Integrating uncertainty in deep neural networks for MRI based stroke analysis

Herzog, Lisa; Murina, Elvis; Dürr, Oliver; Wegener, Susanne; Sick, Beate

doi:10.1016/j.media.2020.101790

Publication type:	Article in scientific journal
Type of review:	Peer review (publication)
Title:	Integrating uncertainty in deep neural networks for MRI based stroke analysis
Authors:	Herzog, Lisa Murina, Elvis Dürr, Oliver Wegener, Susanne Sick, Beate
et. al:	No
DOI:	10.1016/j.media.2020.101790
Published in:	Medical Image Analysis
Volume(Issue):	65
Issue:	101790
Issue Date:	Oct-2020
Publisher / Ed. Institution:	Elsevier
ISSN:	1361-8415 1361-8423
Language:	English
Subjects:	Bayesian convolutional neural networks; Ischemic stroke; Magnetic resonance imaging; Uncertainty; Bayes Theorem; Humans; Magnetic Resonance Imaging; Reproducibility of Results; Neural Networks, Computer; Stroke; eess.IV; Computer Science - Computer Vision and Pattern Recognition; Computer Science - Learning; Quantitative Biology - Quantitative Methods; Statistics - Machine Learning
Subject (DDC):	006: Special computer methods
Abstract:	At present, the majority of the proposed Deep Learning (DL) methods provide point predictions without quantifying the model's uncertainty. However, a quantification of the reliability of automated image analysis is essential, in particular in medicine when physicians rely on the results for making critical treatment decisions. In this work, we provide an entire framework to diagnose ischemic stroke patients incorporating Bayesian uncertainty into the analysis procedure. We present a Bayesian Convolutional Neural Network (CNN) yielding a probability for a stroke lesion on 2D Magnetic Resonance (MR) images with corresponding uncertainty information about the reliability of the prediction. For patient-level diagnoses, different aggregation methods are proposed and evaluated, which combine the individual image-level predictions. Those methods take advantage of the uncertainty in the image predictions and report model uncertainty at the patient-level. In a cohort of 511 patients, our Bayesian CNN achieved an accuracy of 95.33% at the image-level representing a significant improvement of 2% over a non-Bayesian counterpart. The best patient aggregation method yielded 95.89% of accuracy. Integrating uncertainty information about image predictions in aggregation models resulted in higher uncertainty measures to false patient classifications, which enabled to filter critical patient diagnoses that are supposed to be closer examined by a medical doctor. We therefore recommend using Bayesian approaches not only for improved image-level prediction and uncertainty estimation but also for the detection of uncertain aggregations at the patient-level.
URI:	https://arxiv.org/ftp/arxiv/papers/2008/2008.06332.pdf https://digitalcollection.zhaw.ch/handle/11475/30395
Fulltext version:	Published version
License (according to publishing contract):	Licence according to publishing contract
Departement:	School of Engineering
Organisational Unit:	Institute of Data Analysis and Process Design (IDP)
Appears in collections:	Publikationen School of Engineering

Files in This Item:

There are no files associated with this item.

Show full item record

Herzog, L., Murina, E., Dürr, O., Wegener, S., & Sick, B. (2020). Integrating uncertainty in deep neural networks for MRI based stroke analysis. Medical Image Analysis, 65(101790). https://doi.org/10.1016/j.media.2020.101790

Herzog, L. et al. (2020) ‘Integrating uncertainty in deep neural networks for MRI based stroke analysis’, Medical Image Analysis, 65(101790). Available at: https://doi.org/10.1016/j.media.2020.101790.

L. Herzog, E. Murina, O. Dürr, S. Wegener, and B. Sick, “Integrating uncertainty in deep neural networks for MRI based stroke analysis,” Medical Image Analysis, vol. 65, no. 101790, Oct. 2020, doi: 10.1016/j.media.2020.101790.

HERZOG, Lisa, Elvis MURINA, Oliver DÜRR, Susanne WEGENER und Beate SICK, 2020. Integrating uncertainty in deep neural networks for MRI based stroke analysis. Medical Image Analysis [online]. Oktober 2020. Bd. 65, Nr. 101790. DOI 10.1016/j.media.2020.101790. Verfügbar unter: https://arxiv.org/ftp/arxiv/papers/2008/2008.06332.pdf

Herzog, Lisa, Elvis Murina, Oliver Dürr, Susanne Wegener, and Beate Sick. 2020. “Integrating Uncertainty in Deep Neural Networks for MRI Based Stroke Analysis.” Medical Image Analysis 65 (101790). https://doi.org/10.1016/j.media.2020.101790.

Herzog, Lisa, et al. “Integrating Uncertainty in Deep Neural Networks for MRI Based Stroke Analysis.” Medical Image Analysis, vol. 65, no. 101790, Oct. 2020, https://doi.org/10.1016/j.media.2020.101790.