
-
original_040.tgz
81.37 MB
-
original_041.tgz
105.30 MB
-
original_042.tgz
98.73 MB
-
original_043.tgz
72.43 MB
-
original_044.tgz
69.05 MB
-
original_045.tgz
121.91 MB
-
original_046.tgz
56.54 MB
-
original_047.tgz
86.54 MB
-
original_048.tgz
60.73 MB
-
original_049.tgz
67.56 MB
-
original_050.tgz
79.49 MB
-
original_051.tgz
61.83 MB
-
original_052.tgz
85.07 MB
-
original_053.tgz
107.14 MB
-
original_054.tgz
88.16 MB
-
original_055.tgz
107.04 MB
-
original_056.tgz
91.59 MB
-
original_057.tgz
43.72 MB
-
original_058.tgz
78.52 MB
-
original_059.tgz
66.64 MB
-
original_060.tgz
99.41 MB
-
original_061.tgz
83.23 MB
-
original_062.tgz
62.51 MB
-
original_063.tgz
105.93 MB
-
original_064.tgz
64.54 MB
-
original_065.tgz
113.97 MB
-
original_066.tgz
125.88 MB
-
original_067.tgz
96.96 MB
-
original_068.tgz
76.51 MB
-
original_069.tgz
116.71 MB
-
original_070.tgz
69.43 MB
-
original_071.tgz
98.40 MB
-
original_072.tgz
99.72 MB
-
original_073.tgz
108.16 MB
-
original_074.tgz
103.88 MB
-
original_075.tgz
69.72 MB
-
original_076.tgz
73.67 MB
-
original_077.tgz
51.83 MB
-
original_078.tgz
96.29 MB
-
original_079.tgz
80.94 MB
-
original_080.tgz
77.14 MB
-
original_081.tgz
50.77 MB
-
original_082.tgz
67.74 MB
-
original_083.tgz
99.91 MB
-
original_084.tgz
102.57 MB
-
original_085.tgz
77.18 MB
-
original_086.tgz
83.07 MB
-
original_087.tgz
93.59 MB
Jul 29, 2025 version files 294.26 MB
-
holdout_cohort.tgz
51.69 MB
-
README.md
3.93 KB
-
training_cohort.tgz
242.56 MB
Abstract
This repository contains computational meshes of human left atria segmented from late gadolinium enhanced cardiac magnetic resonance imaging (LGE-MRI) for patients receiving a catheter ablation for atrial fibrillation. There are two models per patient (pre-ablation with patterns of fibrotic remodeling, post-ablation with scar created by the procedure) for 82 individuals. These meshes support the replication of our findings while protecting confidential patient information.
Dataset DOI: 10.5061/dryad.kkwh70sg0
These meshes are linked to an accepted paper which will be published around mid-August 2025. Once the paper is published and the embargo period has ended, a link will be added to this repository. If you use these models, please cite this repository our paper.
Description of the data and file structure
This repository contains computational meshes of human left atria segmented from late gadolinium enhanced cardiac magnetic resonance imaging (LGE-MRI) for patients receiving a catheter ablation for atrial fibrillation. There are two models per patient (pre-ablation with patterns of fibrotic remodeling, post-ablation with scar created by the procedure) for 82 individuals. These meshes support the replication of our findings while protecting confidential patient information.
Files and variables
Computational meshes are provided in VTK format. The same deidentified numbers used in the publication to identify the holdout patients correspond to the meshes for ID numbers 1-15 in this dataset. Patients in the original cohort were assigned ID numbers 21-87.
We used this format (vtk_bin
) in the interest of reproducibility. To reduce the total size of the repository, meshes have been downsampled to an average edge length of 0.5 mm; before running simulations using these meshes, please consider resampling to an appropriate resolution (e.g., ~0.2 mm for reaction-diffuction simulations of electrophysiology). These files can be converted to a machine- and human-readable text format compatible with the openCARP
simulation framework using the free software meshtool
(available here), e.g.:
$ /path/to/meshtool convert -imsh ID001_PostAbl -ifmt vtk_bin -omsh ID001_PostAbl -ofmt carp_txt
Once each mesh is converted from carp_bin
to carp_txt
format, the resulting files (.pts
, .elem
, .lon
) can be opened in a text editor and the explicit values of Cartesian coordinates, element specifications, and fiber orientation vectors can be seen.
Attached files are structured as follows.
- holdout_cohort.tgz
- ID001_PreAbl_rev1.vtk
- ID001_PostAbl_rev1.vtk
- ...
- ID015_PreAbl_rev1.vtk
- ID015_PostAbl_rev1.vtk
- training_cohort.tgz
- ID021_PreAbl_rev1.vtk
- ID021_PostAbl_rev1.vtk
- ...
- ID087_PreAbl_rev1.vtk
- ID087_PostAbl_rev1.vtk
Code/software
We used openCARP (https://opencarp.org/) [1] to run simulations. openCARP is open software licensed under the Academic Public License. Please read and abide by the relevant licenses if you choose to use these tools. More information about the licenses for both of these tools is available at this stable link: https://opencarp.org/download/license.
A good option for visualizing these models is NumeriCor Studio for Academia, which is freely available and has binary installers for Linux, Mac, and Windows. .vtk files can also be used with the free/open source Paraview software: https://www.paraview.org/.
References
- Plank, G. et al. The openCARP simulation environment for cardiac electrophysiology. Comput Methods Programs Biomed 208, 106223 (2021). https://doi.org/10.1016/j.cmpb.2021.106223
Human subjects data
This repository contains computational meshes of human left atria segmented from late gadolinium enhanced cardiac magnetic resonance imaging. These meshes support the replication of our findings while protecting confidential patient information. This study was approved by the University of Washington Institutional Review Board. All participants provided written informed consent.
This study retrospectively included patients from University of Washington (UW) Medical Center with documented persistent atrial fibrillation (AFib) or paroxysmal AFib who had already received both pre- and post-procedural LGE-MRI scans and underwent either cryoballoon or radiofrequency (RF) ablation. Cardiac late gadolinium enhanced magnetic resonance images (LGE-MRI) were obtained using previously described protocols for all participants within 90 days prior to their ablation procedure and again 3-6 months post-ablation to quantify the extent of LA fibrosis and scar, respectively. Exclusion criteria for AFib patients included those who had a prior catheter ablation, patients with cardiac implantable electronic devices, severe claustrophobia, renal dysfunction, and contraindications to MRI or gadolinium-based contrast. Scans were performed on the Philips Ingenia system, 15–25 min after contrast injection, using a three-dimensional inversion-recovery, respiration-navigated, ECG-gated, gradient echo pulse sequence. Acquisition parameters included transverse imaging volume with a voxel size of 1.25 × 1.25 × 2.5 mm (reconstructed to 0.625 × 0.625 × 1.25 mm). Scan time was 5–10 minutes dependent on respiration and heart rate.
Patients had clinical assessment and catheter ablation in the UW AFib program. All patients underwent pulmonary vein isolation (PVI), and some had additional substrate modification at the operator’s discretion.
Geometric models were reconstructed from both pre- and post- ablation LGE-MRI scans by Merisight Inc. (Salt Lake City, UT) to assess LA volume and surface area. Geometric models were reconstructed from pre-ablation scans and the relative extent of fibrosis in the LA was quantified via an adaptive histogram thresholding algorithm to determine pre-ablation LGE-MRI derived fibrosis. For post-ablation models, ablation scar was quantified on post-ablation LGE-MRI using previously established methods. Non-rigid registration was used to map LGE-derived post-ablation scar patterns onto existing LA pre-ablation fibrotic models. Hyper-enhancement on post-ablation scans was assumed to be ablation-induced scar; this accounts for the fact that hyperenhancement from ablation scar is at a higher absolute level than that of native fibrosis. Consequently, regions labeled as fibrotic pre-ablation fall below the hyperenhancement threshold in post-procedure scans.