CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning

By Pranav Rajpurkar, Jeremy Irvin, Kaylie Zhu, Brandon Yang, Hershel Mehta, Tony Duan, Daisy Yi Ding, Aarti Bagul, Curtis P. Langlotz, K. Shpanskaya and 2 more. 2017

3,431 citations recorded by the source corpus.

We develop an algorithm that can detect pneumonia from chest X-rays at a level exceeding practicing radiologists. Our algorithm, CheXNet, is a 121-layer convolutional neural network trained on ChestX-ray14, currently the largest publicly available chest X-ray dataset, containing over 100,000 frontal-view X-ray images with 14 diseases. Four practicing academic radiologists annotate a test set, on which we compare the performance of CheXNet to that of radiologists. We find that CheXNet exceeds average radiologist performance on the F1 metric. We extend CheXNet to detect all 14 diseases in ChestX-ray14 and achieve state of the art results on all 14 diseases.

Subfields: Computer Vision and Pattern Recognition, Pulmonary and Respiratory Medicine, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Pulmonary and Respiratory Medicine

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Shortcut learning in deep neural networks (2020, background): “ing pace, DNNs were reported as having achieved human-level object classiﬁ- cation performance [1], beating world-class human Go, Poker, and Starcraft players [2,3], detecting cancer from X-ray scans [4], translating text across languages [5], helping combat climate change [6], and accelerating the ”
Unsolved Problems in ML Safety (2021, background): “As machine learning (ML) systems are deployed in high-stakes environments, such as medical settings [147], roads [185], and command and control centers [39], unsafe ML systems may result in needless loss of life.”
Deep learning predicts hip fracture using confounding patient and healthcare variables (2018, background): “Class imbalance can be partly mitigated by data augmentation strategies.(8,12,28,29,33) Islam et al.”
Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: A cross-sectional study (2018, methodology): “NIH radiographs were labeled automatically using a proprietary natural language processing (NLP) system based on expanding sentences as parse trees and using hand-crafted rules based on the MESH vocabulary to identify statements indicating positive pathology [2].”
What is being transferred in transfer learning? (2020, background): “In some cases there is nontrivial difference in visual forms between the source and the target domain [42, 37].”
Deep Learning Applications in Medical Image Analysis (2018, methodology): “[48] employed a modified DenseNet [49] with 121 convolutional layers called CheXNet to classify 14 different diseases seen on the chest x-rays, using 112,000 images from the ChestXray14 [17] dataset.”
Diagnosis of Pneumonia from Chest X-Ray Images Using Deep Learning (2019, methodology): “CheXNet achieved a f1 score of 0.435 (95% CI 0.387, 0.481), higher than the radiologist average of 0.387 (95% CI 0.330, 0.442).”
Direct Uncertainty Prediction for Medical Second Opinions (2018, background): “Despite this domain expertise, the disagreement issue remains fundamental [2, 1, 5, 12].”
Cross-Domain Transferability of Adversarial Perturbations (2019, methodology): “There are approximately 80k samples in Paintings, 50k in Comics, 1.2 million in ImageNet training set and 10k in ChestX.”
Dual-Sampling Attention Network for Diagnosis of COVID-19 From Community Acquired Pneumonia (2020, methodology): “The CNN networks have been successfully applied to pneumonia diagnosis in CXR images [16], [32].”
Can we trust deep learning models diagnosis? The impact of domain shift in chest radiograph classification (2019, methodology): “Training and testing on ChestX-ray14 achieves results similar to the ones reported on [4].”
Identifying pneumonia in chest X-rays: A deep learning approach (2019, background): “Rajpurkar et al. [19] developed a 121 layer deep convolutional layer network chestX-ray14 dataset.”