In light of a pending national standard requiring breast density notification in mammography reports, an emerging artificial intelligence (AI) tool may help reduce subjectivity and variability in breast density assessments. The Food and Drug Administration (FDA) has granted 510(k) clearance for Saige-Density (DeepHealth/RadNet), an adjunctive AI software that provides automated categorization of breast density based on the American College of Radiology's Breast Imaging Reporting and Data System (BI-RADS) classification. DeepHealth said a retrospective, multicenter study showed a 91.5 percent alignment between Saige-Density assessment and consensus assessment of breast density by five specialists in breast imaging. The Saige-Density AI algorithm was trained on a racially diverse database of over 166,000 images from 30,000 mammography exams across the United States, according to DeepHealth. "Achieving FDA clearance for another important tool in the breast cancer screening process in such a short time frame highlights our aggressive commitment to bringing state-of-the art AI innovation to the breast screening mammography market," noted Gregory Sorenson, M.D., the CEO and co-founder of DeepHealth.

LOS ANGELES, Nov. 28, 2022 (GLOBE NEWSWIRE) -- RadNet, Inc. (NASDAQ: RDNT), a national leader in providing high-quality, cost-effective, fixed-site outpatient diagnostic imaging services today reported that its lung artificial intelligence subsidiary, Aidence, and Google Health, a division of Alphabet, Inc. (NASDAQ: GOOG), announce an agreement to license Google Health's AI research model for lung nodule malignancy prediction on CT imaging. Aidence will develop, validate and bring this model to the market to support the early and accurate diagnosis of lung cancer and the reduction of unnecessary procedures in screening programs. Lung cancer screening with low-dose CT has been shown to significantly reduce lung cancer mortality by as high as 24% for men and 33% for women, according to the 2020 NELSON trial. Screening initiatives are increasingly being implemented in Europe, such as the UK's Targeted Lung Health Checks. In the United States, eligibility criteria have recently been broadened, further reflecting the benefit of lung cancer screening.

Positive or negative Medicare payment adjustments in 2019 will depend on performance to quality and other measures in 2017 under a new program called the Merit-based Incentive Payment System. Doing well on quality measures is important because they comprise 60 percent of a provider's total MIPS score – possibly 85 percent for certain specialties such as radiology. Most of the quality measures in MIPS are based on face-to-face patient encounters and are not particularly applicable to radiology. "We had to take full advantage of the few quality measures that were suitable to us," said Cheryl Sullivan, clinical workflow analyst at RadNet. "Our coding system was only partially effective in retrieving the necessary information for measures and needed to be checked frequently against the clinical record for accuracy. Manual coding is slow, expensive, and still must be checked for accuracy."

We describe a deep learning approach for automated brain hemorrhage detection from computed tomography (CT) scans. Our model emulates the procedure followed by radiologists to analyse a 3D CT scan in real-world. Similar to radiologists, the model sifts through 2D cross-sectional slices while paying close attention to potential hemorrhagic regions. Further, the model utilizes 3D context from neighboring slices to improve predictions at each slice and subsequently, aggregates the slice-level predictions to provide diagnosis at CT level. We refer to our proposed approach as Recurrent Attention DenseNet (RADnet) as it employs original DenseNet architecture along with adding the components of attention for slice level predictions and recurrent neural network layer for incorporating 3D context. The real-world performance of RADnet has been benchmarked against independent analysis performed by three senior radiologists for 77 brain CTs. RADnet demonstrates 81.82% hemorrhage prediction accuracy at CT level that is comparable to radiologists. Further, RADnet achieves higher recall than two of the three radiologists, which is remarkable.