2025 COBRA @ UC Davis

Retrospective and Universal Phantom-Based Calibration of 123I- and 131I-MIBG SPECT Imaging for Quantitative Dosimetry in 131I-MIBG Therapy

Yiran Wang, Brahim Mehadji, Emilie Roncali, Youngho Seo

Purpose: SPECT imaging of pretherapy 123I-MIBG and posttherapy 131I-MIBG is routinely used in 131I-MIBG therapy. However, the acquired images are usually unitless, limiting their utility for imaging-based dosimetry. This work develops a phantom-based calibration method to convert unitless SPECT images into quantitative (Bq/mL) maps for dosimetric purposes. Methods: We performed SPECT phantom scans of NaI solutions of 123I and 131I. Each acquisition was conducted in planar imaging mode. The counts from the phantom images were matched to the known phantom activity to establish a calibration factor, which was then applied to human planar images to determine the total-body activity. The human 3D SPECT images were normalized with the body activity. Results: The calibrated total-body activity ranged 1.8-55 MBq for 123I and 38-1400 MBq for 131I among the 25 scans, enabling conversion of SPECT images into Bq/mL without having to acquire scanner-specific calibration factors prospectively. Conclusion: We developed a phantom-based calibration method to transform unitless SPECT images into quantitative maps. It facilitates imaging-based dosimetry, especially for procedures where imaging is not originally intended for dosimetric analysis.

Photonic Crystals’ effect on Scintillation and Cherenkov photons in scintillator detector

Xuzhi He, Carlotta Trigila, Emilie Roncali

High-Sensitivity Gamma Imaging for 225Ac-Based Targeted Alpha Therapy

Biswajit Das, Baharak Mehrdel, David Goodman, Michael Streicher, Youngho Seo, and Javier Caravaca

225Ac-based radiopharmaceuticals show promise for targeted alpha therapy (TAT) in cancer treatment, but gamma-ray imaging is challenging due to low injected activities and weak branching emissions. In-vivo small-animal imaging is crucial for studying biokinetics and accelerating drug development, yet current SPECT scanners have sensitivity limitations requiring long imaging sessions. Our study explores Compton imaging as an alternative, using a 3D-positioning cadmium zinc telluride (CZT) camera (M400, H3D) for higher sensitivity imaging of 225Ac daughters, 213Bi (440 keV) and 221Fr (218 keV). The system achieved high Compton sensitivity for 213Bi and 221Fr. Imaging of 225Ac sources at sub-μCi levels was demonstrated within a short scan session, while a mouse phantom study showed effective imaging of 213Bi and 221Fr. Our results suggest tumor uptake in-vivo imaging with our system could be achieved in 23 minutes and 90 minutes for 213Bi and 221Fr, respectively. These exposures correspond to a single detector head system and that a full system with four heads will further decrease exposure time.

SHARE-MI – A Roadmap Towards Affordable High-Quality Nuclear Medicine Imagers

Reimund Bayerlein

Recent advancements in positron emission tomography (PET) have led to the development of high-performance total-body and long axial field-of-view (AFOV) scanners, significantly increasing sensitivity and improving image quality. However, these novel imaging systems come with substantial challenges, including high costs for procurement, maintenance, computation, and data storage.

Initiatives like SHARE-MI aim to address these challenges by exploring pathways to make high-performance PET scanners more affordable and widely available, particularly for smaller institutes and low- and middle-income countries (LMICs). By optimizing design, cost structures, and resource allocation, these efforts seek to preserve much of the performance and quality of cutting-edge PET technology while improving accessibility.

This poster aims to stimulate discussion and provide insights into various strategies for mitigating the financial burden associated with modern PET imaging systems. Through this dialogue, we hope to contribute to the ongoing efforts of the SHARE-MI initiative and support the broader adoption of high-performance PET imaging worldwide.

Total-body parametric PET with free-state tracer kinetic modeling: application to imaging of metabolic dysfunction-associated steatohepatitis (MASH) with [18]F-FDG

Quyen Tran, Victoria Lyo, Karen E. Matsukuma, Michael T. Corwin, Benjamin A. Spencer, Valentina Medici, Terry Jones, Simon R. Cherry, Ramsey D. Badawi, Souvik Sarkar, and Guobao Wang

Imaging metabolism is highly significant in metabolic dysfunction-associated steatohepatitis (MASH) which affects about 20 - 30% of the population worldwide with a hallmark of liver inflammation. As compared to static [18]F-fluorodeoxyglucose (FDG) PET imaging, dynamic FDG-PET may improve metabolic quantification by accounting for free-state FDG but is not sensitive to characterize liver metabolism in MASH. Herein, we demonstrate the use of a three-tissue model and the need for high-temporal resolution (HTR) to improve compartmental resolution by further separating free-state FDG in the interstitial space and intracellular space. This HTR method is applied to improve characterization of liver metabolism and systemic evaluation of extrahepatic organs in patients with MASH, demonstrating a unique application of total-body dynamic PET in a major disease.

Acknowledgement: This work is supported in part by NIH under grant R01 DK124803 and R01 CA206187.

Total-Body Multiparametric Dynamic PET: Towards Integrated Multisystem Biomedical Research

Kevin J. Chung, Terry Jones, Abhijit J. Chaudhari, Lorenzo Nardo, Yasser G. Abdelhafez, Benjamin A. Spencer, Negar Omidvari, Quyen Tran, Moon S. Chen Jr, Souvik Sarkar, Valentina Medici, Victoria Lyo, Ramsey D. Badawi, Simon R. Cherry, Guobao Wang

It is increasingly recognized that our most prevalent health challenges, including heart and brain diseases, have multisystemic consequences that exacerbate outcomes. However, comprehensive evaluation remains challenging due to the lack of methods to simultaneously assess multiple systems and pathologies. Total-body positron emission tomography (PET) now allows dynamic imaging of the entire human body with unprecedented sensitivity, offering a new approach to study diseases across systems and scales. Here we develop total-body multiparametric imaging by leveraging the ultra-high performance of total-body PET as a clinical platform for integrated disease assessment. Our approach uses high-temporal resolution imaging and advanced kinetic modeling enabled by total-body PET to quantify blood flow, capillary permeability, and molecular target interactions from a single dynamic scan. We demonstrate its potential for parametric imaging across systems and scales, from total-body blood flow to blood-brain barrier permeability and molecular interactions. Total-body multiparametric PET may therefore allow more holistic studies over conventional approaches, facilitating systems-based perspectives on traditionally single-organ diseases.

Dual-Ended Readout TOF-DOI PET Detectors

Haibo Wang, Jiahao Xie, Jinyi Qi, Simon R. Cherry, and Junwei Du

Two types of time-of-flight (TOF) and depth-of-interaction (DOI) positron emission tomography (PET) detectors based on the dual-ended readout method and lutetium–yttrium oxyorthosilicate (LYSO) arrays with two different pitches were introduced. Both detectors were equipped with 8 × 8 silicon photomultiplier (SiPM) arrays with a 3.2 mm pitch coupled to both ends of 8 × 8 LYSO arrays with a 3.2 mm pitch or 16 × 16 LYSO arrays with a 1.6 mm pitch. All crystals in the LYSO arrays were clearly resolved. Detectors based on the 8 × 8 LYSO arrays achieved an average coincidence time resolution (CTR) of 207 ± 5 ps and an average DOI resolution of 3.9 ± 0.6 mm across the crystals. In contrast, detectors with the 16 × 16 LYSO arrays provided an average CTR of 218 ± 6 ps and an average DOI resolution of 2.6 ± 0.3 mm. The TOF-DOI PET detectors developed in this work are promising candidates for whole-body PET and brain PET scanners, which can offer uniform spatial resolution across the field-of-view (FOV).

Deep Learning-Based Optical Photon Transport with optiGAN: An Alternative to Photon Tracking in GATE 10

Guneet Mummaneni, Carlotta Trigila, Nils Krah, David Sarrut, Emilie Roncali

Optical photon transport simulations are essential for radiation detector development in medical imaging and high-energy physics. While Monte Carlo methods are the gold standard, their high computational cost remains a challenge. This study integrates optiGAN, a GAN-based model for optical photon generation, into GATE 10, the Python-based version of the GATE simulation framework released in November 2024, to accelerate simulations while maintaining accuracy. After validating optical transport components by comparing GATE 9.3 and GATE 10 under identical conditions, results confirmed equivalent performance. Further comparison between full Monte Carlo GATE 10 and GATE 10-optiGAN showed excellent agreement, with Jensen-Shannon similarity exceeding 92% across multiple photon transport parameters. optiGAN accurately captured multimodal photon trajectories, reinforcing its efficiency. Computational analysis demonstrated a ~50% reduction in execution time with GATE 10-optiGAN. This work validates optical simulations in GATE 10 and AI-accelerated photon tracking, enabling large-scale optical simulations.

High-Performance Dual-Ended SiPM Readout for TOF-PET With BGO and LYSO:Ce

Nicolaus Kratochwil, Emilie Roncali, Joshua W. Cates, Gerard Arino-Estrada

Detection time performance is a key aspect for time-of-flight positron emission tomography. With recent advancement in SiPM technology and fast readout electronics, one limiting factor on timing performance is light transport in the crystal. Low-noise and high-frequency dual-ended SiPM readout can be employed to mitigate the aforementioned challenges and has the potential to combine ultra-fast timing with highest gamma-ray detection efficiency. We have studied the timing properties of cerium-doped lutetium-yttrium-oxyorthosilicate (LYSO:Ce) and bismuth germanate (BGO) in a symmetric dual-ended SiPM readout configuration. Coupling 3x3x20mm³ polished BGO crystals to Broadcom SiPMs a CTR of 234ps FWHM (harmonic average) was obtained for all photopeak events. For same-sized LYSO:Ce crystals, the measured CTR value is 98ps, which is in excellent agreement with analytic calculations on the timing limits considering scintillation properties and modeling of light transport. The results demonstrate significant timing improvement with dual-ended readout, both for Cherenkov photons in BGO and for standard scintillation for enhanced diagnostic accuracy in PET imaging.

Design and Simulation of a Scalable Direct X-ray Detector Readout IC

Daniel Fiallo, Michael Farrier, Shiva Abbaszadeh

This work presents the design and simulation of a scalable CMOS readout integrated circuit (ROIC) for direct X-ray detection and VisNIR, enabling high-resolution imaging with micron-scale pixels. The proposed 6T pixel architecture supports two exclusive detection methods: electron collection using a PIN photodiode and hole collection via an amorphous selenium (a-Se) photoconductor layer. A three-side tileable floorplan allows for seamless expansion of detector area, supporting large-scale imaging applications. Simulation results demonstrate the feasibility of achieving low-noise performance, high quantum efficiency, and frame rates exceeding 30 fps with 2×2 binning. The modular system architecture, including FPGA-based readout control and high-speed data acquisition, supports scalable tiled detector configurations with areas exceeding 11 × 23 cm² and resolutions greater than 67 megapixels per tile. This platform is designed to address emerging needs in medical imaging, non-destructive testing, and high-energy physics experiments.

A preliminary phantom study on a dual-panel CZT-based PET system

Kimia Gholami, Mohammadreza Mohseni, Shiva Abbaszadeh

The cadmium zinc telluride (CZT)--based positron emission tomography (PET) scanner, developed at the RIL laboratory at UC Santa Cruz, features a two-panel design with cross-strip, edge-on CZT detectors. This edge-on configuration achieves a high energy resolution of 5.28% at 511 keV. This work presents preliminary results from the experiment, which were obtained without any energy corrections, depth calibration, or other adjustments.

Multi-MeV Gamma Interaction Position Estimation in Large Pure Cherenkov Emitter Crystals

L. Rebolo, M. Backfish, E. Prebys, S. St. James, P.M. Mendes Correia, A.L. Silva1, J. Veloso, G. Ariño-Estrada

In this work, we study the Chernekov light created by prompt-gammas in a monolithic and pure Chernekov emitter crystal to achieve position estimation for collimated Prompt-gamma Imaging (PGI). This approach relies in the exclusive use of Cherenkov ligth and could lead to a novel system concept with an effective background discrimination at a relatively low production cost. The detector consisted of a 25x25x10 mm3 PbF2 crystal coupled to a 8x8 SiPM array and was mounted on a 3D automated stage. The collimator consisted of a 100x100x100 mm3 block with a 1 mm slit to collimate gammas from the Th-228 source. The source and collimator remained fixed and the detector was shifted in increments of 1 mm in the Y-axis direction. 36 positions were acquired, with 2 hour acquisition per position. Two different reconstruction algorithms were tested: Center of Gravity (CoG) and Rise to a Power (RTP). Results showed that Cherenkov light alone can provide spatial resolution, with the RTP achieving the best performance. Future work will focus on optimizing event reconstruction methods, improving SNR and evaluating the performance of this detector concept with different crystal configuration, higher-energy gamma sources and clinical proton beams.