Latent disentanglement in mesh variational autoencoders improves the diagnosis of craniofacial syndromes and aids surgical planning

Simone Foti; Alexander J Rickart; Bongjin Koo; Eimear O' Sullivan; Lara S van de Lande; Athanasios Papaioannou; Roman Khonsari; Danail Stoyanov; N U Owase Jeelani; Silvia Schievano; David J Dunaway; Matthew J Clarkson

doi:10.1016/j.cmpb.2024.108395

Latent disentanglement in mesh variational autoencoders improves the diagnosis of craniofacial syndromes and aids surgical planning

Comput Methods Programs Biomed. 2024 Nov:256:108395. doi: 10.1016/j.cmpb.2024.108395. Epub 2024 Aug 26.

Affiliations

¹ Wellcome/EPSRC Centre for Interventional and Surgical Sciences, University College London, London, UK; Centre For Medical Image Computing, University College London, London, UK; Imperial College London, Department of Computing, London, UK. Electronic address: s.foti@cs.ucl.ac.uk.
² UCL Great Ormond Street Institute of Child Health and Craniofacial Unit, Great Ormond Street Hospital for Children, London, UK.
³ Wellcome/EPSRC Centre for Interventional and Surgical Sciences, University College London, London, UK; Centre For Medical Image Computing, University College London, London, UK; University of California, Santa Barbara, Department of Electrical & Computer Engineering, Santa Barbara, USA.
⁴ UCL Great Ormond Street Institute of Child Health and Craniofacial Unit, Great Ormond Street Hospital for Children, London, UK; Imperial College London, Department of Computing, London, UK.
⁵ Department of Oral and Maxillofacial Surgery, Erasmus Medical Center, Rotterdam, The Netherlands.
⁶ Department of Maxillofacial Surgery and Plastic Surgery, Necker - Enfants Malades University Hospital, Paris, France.
⁷ Wellcome/EPSRC Centre for Interventional and Surgical Sciences, University College London, London, UK; Centre For Medical Image Computing, University College London, London, UK.

PMID: 39213899
DOI: 10.1016/j.cmpb.2024.108395

Abstract

Background and objective: The use of deep learning to undertake shape analysis of the complexities of the human head holds great promise. However, there have traditionally been a number of barriers to accurate modelling, especially when operating on both a global and local level.

Methods: In this work, we will discuss the application of the Swap Disentangled Variational Autoencoder (SD-VAE) with relevance to Crouzon, Apert and Muenke syndromes. The model is trained on a dataset of 3D meshes of healthy and syndromic patients which was increased in size with a novel data augmentation technique based on spectral interpolation. Thanks to its semantically meaningful and disentangled latent representation, SD-VAE is used to analyse and generate head shapes while considering the influence of different anatomical sub-units.

Results: Although syndrome classification is performed on the entire mesh, it is also possible, for the first time, to analyse the influence of each region of the head on the syndromic phenotype. By manipulating specific parameters of the generative model, and producing procedure-specific new shapes, it is also possible to approximate the outcome of a range of craniofacial surgical procedures.

Conclusion: This work opens new avenues to advance diagnosis, aids surgical planning and allows for the objective evaluation of surgical outcomes. Our code is available at github.com/simofoti/CraniofacialSD-VAE.

Keywords: Craniofacial syndromes; Geometric deep learning; Latent disentanglement.

MeSH terms

Algorithms
Craniofacial Abnormalities* / classification
Craniofacial Abnormalities* / diagnosis
Craniofacial Abnormalities* / surgery
Deep Learning
Head
Humans
Imaging, Three-Dimensional
Syndrome