Bijoy Kundu, PhD
A team led by Bijoy Kundu, PhD, a professor in the Department of Radiology and Medical Imaging, has developed iD-PET, an artificial intelligence-driven dynamic PET imaging software platform designed to transform presurgical planning for patients with drug-resistant epilepsy. Kundu has developed the platform in collaboration with Mark Quigg, MD, a professor in the Department of Neurology. The condition affects nearly a third of the 1.6% of Americans living with epilepsy whose seizures remain uncontrolled despite medication. Despite strong evidence supporting epilepsy surgery as a transformative treatment, fewer than 1% of eligible patients undergo the procedure, largely because standard imaging fails to identify the seizure focus in nearly half of all cases.
iD-PET addresses that gap by moving beyond conventional static FDG-PET imaging to map the dynamic rate of glucose uptake, metabolism and decay in the brain, revealing epileptic zones invisible to standard techniques. Preliminary results show iD-PET detects focal abnormalities in patients whose standard PET scans appear normal, with the potential to significantly expand access to life-changing surgical treatment.
The team also built a deep learning pipeline that derives a patient’s blood input function directly from imaging, removing the need for the invasive arterial blood sampling that dynamic PET has traditionally required.
With support from the Ivy Biomedical Innovation Fund, the UVA Brain Institute, and the Virginia Innovation Partnership Corporation’s Commonwealth Commercialization Fund, the team enrolled 50 subjects and developed novel machine learning and quantification algorithms that generate clinician-ready parametric brain maps for non-invasive, automated epileptic zone localization.
Now backed by a renewed $100,000 Ivy Biomedical Innovation Fund award granted May 1, 2026, the team is advancing iD-PET into a fully autonomous, AI-driven, turnkey clinical software platform, with plans to pursue FDA Breakthrough Device designation.
Filed Under: Research