Deep Neural Networks Improve Radiologists’ Performance in Breast Cancer Screening

By Nan Wu, Jason Phang, Jungkyu Park, Yiqiu Shen, Zhe Huang, Masha Zorin, Stanislaw Jastrzebski, Thibault Févry, Joe Katsnelson, Eric Kim and 22 more. 2019

596 citations recorded by the source corpus.

We present a deep convolutional neural network for breast cancer screening exam classification, trained, and evaluated on over 200000 exams (over 1000000 images). Our network achieves an AUC of 0.895 in predicting the presence of cancer in the breast, when tested on the screening population. We attribute the high accuracy to a few technical advances. 1) Our network’s novel two-stage architecture and training procedure, which allows us to use a high-capacity patch-level network to learn from pixel-level labels alongside a network learning from macroscopic breast-level labels. 2) A custom ResNet-based network used as a building block of our model, whose balance of depth and width is optimized for high-resolution medical images. 3) Pretraining the network on screening BI-RADS classification, a related task with more noisy labels. 4) Combining multiple input views in an optimal way among a number of possible choices. To validate our model, we conducted a reader study with 14 readers, each reading 720 screening mammogram exams, and show that our model is as accurate as experienced radiologists when presented with the same data. We also show that a hybrid model, averaging the probability of malignancy predicted by a radiologist with a prediction of our neural network, is more accurate than either of the two separately. To further understand our results, we conduct a thorough analysis of our network’s performance on different subpopulations of the screening population, the model’s design, training procedure, errors, and properties of its internal representations. Our best models are publicly available at https://github.com/nyukat/breast_cancer_classifier.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

A case-based interpretable deep learning model for classification of mass lesions in digital mammography (2021, methodology): “We cannot compare with papers focused on detection because our technique works on diagnosis of an already detected lesion [35, 54].”
Predicting cancer outcomes with radiomics and artificial intelligence in radiology (2021, background): “In several high-profile studies, CNN-based models have even surpassed the performance of expert human readers in interpreting chest radiography 84 and CT 85 , and digital mammography 6,86 .”
AI in health and medicine (2022, background): “AI systems have achieved considerable improvements in accuracy for radiology tasks, including mammography interpretation 21,22 , cardiac function assessment 23,24 and lung cancer screening 25 , tackling not only diagnosis but also risk prediction and treatment 26 .”
Key challenges for delivering clinical impact with artificial intelligence (2019, background): “radiographs [5–9], detecting cancer in mammograms [10, 11], analysing computer tomography scans [12–15], identifying brain tumours on magnetic resonance images [16], and predicting development of Alzheimer’s disease from positron emission tomography [17].”
Health system-scale language models are all-purpose prediction engines (2023, methodology): “1b,c), we pretrained and fine-tuned an LLM for each downstream task using a bidirectional encoder model known as BERT (Bidirectional Encoder Representation with Trans-former) and a masked language modelling (MLM) objective on the NYU Notes dataset 11 until the validation loss plateaued.”
How Machine Learning will Transform Biomedicine (2020, unknown): “For example, deep learning-based CAD software was able to detect diabetic retinopathy at high levels of accuracy (Gulshan et al., 2016) and to retrospectively identify invasive and in-situ breast cancer of all grades similar to radiologists (McKinney et al., 2020; Wu et al., 2019).”