Altmetric

Deep learning cardiac motion analysis for human survival prediction

File Description SizeFormat 
bello NMI.pdfAccepted version5.77 MBAdobe PDFView/Open
Title: Deep learning cardiac motion analysis for human survival prediction
Authors: Bello, G
Dawes, T
Duan, J
Biffi, C
Simoes Monteiro de Marvao, A
Howard, L
Gibbs, S
Wilkins, M
Cook, S
Rueckert, D
O'Regan, D
Item Type: Journal Article
Abstract: Motion analysis is used in computer vision to understand the behaviour of moving objects in sequences of images. Optimizing the interpretation of dynamic biological systems requires accurate and precise motion tracking as well as efficient representations of high-dimensional motion trajectories so that these can be used for prediction tasks. Here we use image sequences of the heart, acquired using cardiac magnetic resonance imaging, to create time-resolved three-dimensional segmentations using a fully convolutional network trained on anatomical shape priors. This dense motion model formed the input to a supervised denoising autoencoder (4Dsurvival), which is a hybrid network consisting of an autoencoder that learns a task-specific latent code representation trained on observed outcome data, yielding a latent representation optimized for survival prediction. To handle right-censored survival outcomes, our network used a Cox partial likelihood loss function. In a study of 302 patients, the predictive accuracy (quantified by Harrell’s C-index) was significantly higher (P = 0.0012) for our model C = 0.75 (95% CI: 0.70–0.79) than the human benchmark of C = 0.59 (95% CI: 0.53–0.65). This work demonstrates how a complex computer vision task using high-dimensional medical image data can efficiently predict human survival.
Issue Date: 11-Feb-2019
Date of Acceptance: 8-Jan-2019
URI: http://hdl.handle.net/10044/1/66986
DOI: https://dx.doi.org/10.1038/s42256-019-0019-2
ISSN: 2522-5839
Publisher: Nature Research
Start Page: 95
End Page: 104
Journal / Book Title: Nature Machine Intelligence
Volume: 1
Copyright Statement: © The Author(s), under exclusive licence to Springer Nature Limited 2019
Sponsor/Funder: British Heart Foundation
Engineering & Physical Science Research Council (EPSRC)
Imperial College London
Imperial College Healthcare NHS Trust- BRC Funding
British Heart Foundation
Imperial College Healthcare NHS Trust- BRC Funding
Funder's Grant Number: RE/08/002/23906
EP/P001009/1
RDC04
NH/17/1/32725
RDB02
Keywords: cs.LG
cs.CV
stat.ML
Publication Status: Published
Appears in Collections:Computing
Clinical Sciences
Imaging Sciences
National Heart and Lung Institute
Molecular Sciences
Department of Medicine
Faculty of Medicine



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Creative Commonsx