Dynamic [11C]-Pittsburgh compound B (PiB) PET provides temporal information for studying amyloid-related diseases, though the long acquisition results in patient discomfort and additional operational costs. We proposed Modeling ODEs with Visually Interpretable Evolution (MOVIE), a generative framework for synthesizing late-phase [11C]-PiB PET images from a short 1-min early-phase acquisition, such that multiple biomarker information can be obtained in a single session. 78 dynamic PiB PET scans from a mixed cohort of cerebral amyloid angiopathy, Alzheimer’s disease, and hypertension patients were used to obtain the early- and late-phase images. We incorporated neural ordinary or stochastic differential equations (ODEs/SDEs) in MOVIE as well as utilizing mid-phase frames to guide image prediction. The metrics PSNR, SSIM, LPIPS, effect size, and SUVR-based regional analyses were used for quantitative evaluation; classification performance for amyloid positivity was assessed using ROC AUC and decision-curve analysis (DCA). The step-by-step synthesis of images from the early- to the late-phase was visualized. MOVIE achieved quantitative gains with improved metrics and outperformed a UNet baseline model (highest PSNR: 41.46 ± 2.96 and SSIM: 0.969 ± 0.0093 for the SDE model with effect size > 0.8). Regarding classification performance, the covariate-adjusted AUC for amyloid positivity in the global Centiloid composite ROI was higher in the MOVIE (SDE) implementation (AUC = 0.881) than the baseline, and also yielded a net benefit above the baseline under DCA. Visualizations of the step-by-step generative trajectories showed biologically plausible early-to-late evolution of uptake spatial patterns. MOVIE is, to our knowledge, the first framework to couple early-to-late PET synthesis with a temporally continuous, physiology-guided generative model. By moving beyond black-box regression, MOVIE shows potential to reduce scan time, alleviate hospital operations, improve interpretability, and support clinical decision-making in amyloid PET imaging.
Radiation emitted by therapeutic radionuclides deposit energy over micrometer-scale distances, yet practical patient dosimetry for radiopharmaceutical therapy can only currently be performed using gamma camera images whose resolution limits are much larger. Gamma camera-based dosimetry calculations may be adequate for some tissues, where the heterogeneity of the actual source microdistribution is not impactful. However, the difference in spatial scale between what can be measured and the actual radiopharmaceutical microdistribution may obscure biologically relevant dose heterogeneity, e.g., within tumors or normal organs that non-uniformly express the molecular target. This is true for widely used β-emitting radionuclides such as 177Lu that exhibit low linear energy transfer and particle ranges of hundreds of micrometers, enabling cross-fire radiation between neighboring cells. The challenge becomes more pronounced for α-emitters such as 211At that deposit energy over shorter distances of only tens of micrometers with LET values two to three orders of magnitude higher than β-particles. Understanding how radiation dose is impacted by the complex interactions among radionuclide emission characteristics, tumor microarchitecture, intercellular spacing, and heterogeneous targeting requires modeling frameworks capable of resolving radiation transport at cellular scales within multicellular tumor environments. At the single-cell level, the self-dose to the nucleus per decay was approximately 120-fold higher for 211At than for 177Lu due to the higher LET. Energy confinement analyses demonstrated that more than 90
to simulate amyloid PET (A-PET) studies using an anatomical brain phantom (StepBrain—SB) with varying gray-to-white matter ratios (GWR), to quantify partial volume effects (PVE)-related bias and assess the accuracy of PVE correction (PVEc) methods. SB was scanned to simulate clinical A-PET studies across GWR from 0 to 120
The full-ring semiconductor SPECT/CT systems utilize a focus mode to enhance image quality and reduce acquisition time. However, the impact of arm positioning on this mode remains under-investigated. This study evaluated the effects on image quality during electrocardiography-gated myocardial perfusion SPECT acquisition under different imaging modes. We used an electrocardiography-gated myocardial phantom with the Tc-99 m solution adjusted to simulate myocardial concentration. The upper limb portion of the whole-body phantom was attached to simulate the arms-down position and removed to simulate the arms-up position. Acquisitions were performed in both the arms-up and arms-down positions using two imaging modes: a uniform mode covering the body contour and a focus mode targeting the heart. Contrast analysis, polar map assessment, and cardiac functional analysis were performed and compared across different conditions. Under all conditions, the number of myocardial counts was higher in Uniform Mode than in Focus Mode; however, the number of ventricular cavity counts was significantly reduced in Focus Mode. When using CTAC, the mean myocardial count in Focus Mode was 11,075 ± 1064 (arms raised) and 9467 ± 843 (arms lowered), whereas the ventricular cavity counts were minimized to 261 ± 191 and 631 ± 318, respectively. As a result, Focus Mode consistently achieved superior contrast ratios, recording a maximum value of 0.98 ± 0.03 in the arm-raised position using CTAC + SC. Polar map analysis revealed that focus mode improved myocardial uniformity and mitigated uptake reductions in the septal and lateral walls compared to uniform mode, particularly under the arms-down condition. Quantitative gated SPECT (QGS) showed that while EDV was underestimated across all modes, focus mode provided EF values (27.7 [range:27.6–27.8]
The introduction of ring-type semiconductor single-photon emission computed tomography (SPECT)/computed tomography (CT) systems has generated growing interest in the clinical application of dynamic SPECT imaging. Nevertheless, the efficacy of dynamic SPECT in clinical practice has not yet been sufficiently demonstrated. This study assessed the imaging characteristics of dynamic SPECT using a ring-type semiconductor SPECT/CT system (ring-type system) through phantom experiments and determined the optimal parameters for image reconstruction and Becquerel calibration factors (BCFs). Two 40-mL balloon tubes were placed at the central and periphery of a cylindrical phantom. A Tc-99m solution was infused at a constant rate using a syringe pump to simulate dynamic changes in radioactivity over time. Dynamic SPECT imaging was performed using a ring-type system, and dynamic planar imaging was also performed using an Anger-type system for comparison. Radioactivity concentrations in the balloons were measured during dynamic imaging to generate time–activity curves (TACs). Dynamic SPECT data were reconstructed using acquisition times of 5 and 15 s and eight different values of β (the noise suppression parameter in the Bayesian penalty likelihood method): 0.005–0.2. The validity of the quantitative evaluation was examined, including the necessity of specific conversion factors for radioactivity concentration. The optimal reconstruction parameters for dynamic SPECT were explored by comparing the TACs obtained from the two systems. The TACs obtained from planar imaging demonstrated superior linearity compared to those obtained from SPECT. The TACs in SPECT tended to show significant variability, particularly during short acquisition times. In contrast, when two adjacent balloons overlapped, planar imaging produced a merged curve, whereas SPECT resolved them separately. A comparison of images at various β values revealed that under low-count-rate conditions, high β values (0.12–0.2) resulted in excessive noise suppression and signal loss. Regarding the BCF used to calculate the radioactivity concentration from the reconstructed images, its value was highly dependent on the acquisition and reconstruction conditions. The results closest to the theoretical values were obtained under conditions with a β value of 0.02. The TACs obtained using a ring-type system in dynamic SPECT demonstrated reduced linearity and greater fluctuations than those acquired with an Anger-type system. We successfully demonstrated that dynamic SPECT enables accurate separation and measurement of targets even when two adjacent targets overlap in planar images. We also identified image reconstruction parameters and BCFs that are appropriate for dynamic SPECT.
