Can we trust deep learning models diagnosis? The impact of domain shift in chest radiograph classification

@article{Pooch2020CanWT,
  title={Can we trust deep learning models diagnosis? The impact of domain shift in chest radiograph classification},
  author={E. Pooch and Pedro L. Ballester and Rodrigo C. Barros},
  journal={ArXiv},
  year={2020},
  volume={abs/1909.01940}
}
While deep learning models become more widespread, their ability to handle unseen data and generalize for any scenario is yet to be challenged. In medical imaging, there is a high heterogeneity of distributions among images based on the equipment that generates them and their parametrization. This heterogeneity triggers a common issue in machine learning called domain shift, which represents the difference between the training data distribution and the distribution of where a model is employed… Expand
An adversarial approach for the robust classification of pneumonia from chest radiographs
TLDR
Improved out-of-hospital generalization performance of a pneumonia classifier is demonstrated by training a model that is invariant to the view position of chest radiographs (anterior-posterior vs. posterior-anterior). Expand
Computer-aided abnormality detection in chest radiographs in a clinical setting via domain-adaptation
TLDR
The experimental results show the proposed domain-shift detection and removal method’s effectiveness in deploying a pre-trained DL model for abnormality detection in chest radiographs in a clinical setting. Expand
The reliability of a deep learning model in clinical out-of-distribution MRI data: a multicohort study
TLDR
This is the most comprehensive study to date investigating the domain shift in deep learning on MRI data, and it is recommended that rigorous evaluation of DL models on clinical data prior to being certified for deployment is advocated. Expand
Advancing Medical Imaging Informatics by Deep Learning-Based Domain Adaptation
TLDR
DA has emerged as a promising solution to deal with the lack of annotated training data, especially for segmentation tasks and among various DA approaches, domain transformation (DT) and latent feature-space transformation (LFST) are discussed. Expand
Deep Learning for Automatic Classification of Multi-Modal Information Corresponding to Chest Radiology Reports
Hospitals frequently use chest X-ray images, which can provide a detailed view of a patient’s thorax, as a radiology examination technique for the diagnosis of cardiac and pulmonary diseases.Expand
Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph
TLDR
It is concluded that the recalibration method can transfer models from digital CXRs to CXR photographs, which is expected to help physicians’ clinical works. Expand
Interpretable and Annotation-Efficient Learning for Medical Image Computing: Third International Workshop, iMIMIC 2020, Second International Workshop, MIL3ID 2020, and 5th International Workshop, LABELS 2020, Held in Conjunction with MICCAI 2020, Lima, Peru, October 4–8, 2020, Proceedings
TLDR
It is shown that CNNs permit highly accurate discrimination of the fracture status of individual vertebrae and the benefit of using attribution maps to explain a network’s decision is evaluated. Expand
Domain Adaptation for Medical Image Analysis: A Survey
  • Hao Guan, Mingxia Liu
  • Computer Science, Medicine
  • IEEE transactions on bio-medical engineering
  • 2021
TLDR
A survey of the recent advances of domain adaptation methods in medical image analysis to enable researchers to gain a better understanding of the current status, challenges and future directions of this energetic research field. Expand
Fine-Tuning Approach for Segmentation of Gliomas in Brain Magnetic Resonance Images with a Machine Learning Method to Normalize Image Differences among Facilities
TLDR
It is demonstrated that the performance degradation of machine learning models for segmenting gliomas in brain magnetic resonance images caused by domain shift can be improved by fine-tuning methods with a small number of cases from each facility, although the data obtained appeared to be biased. Expand
A critic evaluation of methods for COVID-19 automatic detection from X-ray images
TLDR
It is deduced that several testing protocols for the recognition are not fair and that the neural networks are learning patterns in the dataset that are not correlated to the presence of COVID-19. Expand
...
1
2
3
4
...

References

SHOWING 1-10 OF 27 REFERENCES
CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparison
TLDR
A labeler is designed to automatically detect the presence of 14 observations in radiology reports, capturing uncertainties inherent in radiograph interpretation, in CheXpert, a large dataset that contains 224,316 chest radiographs of 65,240 patients. Expand
Semi-supervised learning with generative adversarial networks for chest X-ray classification with ability of data domain adaptation
TLDR
For cardiac abnormality classification in chest X-rays, it is demonstrated that an order of magnitude less data is required with semi-supervised learning generative adversarial networks than with conventional supervised learning convolutional neural networks. Expand
Semantic-Aware Generative Adversarial Nets for Unsupervised Domain Adaptation in Chest X-ray Segmentation
TLDR
A novel unsupervised domain adaptation approach for segmentation tasks by designing semantic-aware generative adversarial networks (GANs) that transform the test image into the appearance of source domain, with the semantic structural information being well preserved. Expand
Unsupervised Reverse Domain Adaptation for Synthetic Medical Images via Adversarial Training
TLDR
This work proposes a novel framework that uses a reverse flow, where adversarial training is used to make real medical images more like synthetic images, and clinically-relevant features are preserved via self-regularization to improve structural similarity of endoscopy depth estimation. Expand
CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
TLDR
An algorithm that can detect pneumonia from chest X-rays at a level exceeding practicing radiologists is developed, and it is found that CheXNet exceeds average radiologist performance on the F1 metric. Expand
Exploring large scale public medical image datasets
TLDR
Visual inspection of images is a necessary component of understanding large image datasets and teams producing public datasets should perform this important quality control procedure and include a thorough description of their findings, along with an explanation of the data generating procedures and labeling rules, in the documentation for their datasets. Expand
Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss
TLDR
This paper proposes an unsupervised domain adaptation framework with adversarial learning for cross-modality biomedical image segmentations, based on a dilated fully convolutional network for pixel-wise prediction. Expand
PadChest: A large chest x-ray image dataset with multi-label annotated reports
TLDR
This dataset includes more than 160,000 images obtained from 67,000 patients that were interpreted and reported by radiologists at San Juan Hospital (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography. Expand
Deep learning for segmentation of brain tumors: Impact of cross‐institutional training and testing
TLDR
There is a very strong effect of selecting data for training on performance of CNNs in a multi-institutional setting, and the reasons behind this effect requires additional comprehensive investigation. Expand
MIMIC-CXR: A large publicly available database of labeled chest radiographs
TLDR
MIMic-CXR-JPG is derived entirely from the MIMIC-C XR database, and aims to provide a convenient processed version of MIMICS CXR, as well as to provided a standard reference for data splits and image labels. Expand
...
1
2
3
...