We present a deep convolutional neural network for breast cancer screening exam classification, trained and evaluated on over 200,000 exams (over 1,000,000 images). Our network achieves an AUC of 0.895 in predicting whether there is a cancer in the breast, when tested on the screening population. We attribute the high accuracy of our model to a two-stage training procedure, which allows us to use a very high-capacity patch-level network to learn from pixel-level labels alongside a network learning from macroscopic breast-level labels. To validate our model, we conducted a reader study with 14 readers, each reading 720 screening mammogram exams, and find our model to be as accurate as experienced radiologists when presented with the same data. Finally, we show that a hybrid model, averaging probability of malignancy predicted by a radiologist with a prediction of our neural network, is more accurate than either of the two separately. To better understand our results, we conduct a thorough analysis of our network's performance on different subpopulations of the screening population, model design, training procedure, errors, and properties of its internal representations.deep learning | deep convolutional neural networks | breast cancer screening | mammography B reast cancer is the second leading cancer-related cause of death among women in the US. In 2014, over 39 million screening and diagnostic mammography exams were performed in the US. It is estimated that in 2015 232,000 women were diagnosed with breast cancer and approximately 40,000 died from it (1). Although mammography is the only imaging test that has reduced breast cancer mortality (2-4), there has been discussion regarding the potential harms of screening, including false positive recalls and associated false positive biopsies. The vast majority of the 10-15% of women asked to return following an inconclusive screening mammogram undergo another mammogram and/or ultrasound for clarification. After the additional imaging exams, many of these findings are determined as benign and only 10-20% are recommended to undergo a needle biopsy for further work-up. Among these, only 20-40% yield a diagnosis of cancer (5). Evidently, there is an unmet need to shift the balance of routine breast cancer screening towards more benefit and less harm.Traditional computer-aided detection (CAD) in mammography is routinely used by radiologists to assist with image interpretation, despite multicenter studies showing these CAD programs do not improve their diagnostic performance (6).These CAD programs typically use handcrafted features to mark sites on a mammogram that appear distinct from normal tissue structures. The radiologist decides whether to recall these findings, determining clinical significance and actionability. Recent developments in deep learning (7)-in particular, deep convolutional neural networks (CNNs) (8-12)-open possibilities for creating a new generation of CAD-like tools.This paper makes several contributions. Primarily, we train and evaluate a set of stro...

Though consistently shown to detect mammographically occult cancers, breast ultrasound has been noted to have high false-positive rates. In this work, we present an AI system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. Developed on 288,767 exams, consisting of 5,442,907 B-mode and Color Doppler images, the AI achieves an area under the receiver operating characteristic curve (AUROC) of 0.976 on a test set consisting of 44,755 exams. In a retrospective reader study, the AI achieves a higher AUROC than the average of ten board-certified breast radiologists (AUROC: 0.962 AI, 0.924 ± 0.02 radiologists). With the help of the AI, radiologists decrease their false positive rates by 37.3% and reduce requested biopsies by 27.8%, while maintaining the same level of sensitivity. This highlights the potential of AI in improving the accuracy, consistency, and efficiency of breast ultrasound diagnosis.

Preattentive texture segregation was examined using textures composed of randomly placed, oriented line segments. A difference in texture element orientation produced an illusory, or orientation-defined, texture edge. Subjects discriminated between two textures, one with a straight texture edge and one with a "wavy" texture edge. Across conditions the orientation of the texture elements and the orientation of the texture edge varied. Although the orientation difference across the texture edge (the "texture gradient") is an important determinant of texture segregation performance, it is not the only one. Evidence from several experiments suggests that configural effects are also important. That is, orientation-defined texture edges are strongest when the texture elements (on one side of the edge) are parallel to the edge. This result is not consistent with a number of texture segregation models including feature- and filter-based models. One possible explanation is that the second-order channel used to detect a texture edge of a particular orientation gives greater weight to first-order input channels of that same orientation.

Instantaneous texture discrimination performance was examined for different texture stimuli to uncover the use of edge-based and region-based texture analysis mechanisms. Textures were composed of randomly placed, short, oriented line segments. Line segment orientation was chosen randomly using a Gaussian distribution (described by a mean and a standard deviation). One such distribution determined the orientations on the left side of the image, and a second distribution was used for the right side. The two textures either abutted to form an edge or were separated by a blank region. A texture difference in mean orientation led to superior discrimination performance when the textures abutted. On the other hand, when the textures differed in the standard deviation of the orientation distribution, performance was similar in the two conditions. These results suggest that edge-based texture analysis mechanisms were used (i.e. were the most sensitive) in the abutting difference-in-mean case, but region-based texture analysis mechanisms were used in the other three cases.

Screening mammography should not be delayed after COVID-19 vaccination because axillary adenopathy is a common imaging finding and persists for as long as 43 weeks.