Two NVIDIA Inception startups, iSono Health and Whiterabbit.ai, are rolling out FDA-cleared AI tools for breast cancer screening and risk assessment, aiming to ease a screening bottleneck that's about to get worse. A projected shortfall of tens of thousands of radiologists over the next decade is straining the system's capacity to read mammograms, even as about 40 million are performed in the U.S. each year. Both companies lean on NVIDIA's AI infrastructure to speed up their pipelines.
ATUSA's two-minute scan
ISono Health's ATUSA platform captures a standardized breast volume in roughly two minutes per breast. A conventional handheld ultrasound can take up to 45 minutes. The company's AI is trained on thousands of full-breast scans — more than 1.5 million ultrasound frames — and uses NVIDIA GPU acceleration plus open source medical imaging tools to produce a 3D scan. iSono says that scan is 28% more sensitive than a handheld 2D ultrasound.
📊 Market Data Snapshot
ATUSA is commercially available through partner clinics in California, Texas, Georgia, Tennessee and Washington D.C. A multicenter clinical study with 3,200 patients is underway, led by research sites at UC Davis and Vanderbilt University Medical Center. Neda Razavi, iSono's CEO, has previously pointed to access gaps as a core motivation for the platform.
Whiterabbit's density and risk play
Whiterabbit.ai's FDA-cleared WRDensity software automatically assesses breast density from mammograms and has been used in the care of hundreds of thousands of patients. The company also developed WRRisk, a clinical decision support tool that estimates a patient's long-term risk of developing breast cancer.
Jason Su, Whiterabbit's CEO, has said the goal is to catch risk earlier. The company trains its AI models on a cluster of NVIDIA GPUs housed at Washington University in St. Louis, supplemented by extra GPU capacity in the cloud. Inference runs on NVIDIA GPUs deployed directly in the clinic — a setup that keeps data local and latency low.
The screening gap neither tool solves alone
Breast cancer is the most commonly diagnosed cancer among American women. A majority of women over 40 skip the recommended annual screening. Faster, cheaper imaging could help, but adoption depends on clinic economics and patient follow-through. The radiologist shortage is the harder constraint: AI can triage and prioritize, but someone still has to sign off on the results.
What the FDA clearances signal
The FDA has cleared both ATUSA and WRDensity, which matters beyond these two products. It sets a precedent for approving AI-based medical devices that rely on GPU-accelerated inference, and it raises the validation bar for anyone claiming clinical utility. For crypto AI projects that gesture at medical use cases without multicenter trials or real patient data, that bar is now visibly higher. The agency hasn't announced any enforcement action against crypto AI tokens, and there's no indication it's planning one. But the contrast between iSono's 3,200-patient study and the typical AI-token white paper is hard to miss.
ISono's multicenter study is the next concrete milestone to watch. Results from the UC Davis and Vanderbilt sites will determine whether the 28% sensitivity advantage holds up in a broader patient population — and whether payers start treating two-minute breast ultrasound as a standard option rather than a boutique one.
