Clinically applicable deep learning for diagnosis and referral in retinal disease

By Jeffrey De Fauw, J. Ledsam, Bernardino Romera-Paredes, Stanislav Nikolov, Nenad Tomašev, Sam Blackwell, Harry Askham, Xavier Glorot, Brendan O'Donoghue, D. Visentin and 24 more. 2018

2,402 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Ophthalmology, 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 Biomedical Engineering

In Computer Vision and Pattern Recognition

In Ophthalmology

Emily Y. Chew

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Data-Efficient Image Recognition with Contrastive Predictive Coding (2019, methodology): “erefore continued by adding a ﬁxed, generic augmentation scheme using the primitives from Cubuk et al. [10] (e.g. shearing, rotation, etc), as well as random elastic deformations and color transforms [11] (+4.5% accuracy). Note that these augmentations introduce some inductive bias about content-pre”
Contrastive Learning of Medical Visual Representations from Paired Images and Text (2020, background): “Medical image understanding has the potential to transform healthcare and has seen rapid progress with deep learning (Gulshan et al., 2016; Esteva et al., 2017; De Fauw et al., 2018; Rajpurkar et al., 2018b).”
Wasserstein Fair Classification (2019, background): “…for individuals and society, such as health care, criminal risk assessment, social services, hiring, ﬁnancial lending, and online advertising (De Fauw et al., 2018; Dieterich et al., 2016; Eubanks, 2018; Hoffman et al., 2018; Malekipir-bazari and Aksakalli, 2015; Perlich et al., 2014), is…”
Improving the accuracy of medical diagnosis with causal machine learning (2020, methodology): “…and using Bayes rule to estimate the posterior, P ð D jE ; θ Þ ¼ P ðEj D ; θ Þ P ð D ; θ Þ P ðE ; θ Þ : Examples of discriminative diagnostic models include neural network and deep learning models 8,10,15,33,34 , whereas generative models are typically Bayesian networks 18,19,21,22,27,35,36 .”
An overview of deep learning in medical imaging focusing on MRI (2018, background): “Not only medical technology companies, but also for example Google Brain [28–30],4 DeepMind [31],5 Microsoft [32,33]6 and IBM [34].”
Concept Bottleneck Models (2020, background): “More recently, deep neural networks with concept bottlenecks have re-emerged as targeted tools for solving particular tasks, e.g., Fauw et al. (2018) for retinal disease diagnosis, Yi et al. (2018) for visual question-answering, and Bucher et al. (2018) for content-based image retrieval.”
Should We Trust Algorithms? (2020, methodology): “…DeepMind’s eye diagnosis system developed with Moorfields Eye Hospital is based on a deep-learning algorithm, but one that has been deliberately structured in layers to help visually explain intermediate steps between the raw image and diagnosis and triage recommendation (De Fauw et al., 2018).”
Explainable machine learning in deployment (2019, background): “Machine learning (ML) models are being increasingly embedded into many aspects of daily life, such as healthcare [16], finance [26], and social media [5].”
Deep learning predicts hip fracture using confounding patient and healthcare variables (2018, background): “Clinicians are better at interpreting images when they consider a patient’s clinical context.(27,38) As clinicians are frequently considering other evidence sources not available to the model (e.”
Are Labels Required for Improving Adversarial Robustness? (2019, methodology): “System based on these techniques are now being developed and deployed for a wide variety of applications, from recommending/ranking content on the web [11, 18] to autonomous driving [6] and even in medical diagnostics [12].”
Key challenges for delivering clinical impact with artificial intelligence (2019, background): “One recent approach replaced end-to-end classification with a twostage architecture comprising segmentation and classification, allowing the clinician to interrogate the segmentation map to understand the basis of the subsequent classification [24].”
REFUGE Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs (2019, neutral): “This has been recently applied by De Fauw et al. (2018) for retinal OCT analysis with a significant success.”
Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis (2020, background): “Given a frontal X-ray image, the CAD system predicts the presence or absence of the 14 thoracic diseases.”