Summary
NVIDIA’s October 2026 blog profiles four startup companies developing AI tools for breast imaging, risk assessment and treatment planning. The systems are separate products and projects, with FDA clearance specified for two of them.
NVIDIA’s October 5, 2026 blog profiles four companies in its Inception startup programme that are developing AI tools for different stages of breast-cancer care: ultrasound imaging, mammogram assessment, treatment-response prediction and 3D tumour visualisation. The companies’ systems are separate products and projects, not one integrated clinical platform.
The examples range from products NVIDIA identifies as FDA-cleared to research and other tools whose regulatory status is not specified in the post. Their roles also differ: some process medical images, while others estimate risk or treatment outcomes for clinicians to consider.
Imaging and screening tools
iSono Health’s ATUSA is a wearable, automated 3D quantitative ultrasound system. NVIDIA says it captures a standardised breast volume in about two minutes per breast, compared with up to 45 minutes for conventional handheld ultrasound. The system’s AI automates image acquisition and was trained on thousands of full-breast scans containing more than 1.5 million ultrasound frames.
The blog identifies ATUSA as FDA-cleared. iSono Health reports that it is 28% more sensitive than handheld 2D ultrasound. The company says the automated, repeatable scans could also help clinicians compare breast tissue across visits. A multicentre clinical study involving 3,200 patients is underway to further validate the platform, with lead research sites at UC Davis and Vanderbilt University Medical Center.
Whiterabbit.ai’s FDA-cleared WRDensity software assesses breast density from mammograms and, according to the blog, has been used in the care of hundreds of thousands of patients. The company also describes WRRisk, software for estimating long-term breast-cancer risk. Separately, it is researching AI intended to help radiologists find more cancers and automate the screening of mammograms classified as negative. The post presents those capabilities as a development effort, not as a cleared product.
From pathology slides to treatment predictions
Ataraxis AI uses digital pathology slides and standard clinical variables to estimate treatment response and recurrence risk. One model predicts whether presurgical chemotherapy is likely to shrink a tumour to a response before surgery. Another estimates five-year recurrence risk after surgery and the likely benefit of chemotherapy as a next step. NVIDIA says the models have been validated across more than 10 institutions and multiple clinical trials, and are in active clinical use.
SimBioSys combines imaging and clinical information to create 3D models of breast tumours and other soft tissue, which it says can help guide surgery and inform treatment planning. Its recurrence-risk tool uses volumetric breast MRI data alongside tumour pathology and clinical data. The company says its platform can also incorporate genomic testing when applicable, bringing together information that would otherwise be assessed in separate forms.
These systems are described as tools to support clinical assessment, rather than as replacements for clinicians. NVIDIA’s post says certain technologies in the article are investigational and have not been approved by the U.S. FDA for commercial use, but does not specify which ones. It separately identifies ATUSA and WRDensity as FDA-cleared.