Abstract
Cardiovascular diseases remain the leading cause of global mortality. Patient-specific 3D cardiac models inform pre-operative planning and personalized simulations, yet bi-atrial mesh reconstruction from routine cardiac MRI (CMR) is challenging due to sparse 2D coverage. We present an automated pipeline for reconstructing 3D bi-atrial surface meshes from sparse 2-chamber and 4-chamber long-axis CMR segmen- tations. An augmented Gaussian Process (GP) prior combines a data- driven PCA covariance with a radial basis function kernel to extend shape flexibility beyond the training distribution. A heteroscedastic observation model assigns distinct noise levels to contour points and mesh- level regularization constraints, with patient-specific meshes inferred using GP-Iterative Closest Point (GP-ICP) fitting. Evaluated on 500 UK Biobank subjects, the pipeline completes in <1 min per subject on a single CPU core. Reconstructed meshes showed comparable inter-patient shape variance on the test set (pairwise mean surface distance: 4.42 ± 1.40 mm vs. 5.30 ± 2.02 mm for raw contours). Reconstructed left and right atrial volumes (LA: 118.74 ± 15.33 mL; RA: 79.27 ± 10.42 mL) are within clinical ranges reported for atrial fibrillation cohorts. This work enables fast, population-scale bi-atrial reconstruction from sparse routine CMR acquisitions. Code and models will be made publicly available.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/STACOM2026_012.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/profile?id=~Semhar_Misghina1
BibTex
@InProceedings{MisSem_Automated_MICCAISAT2026,
author = { Misghina, Semhar B. AND Cheng, Fuyu AND Chen, Yang AND Slabaugh, Gregory AND Munroe, Patricia AND Petersen, Steffen E. AND Aung, Nay AND Roney, Caroline H.},
title = { { Automated Reconstruction of Patient-Specific 3D Bi-Atrial Meshes from Sparse 2D Cine CMR } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17267},
month = {pending},
page = {pending}
}