HOPPR has launched a foundation model that processes 3D chest CT volumes and generates narrative language describing findings across pulmonary, mediastinal, cardiac, upper abdominal, osseous, and soft-tissue regions.
The EF Chest CT Narrative Model was trained on a proprietary dataset of chest CT studies from multiple U.S. clinical sites, with an emphasis on ensuring less common but serious conditions such as aortic injury, pulmonary embolism, rib fractures, and pneumothorax were well represented in training, according to the company. It is available through HOPPR Forward Deployed Services, which provides support from evaluation through integration.
With the release, HOPPR's foundation model portfolio now spans chest x-ray, mammography, and chest CT, covering both AI-assisted classification and narrative-language-generation tasks. The models are available on the HOPPR AI Foundry alongside third-party models from NVIDIA, Google, Microsoft, and Stanford AIMI, among others, the company said.
