A Geometric Deep Learning Framework for Generation of Virtual Left Ventricles as Graphs

Soodeh Kalaie, Andrew J. Bulpitt, Alejandro F. Frangi, Ali Gooya
Medical Imaging with Deep Learning, PMLR 227:426-443, 2024.

Abstract

Generative statistical models have a wide range of applications in the modelling of anatomies. In-silico clinical trials of medical devices, for instance, require the development of virtual populations of anatomy that capture enough variability while remaining plausible. Model construction and use are heavily influenced by the correspondence problem and establishing shape matching over a large number of training data.This study focuses on generating virtual cohorts of left ventricle geometries resembling different-sized shape populations, suitable for in-silico experiments. We present an unsupervised data-driven probabilistic generative model for shapes. This framework incorporates an attention-based shape matching procedure using graph neural networks, coupled with a $\beta-$VAE generation model, eliminating the need for initial shape correspondence. Left ventricle shapes derived from cardiac magnetic resonance images available in the UK Biobank are utilized for training and validating the framework. We investigate our method’s generative capabilities in terms of generalisation and specificity and show that it is able to synthesise virtual populations of realistic shapes with volumetric measurements in line with actual clinical indices. Moreover, results show our method outperforms joint registration-PCA-based models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v227-kalaie24a, title = {A Geometric Deep Learning Framework for Generation of Virtual Left Ventricles as Graphs}, author = {Kalaie, Soodeh and Bulpitt, Andrew J. and Frangi, Alejandro F. and Gooya, Ali}, booktitle = {Medical Imaging with Deep Learning}, pages = {426--443}, year = {2024}, editor = {Oguz, Ipek and Noble, Jack and Li, Xiaoxiao and Styner, Martin and Baumgartner, Christian and Rusu, Mirabela and Heinmann, Tobias and Kontos, Despina and Landman, Bennett and Dawant, Benoit}, volume = {227}, series = {Proceedings of Machine Learning Research}, month = {10--12 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v227/kalaie24a/kalaie24a.pdf}, url = {https://proceedings.mlr.press/v227/kalaie24a.html}, abstract = {Generative statistical models have a wide range of applications in the modelling of anatomies. In-silico clinical trials of medical devices, for instance, require the development of virtual populations of anatomy that capture enough variability while remaining plausible. Model construction and use are heavily influenced by the correspondence problem and establishing shape matching over a large number of training data.This study focuses on generating virtual cohorts of left ventricle geometries resembling different-sized shape populations, suitable for in-silico experiments. We present an unsupervised data-driven probabilistic generative model for shapes. This framework incorporates an attention-based shape matching procedure using graph neural networks, coupled with a $\beta-$VAE generation model, eliminating the need for initial shape correspondence. Left ventricle shapes derived from cardiac magnetic resonance images available in the UK Biobank are utilized for training and validating the framework. We investigate our method’s generative capabilities in terms of generalisation and specificity and show that it is able to synthesise virtual populations of realistic shapes with volumetric measurements in line with actual clinical indices. Moreover, results show our method outperforms joint registration-PCA-based models.} }
Endnote
%0 Conference Paper %T A Geometric Deep Learning Framework for Generation of Virtual Left Ventricles as Graphs %A Soodeh Kalaie %A Andrew J. Bulpitt %A Alejandro F. Frangi %A Ali Gooya %B Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ipek Oguz %E Jack Noble %E Xiaoxiao Li %E Martin Styner %E Christian Baumgartner %E Mirabela Rusu %E Tobias Heinmann %E Despina Kontos %E Bennett Landman %E Benoit Dawant %F pmlr-v227-kalaie24a %I PMLR %P 426--443 %U https://proceedings.mlr.press/v227/kalaie24a.html %V 227 %X Generative statistical models have a wide range of applications in the modelling of anatomies. In-silico clinical trials of medical devices, for instance, require the development of virtual populations of anatomy that capture enough variability while remaining plausible. Model construction and use are heavily influenced by the correspondence problem and establishing shape matching over a large number of training data.This study focuses on generating virtual cohorts of left ventricle geometries resembling different-sized shape populations, suitable for in-silico experiments. We present an unsupervised data-driven probabilistic generative model for shapes. This framework incorporates an attention-based shape matching procedure using graph neural networks, coupled with a $\beta-$VAE generation model, eliminating the need for initial shape correspondence. Left ventricle shapes derived from cardiac magnetic resonance images available in the UK Biobank are utilized for training and validating the framework. We investigate our method’s generative capabilities in terms of generalisation and specificity and show that it is able to synthesise virtual populations of realistic shapes with volumetric measurements in line with actual clinical indices. Moreover, results show our method outperforms joint registration-PCA-based models.
APA
Kalaie, S., Bulpitt, A.J., Frangi, A.F. & Gooya, A.. (2024). A Geometric Deep Learning Framework for Generation of Virtual Left Ventricles as Graphs. Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 227:426-443 Available from https://proceedings.mlr.press/v227/kalaie24a.html.

Related Material