FDA clears AI software that turns head and neck scans into 3D artery views

The FDA has granted 510(k) clearance to Deep Reasoning AI’s NeuroDR, software that turns CT angiography scans into three-dimensional views of head and neck arteries. The tool runs after a scan finishes and sends its output into hospital imaging systems. Deep Reasoning AI and imaging-services partner 3DR Labs now plan to supply it to US hospitals and imaging centers.

From a completed scan to artery views

NeuroDR processes neurovascular computed tomography angiography, or CTA, scans. It operates automatically in the background once scanning is complete. Its processing removes soft tissue, bone and veins from the displayed images, leaving views focused on the left and right internal carotid and vertebral arteries. The resulting visualizations are intended to help clinicians inspect those arteries.

According to Deep Reasoning AI, the software produces several image formats: 3D maximum intensity projections, 3D volume renderings and curved planar reformatted images of the arteries. Diagnostic Imaging describes the software as an adjunctive tool for viewing head and neck arteries.

The company also says NeuroDR separates the anterior and posterior circulations into distinct displays, allowing clinicians to examine each without overlapping anatomy. It delivers the visualizations directly to hospital picture archiving and communication systems as DICOM medical images.

The clinical role is assistance: physicians can use NeuroDR to help verify whether lesions they have identified are present. The software also supports evaluation, documentation, monitoring and treatment planning.

A 105-day path through regulatory review

Innolitics managed the regulatory pathway, including cybersecurity and software documentation, as well as interactions with the FDA. Those interactions included a live demonstration during an early orientation meeting.

The team bypassed a formal pre-submission meeting, a decision Medical Device Network attributed to familiarity with regulator expectations from earlier cleared technologies. Innolitics also designed the performance studies, while 3DR Labs conducted them.

Deep Reasoning AI was founded by Cornell University AI professors. The company describes its approach as combining deep learning with anatomical reasoning to construct patient-specific 3D models despite differences in anatomy, imaging artifacts or metal implants.