Uncertainty-Aware Training for Cardiac Resynchronisation Therapy Response Prediction

TY - CHAP

T1 - Uncertainty-Aware Training for Cardiac Resynchronisation Therapy Response Prediction

AU - Dawood, Tareen

AU - Chen, Chen

AU - Andlauer, Robin

AU - Sidhu, Baldeep S.

AU - Ruijsink, Bram

AU - Gould, Justin

AU - Porter, Bradley

AU - Elliott, Mark

AU - Mehta, Vishal

AU - Rinaldi, C. Aldo

AU - Puyol-Antón, Esther

AU - Razavi, Reza

AU - King, Andrew P.

N1 - Funding Information: This work was supported by the NIHR Guys and St Thomas Biomedical Research Centre and the Kings DRIVE Health CDT for Data-Driven Health. This research has been conducted using the UK Biobank Resource under Application Number 17806. The work was also supported by the EPSRC through the SmartHeart Programme Grant (EP/P001009/1). Funding Information: Acknowledgements. This work was supported by the NIHR Guys and St Thomas Biomedical Research Centre and the Kings DRIVE Health CDT for Data-Driven Health. This research has been conducted using the UK Biobank Resource under Application Number 17806. The work was also supported by the EPSRC through the SmartHeart Programme Grant (EP/P001009/1). Publisher Copyright: © 2022, Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - Evaluation of predictive deep learning (DL) models beyond conventional performance metrics has become increasingly important for applications in sensitive environments like healthcare. Such models might have the capability to encode and analyse large sets of data but they often lack comprehensive interpretability methods, preventing clinical trust in predictive outcomes. Quantifying uncertainty of a prediction is one way to provide such interpretability and promote trust. However, relatively little attention has been paid to how to include such requirements into the training of the model. In this paper we: (i) quantify the data (aleatoric) and model (epistemic) uncertainty of a DL model for Cardiac Resynchronisation Therapy response prediction from cardiac magnetic resonance images, and (ii) propose and perform a preliminary investigation of an uncertainty-aware loss function that can be used to retrain an existing DL image-based classification model to encourage confidence in correct predictions and reduce confidence in incorrect predictions. Our initial results are promising, showing a significant increase in the (epistemic) confidence of true positive predictions, with some evidence of a reduction in false negative confidence.

AB - Evaluation of predictive deep learning (DL) models beyond conventional performance metrics has become increasingly important for applications in sensitive environments like healthcare. Such models might have the capability to encode and analyse large sets of data but they often lack comprehensive interpretability methods, preventing clinical trust in predictive outcomes. Quantifying uncertainty of a prediction is one way to provide such interpretability and promote trust. However, relatively little attention has been paid to how to include such requirements into the training of the model. In this paper we: (i) quantify the data (aleatoric) and model (epistemic) uncertainty of a DL model for Cardiac Resynchronisation Therapy response prediction from cardiac magnetic resonance images, and (ii) propose and perform a preliminary investigation of an uncertainty-aware loss function that can be used to retrain an existing DL image-based classification model to encourage confidence in correct predictions and reduce confidence in incorrect predictions. Our initial results are promising, showing a significant increase in the (epistemic) confidence of true positive predictions, with some evidence of a reduction in false negative confidence.

KW - Awareness

KW - Cardiac resynchronisation therapy

KW - Reliability

KW - Trust

KW - Uncertainty

UR - http://www.scopus.com/inward/record.url?scp=85124009026&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-93722-5_21

DO - 10.1007/978-3-030-93722-5_21

M3 - Conference paper

AN - SCOPUS:85124009026

SN - 9783030937218

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 189

EP - 198

BT - Statistical Atlases and Computational Models of the Heart. Multi-Disease, Multi-View, and Multi-Center Right Ventricular Segmentation in Cardiac MRI Challenge - 12th International Workshop, STACOM 2021, Held in Conjunction with MICCAI 2021, Revised Selected Papers

A2 - Puyol Antón, Esther

A2 - Young, Alistair

A2 - Suinesiaputra, Avan

A2 - Pop, Mihaela

A2 - Martín-Isla, Carlos

A2 - Sermesant, Maxime

A2 - Camara, Oscar

A2 - Lekadir, Karim

PB - Springer Science and Business Media Deutschland GmbH

T2 - 12th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2021 held in conjunction with MICCAI 2021

Y2 - 27 September 2021 through 27 September 2021

ER -