Multi-vendor evaluation of artificial intelligence as an independent reader for double reading in breast cancer screening on 275,900 mammograms.

BACKGROUND: Double reading (DR) in screening mammography increases cancer detection and lowers recall rates, but has sustainability challenges due to workforce shortages. Artificial intelligence (AI) as an independent reader (IR) in DR may provide a cost-effective solution with the potential to improve screening performance. Evidence for AI to generalise across different patient populations, screening programmes and equipment vendors, however, is still lacking. METHODS: This retrospective study simulated DR with AI as an IR, using data representative of real-world deployments (275,900 cases, 177,882 participants) from four mammography equipment vendors, seven screening sites, and two countries. Non-inferiority and superiority were assessed for relevant screening metrics. RESULTS: DR with AI, compared with human DR, showed at least non-inferior recall rate, cancer detection rate, sensitivity, specificity and positive predictive value (PPV) for each mammography vendor and site, and superior recall rate, specificity, and PPV for some. The simulation indicates that using AI would have increased arbitration rate (3.3% to 12.3%), but could have reduced human workload by 30.0% to 44.8%. CONCLUSIONS: AI has potential as an IR in the DR workflow across different screening programmes, mammography equipment and geographies, substantially reducing human reader workload while maintaining or improving standard of care. TRIAL REGISTRATION: ISRCTN18056078 (20/03/2019; retrospectively registered).

Generalisable deep learning method for mammographic density prediction across imaging techniques and self-reported race.

BACKGROUND: Breast density is an important risk factor for breast cancer complemented by a higher risk of cancers being missed during screening of dense breasts due to reduced sensitivity of mammography. Automated, deep learning-based prediction of breast density could provide subject-specific risk assessment and flag difficult cases during screening. However, there is a lack of evidence for generalisability across imaging techniques and, importantly, across race. METHODS: This study used a large, racially diverse dataset with 69,697 mammographic studies comprising 451,642 individual images from 23,057 female participants. A deep learning model was developed for four-class BI-RADS density prediction. A comprehensive performance evaluation assessed the generalisability across two imaging techniques, full-field digital mammography (FFDM) and two-dimensional synthetic (2DS) mammography. A detailed subgroup performance and bias analysis assessed the generalisability across participants' race. RESULTS: Here we show that a model trained on FFDM-only achieves a 4-class BI-RADS classification accuracy of 80.5% (79.7-81.4) on FFDM and 79.4% (78.5-80.2) on unseen 2DS data. When trained on both FFDM and 2DS images, the performance increases to 82.3% (81.4-83.0) and 82.3% (81.3-83.1). Racial subgroup analysis shows unbiased performance across Black, White, and Asian participants, despite a separate analysis confirming that race can be predicted from the images with a high accuracy of 86.7% (86.0-87.4). CONCLUSIONS: Deep learning-based breast density prediction generalises across imaging techniques and race. No substantial disparities are found for any subgroup, including races that were never seen during model development, suggesting that density predictions are unbiased.

Women with dense breasts have a higher risk of breast cancer. For dense breasts, it is also more difficult to spot cancer in mammograms, which are the X-ray images commonly used for breast cancer screening. Thus, knowing about an individual's breast density provides important information to doctors and screening participants. This study investigated whether an artificial intelligence algorithm (AI) can be used to accurately determine the breast density by analysing mammograms. The study tested whether such an algorithm performs equally well across different imaging devices, and importantly, across individuals from different self-reported race groups. A large, racially diverse dataset was used to evaluate the algorithm's performance. The results show that there were no substantial differences in the accuracy for any of the groups, providing important assurances that AI can be used safely and ethically for automated prediction of breast density.

Automatic correction of performance drift under acquisition shift in medical image classification.

Image-based prediction models for disease detection are sensitive to changes in data acquisition such as the replacement of scanner hardware or updates to the image processing software. The resulting differences in image characteristics may lead to drifts in clinically relevant performance metrics which could cause harm in clinical decision making, even for models that generalise in terms of area under the receiver-operating characteristic curve. We propose Unsupervised Prediction Alignment, a generic automatic recalibration method that requires no ground truth annotations and only limited amounts of unlabelled example images from the shifted data distribution. We illustrate the effectiveness of the proposed method to detect and correct performance drift in mammography-based breast cancer screening and on publicly available histopathology data. We show that the proposed method can preserve the expected performance in terms of sensitivity/specificity under various realistic scenarios of image acquisition shift, thus offering an important safeguard for clinical deployment.

Dislocation loop formation by swift heavy ion irradiation of metals.

A coupled two-temperature, molecular dynamics methodology is used to simulate the structural evolution of bcc metals (Fe and W) and fcc metals (Cu and Ni) following irradiation by swift heavy ions. Electronic temperature dependent electronic specific heat capacities and electron-phonon coupling strengths are used to capture the full effects of the variation in the electronic density of states. Tungsten is found to be significantly more resistant to damage than iron, due both to the higher melting temperature and the higher thermal conductivity. Very interesting defect structures, quite different from defects formed in cascades, are found to be created by swift heavy ion irradiation in the bcc metals. Isolated vacancies form a halo around elongated interstitial dislocation loops that are oriented along the ion path. Such configurations are formed by rapid recrystallization of the molten cylindrical region that is created by the energetic ion. Vacancies are created at the recrystallization front, resulting in excess atoms at the core which form interstitial dislocation loops on completion of crystallization. These unique defect structures could, potentially, be used to create metal films with superior mechanical properties and interesting nanostructures.

The influence of the electronic specific heat on swift heavy ion irradiation simulations of silicon.

The swift heavy ion (SHI) irradiation of materials is often modelled using the two-temperature model. While the model has been successful in describing SHI damage in metals, it fails to account for the presence of a bandgap in semiconductors and insulators. Here we explore the potential to overcome this limitation by explicitly incorporating the influence of the bandgap in the parameterisation of the electronic specific heat for Si. The specific heat as a function of electronic temperature is calculated using finite temperature density functional theory with three different exchange correlation functionals, each with a characteristic bandgap. These electronic temperature dependent specific heats are employed with two-temperature molecular dynamics to model ion track creation in Si. The results obtained using a specific heat derived from density functional theory showed dramatically reduced defect creation compared to models that used the free electron gas specific heat. As a consequence, the track radii are smaller and in much better agreement with experimental observations. We also observe a correlation between the width of the band gap and the track radius, arising due to the variation in the temperature dependence of the electronic specific heat.