Corpus ID: 237592831

Uncertainty-Aware Training for Cardiac Resynchronisation Therapy Response Prediction

@article{Dawood2021UncertaintyAwareTF,
  title={Uncertainty-Aware Training for Cardiac Resynchronisation Therapy Response Prediction},
  author={Tareen Dawood and Chen Chen and Robin Andlauer and Baldeep Singh Sidhu and Bram Ruijsink and Justin S. Gould and Bradley Porter and Mark K. Elliott and Vishal S. Mehta and Christopher A. Rinaldi and Esther Puyol-Ant'on and Reza Razavi and Andrew P. King},
  journal={ArXiv},
  year={2021},
  volume={abs/2109.10641}
}
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… Expand

Figures and Tables from this paper

References

SHOWING 1-10 OF 16 REFERENCES
Uncertainty-Aware Training of Neural Networks for Selective Medical Image Segmentation
TLDR
A novel method is presented that considers such uncertainty in the training process to maximize the accuracy on the confident subset rather than the Accuracy on the whole dataset. Expand
A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
TLDR
This study reviews recent advances in UQ methods used in deep learning and investigates the application of these methods in reinforcement learning (RL), and outlines a few important applications of UZ methods. Expand
A framework for combining a motion atlas with non‐motion information to learn clinically useful biomarkers: Application to cardiac resynchronisation therapy response prediction
TLDR
Using a cohort of 34 patients selected for CRT using conventional criteria, results show that the combination of motion and non‐motion data enables CRT response to be predicted with 91.2% accuracy, which compares favourably with the current state‐of‐the‐art inCRT response prediction. Expand
Cardiac Resynchronization Therapy in Non-Ischemic Cardiomyopathy: Role of Multimodality Imaging
TLDR
This review provides an up-to-date synthesis of the latest evidence of CRT use in non-ischemic cardiomyopathy and highlights the potential additional value of multimodality imaging for improving CRT response in this population. Expand
A Probabilistic U-Net for Segmentation of Ambiguous Images
TLDR
A generative segmentation model based on a combination of a U-Net with a conditional variational autoencoder that is capable of efficiently producing an unlimited number of plausible hypotheses and reproduces the possible segmentation variants as well as the frequencies with which they occur significantly better than published approaches. Expand
Electromechanical Model to Predict Cardiac Resynchronization Therapy
  • M. Albatat, D. King, +5 authors I. Balasingham
  • Medicine, Computer Science
  • 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
  • 2018
TLDR
A pipeline for improved CRT-therapy is introduced by creating an electromechanical model using patient-specific geometric parameters allowing individualization of therapy, which successfully mimics expected changes when variables for tension, stiffness, and conduction are entered. Expand
Cardiac Resynchronization Therapy Optimization: A Comprehensive Approach
TLDR
An updated overview of the electropathophysiology of myocardial dysfunction in ventricular conduction delay and the diagnostic approaches involving the use of multiple modalities is provided. Expand
UK Biobank’s cardiovascular magnetic resonance protocol
TLDR
The CMR protocol applied in UK Biobank’s pilot phase is described, which will be extended into the main phase with three centres using the same equipment and protocols. Expand
Selective Classification for Deep Neural Networks
TLDR
A method to construct a selective classifier given a trained neural network, which allows a user to set a desired risk level and the classifier rejects instances as needed, to grant the desired risk (with high probability). Expand
A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
TLDR
A simple baseline that utilizes probabilities from softmax distributions is presented, showing the effectiveness of this baseline across all computer vision, natural language processing, and automatic speech recognition, and it is shown the baseline can sometimes be surpassed. Expand
...
1
2
...