Emilie Roncali, “Searching for Speed: Fast Radiation Detectors and Simulations for Nuclear Imaging”

Overcoming the symbolic barrier of 100 ps, time-of-flight detectors transformed Positron Emission Tomography (PET) imaging, enabling precise localization of radiotracers and approaching reconstruction-less PET. Such a leap in performance stems from breakthrough technologies (new photodetectors, electronics) and harnessing ultrafast Cerenkov photons.

Meanwhile, clinical implementation of radiopharmaceutical therapy drives new requirements in single photon emission computed tomography (SPECT) to image better and faster. Next-generation radiation detectors must deliver greater quantification, sensitivity to allow fast low count imaging, be compatible with high energy isotopes, and capture larger fields of view— all critical for therapeutic monitoring.

Monte Carlo simulations are essential for radiation detector design, modeling particle interactions, energy deposition, and scintillation processes across complex geometries. These tools give access to information not captured experimentally to optimize critical detector parameters including light propagation, collection efficiency, timing performance, and energy resolution prior to prototyping. System-level simulations are also needed to evaluate how detector innovation impacts overall system performance and ultimately image contrast and signal-to-noise ratio.

However, each high-fidelity model represents a massive computational burden. A simulation tracking individual photons through scintillators, optical interfaces, and photodetectors can consume hundreds of CPU-hours for statistically meaningful results, thus prohibiting comprehensive system simulations. Generative AI solutions can offer unprecedented acceleration compared to traditional Monte Carlo method, and showed promise in SPECT and optical simulation applications, with massive speed gains while maintaining accuracy.
This presentation will discuss recent accomplishments in radiation detectors and simulations as well as new technology demands creating exciting opportunities for research in radiation detection technology.

Dr. Emilie Roncali is an Associate Professor in the Department of Biomedical Engineering and a courtesy appointment in Radiology at UC Davis. She received her Ph.D. in Biomedical Engineering from the Ecole Centrale de Paris, France before joining Dr. Simon Cherry’s group at UC Davis for her postdoctoral training. Dr. Roncali’s research involves radiation detector development and simulations, with an emphasis on nuclear imaging, dosimetry for cancer radiopharmaceutical therapy (liver cancer radioembolization and lutetium 177 therapies). Dr Roncali’s current research on radiation detectors focuses on time-of-flight PET detector modeling, Cerenkov light, and AI-accelerated Monte Carlo simulations of radiation detectors. Her group has pioneered the optiGAN, a generative adversarial network framework to accelerate optical tracking in detectors, which is freely available within the simulation software GATE. She serves in the GATE steering committee since 2013, and her group was the third largest contributor to the development of the new python GATE released in December 2024. Dr. Roncali is also a member of two working groups for the National Cancer Institute (NCI) and of the Society of Nuclear Medicine and Molecular Imaging AI Dosimetry taskforce. Her contributions to women in medical imaging sciences have been recognized in 2022 with the Faber award from the Society of Nuclear Medicine and Molecular Imaging.