Researchers at Washington University in St. Louis have built a portable PET scanner that can deliver real-time images at a patient's bedside—a breakthrough that could transform how hospitals guide biopsies, tumor ablations, and other minimally invasive procedures.
Current interventional procedures rely mainly on anatomical imaging like ultrasound, X-ray fluoroscopy, and CT scans for guidance. Studies have shown that interventional radiology procedures guided by dedicated PET/CT systems achieve significantly higher accuracy, but the cost is prohibitive for most hospitals, leaving this precision tool largely unavailable. The new portable point-of-care PET system, with its robotic arm that positions detector panels at arbitrary locations to image any organ, directly addresses this gap.
The technology works through an elegant approach to image reconstruction. Rather than waiting until a full scan is complete to generate images—the conventional method—the portable system updates images in real time as data arrives. Researchers demonstrated this in a phantom study with three clusters of radiotracer-filled rods imaged from six user-selected positions. They began with five iterations using data from the first position, then added single-iteration updates as each new scanning position provided fresh data. Because data acquisition time far outpaced reconstruction time, images continuously improved and became clearer as scanning proceeded.
The results were striking: phantom structures became clearly distinguishable after just three to four positions, comparable in quality to conventional reconstruction methods. This suggests that operators could terminate scanning early if imaging tasks are already fulfilled, or continue acquiring additional positions to further improve image quality—a flexible, adaptive workflow that fits real clinical environments rather than rigid laboratory protocols.
"A portable PET device with real-time imaging capability could bring vast information and benefits from molecular imaging to interventional radiology procedures," said Yuan-Chuan Tai, Ph.D., senior author at Washington University. His team engineered the system with a robotic arm that enables clinicians to position detector panels wherever needed, imaging any organ of interest. Xiyan Li, a graduate researcher in the university's Imaging Science doctoral program, emphasized that this represents more than incremental improvement. "This proposed approach better supports interactive and adaptive imaging workflows at the bedside," Li said. "It represents a paradigm shift that offers new avenues to deploy novel molecular imaging applications."
The current study used a benchtop prototype, presented at the Society of Nuclear Medicine and Molecular Imaging 2026 Annual Meeting. But the team is already building a more refined prototype suitable for initial human imaging studies, with those trials set to begin in 2027. That timeline means patients could soon benefit from molecular imaging guidance that was previously confined to major medical centers—bringing higher accuracy interventional procedures to a far wider range of hospitals and clinical settings.