3D deformable registration of longitudinal abdominopelvic CT images using unsupervised deep learning

Maureen van Eijnatten (Corresponding author), Leonardo Rundo, K. Joost Batenburg, Felix Lucka, Emma Beddowes, Carlos Caldas, Ferdia A. Gallagher, Evis Sala, Carola Bibiane Schönlieb, Ramona Woitek

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

9 Citaten (Scopus)
216 Downloads (Pure)

Samenvatting

This study investigates the use of the unsupervised deep learning framework VoxelMorph for deformable registration of longitudinal abdominopelvic CT images acquired in patients with bone metastases from breast cancer. The CT images were refined prior to registration by automatically removing the CT table and all other extra-corporeal components. To improve the learning capabilities of VoxelMorph when only a limited amount of training data is available, a novel incremental training strategy is proposed based on simulated deformations of consecutive CT images. In a 4-fold cross-validation scheme, the incremental training strategy achieved significantly better registration performance compared to training on a single volume. Although our deformable image registration method did not outperform iterative registration using NiftyReg (considered as a benchmark) in terms of registration quality, the registrations were approximately 300 times faster. This study showed the feasibility of deep learning based deformable registration of longitudinal abdominopelvic CT images via a novel incremental training strategy based on simulated deformations.
Originele taal-2Engels
Artikelnummer106261
Aantal pagina's14
TijdschriftComputer Methods and Programs in Biomedicine
Volume208
DOI's
StatusGepubliceerd - sep. 2021

Bibliografische nota

Funding Information:
This work was also partially supported by The Mark Foundation for Cancer Research and Cancer Research UK Cambridge Centre [C9685/A25177], the CRUK National Cancer Imaging Translational Accelerator (NCITA) [C42780/A27066], and the Wellcome Trust Innovator Award [RG98755]. Additional support was also provided by the National Institute of Health Research (NIHR) Cambridge Biomedical Research Centre [BRC-1215-20014] and the Cambridge Mathematics of Information in Healthcare (CMIH) [funded by the EPSRC grant EP/T017961/1]. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care.

Funding Information:
CBS acknowledges support from the Leverhulme Trust project on ‘Breaking the non-convexity barrier’, EPSRC grant Nr. EP/M00483X/1, EP/S026045/1, the EPSRC Centre Nr. EP/N014588/1, the RISE projects CHiPS and NoMADS, the Cantab Capital Institute for the Mathematics of Information and the Alan Turing Institute.

Funding Information:
MvE acknowledges financial support from the Netherlands Organisation for Scientific Research (NWO) [project number 639.073.506] and the Royal Dutch Academy of Sciences (KNAW) [Van Leersum Grant 2018]. This work was also partially supported by The Mark Foundation for Cancer Research and Cancer Research UK Cambridge Centre [C9685/A25177], the CRUK National Cancer Imaging Translational Accelerator (NCITA) [C42780/A27066], and the Wellcome Trust Innovator Award [RG98755]. Additional support was also provided by the National Institute of Health Research (NIHR) Cambridge Biomedical Research Centre [BRC-1215-20014] and the Cambridge Mathematics of Information in Healthcare (CMIH) [funded by the EPSRC grant EP/T017961/1]. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care. CBS acknowledges support from the Leverhulme Trust project on ?Breaking the non-convexity barrier?, EPSRC grant Nr. EP/M00483X/1, EP/S026045/1, the EPSRC Centre Nr. EP/N014588/1, the RISE projects CHiPS and NoMADS, the Cantab Capital Institute for the Mathematics of Information and the Alan Turing Institute. Microsoft Radiomics was provided to the Addenbrooke's Hospital (Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK) by the Microsoft InnerEye project.

Funding Information:
MvE acknowledges financial support from the Netherlands Organisation for Scientific Research (NWO) [project number 639.073.506] and the Royal Dutch Academy of Sciences (KNAW) [Van Leersum Grant 2018].

Publisher Copyright:
© 2021 The Author(s)

Financiering

FinanciersFinanciernummer
CRUK National Cancer Imaging Translational Accelerator
Cancer Research UKC9685/A25177
Cantab Capital Institute for the Mathematics of Information
Microsoft InnerEye
NCITAC42780/A27066
Royal Dutch Academy of Sciences
Wellcome TrustRG98755
Alan Turing Institute
Mark Foundation For Cancer Research
Engineering and Physical Sciences Research CouncilEP/M00483X/1, EP/N014588/1, EP/T017961/1
National Institute for Health Research
Leverhulme TrustEP/S026045/1
Koninklijke Nederlandse Akademie van Wetenschappen
Nederlandse Organisatie voor Wetenschappelijk Onderzoek639.073.506
UCLH Biomedical Research CentreBRC-1215-20014

    Vingerafdruk

    Duik in de onderzoeksthema's van '3D deformable registration of longitudinal abdominopelvic CT images using unsupervised deep learning'. Samen vormen ze een unieke vingerafdruk.

    Citeer dit