Quantifying and Leveraging Classification Uncertainty for Chest Radiograph Assessment

The interpretation of chest radiographs is an essential task for the detection of thoracic diseases and abnormalities. However, it is a challenging problem with high inter-rater variability and inherent ambiguity due to inconclusive evidence in the data, limited data quality or subjective definitions of disease appearance. Current deep learning solutions for chest radiograph abnormality classification are typically limited to providing probabilistic predictions, relying on the capacity of learning models to adapt to the high degree of label noise and become robust to the enumerated causal factors. In practice, however, this leads to overconfident systems with poor generalization on unseen data. To account for this, we propose an automatic system that learns not only the probabilistic estimate on the presence of an abnormality, but also an explicit uncertainty measure which captures the confidence of the system in the predicted output. We argue that explicitly learning the classification uncertainty as an orthogonal measure to the predicted output, is essential to account for the inherent variability characteristic of this data. Experiments were conducted on two datasets of chest radiographs of over 85,000 patients. Sample rejection based on the predicted uncertainty can significantly improve the ROC-AUC, e.g., by 8% to 0.91 with an expected rejection rate of under 25%. Eliminating training samples using uncertainty-driven bootstrapping, enables a significant increase in robustness and accuracy. In addition, we present a multi-reader study showing that the predictive uncertainty is indicative of reader errors.

[1]  Murat Sensoy,et al.  Evidential Deep Learning to Quantify Classification Uncertainty , 2018, NeurIPS.

[2]  Liang Zheng,et al.  Thorax disease classification with attention guided convolutional neural network , 2020, Pattern Recognit. Lett..

[3]  Dmitry P. Vetrov,et al.  Variational Dropout Sparsifies Deep Neural Networks , 2017, ICML.

[4]  Audun Jøsang,et al.  Subjective Logic: A Formalism for Reasoning Under Uncertainty , 2016 .

[5]  A. Ng,et al.  Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists , 2018, PLoS medicine.

[6]  Ronald M. Summers,et al.  ChestX-ray: Hospital-Scale Chest X-ray Database and Benchmarks on Weakly Supervised Classification and Localization of Common Thorax Diseases , 2019, Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics.

[7]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Charles Blundell,et al.  Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles , 2016, NIPS.

[9]  J. Gohagan,et al.  The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. , 2000, Controlled clinical trials.

[10]  Dorin Comaniciu,et al.  Learning to recognize Abnormalities in Chest X-Rays with Location-Aware Dense Networks , 2018, CIARP.

[11]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[12]  Li Yao,et al.  Weakly Supervised Medical Diagnosis and Localization from Multiple Resolutions , 2018, ArXiv.

[13]  Yi Yang,et al.  Diagnose like a Radiologist: Attention Guided Convolutional Neural Network for Thorax Disease Classification , 2018, ArXiv.

[14]  Arthur P. Dempster,et al.  A Generalization of Bayesian Inference , 1968, Classic Works of the Dempster-Shafer Theory of Belief Functions.