PURPOSE:To date, some studies have employed deep learning techniques to directly generate dynamic positron emission tomography (PET) parametric images from static PET. Compared with traditional methods, this approach requires only a single PET/computed tomography (CT) scan. However, current methods tend to employ static PET images captured at fixed scanning times without considering whether static PET images acquired at different scanning times impact the quality of the dynamic PET parametric images generated via deep learning. METHODS:A single-frame image from dynamic [Formula: see text]-FAPI-04 total-body PET can actually be regarded as a static PET image acquired during that specific period of the PET/CT scan. We extracted 5 frames of dynamic [Formula: see text]-FAPI-04 total-body PET at equal intervals, specifically frames 50, 76, 80, 86, and 92. Each frame of the dynamic [Formula: see text]-FAPI-04 total-body PET images was subsequently input into the deep learning model to obtain dynamic [Formula: see text]-FAPI-04 total-body PET parametric images. By comparing and analyzing the quality of dynamic PET parametric images, we attempted to determine the optimal scan time for static PET. RESULTS:The experimental results revealed that the dynamic [Formula: see text]-FAPI-04 total-body PET parametric image generated from the 58th frame of the dynamic PET image via the deep learning model exhibited the poorest performance. This may be because the diffusion of the radioactive tracer was not stable at the time of PET/CT scanning in frame 58, but it was relatively stable from frames 70-92. CONCLUSION:Static PET images acquired at different scanning times do indeed affect the quality of the dynamic [Formula: see text]-FAPI-04 total-body PET parametric images generated via deep learning. Specifically, when the radioactive tracer is unstable during the early stage of scanning, the quality of the dynamic [Formula: see text]-FAPI-04 total-body PET parametric images generated from static PET images via deep learning appears to be inferior.

To date, some studies have employed deep learning techniques to directly generate dynamic positron emission tomography (PET) parametric images from static PET. Compared with traditional methods, this approach requires only a single PET/computed tomography (CT) scan. However, current methods tend to employ static PET images captured at fixed scanning times without considering whether static PET images acquired at different scanning times impact the quality of the dynamic PET parametric images generated via deep learning. A single-frame image from dynamic ^68Ga -FAPI-04 total-body PET can actually be regarded as a static PET image acquired during that specific period of the PET/CT scan. We extracted 5 frames of dynamic ^68Ga -FAPI-04 total-body PET at equal intervals, specifically frames 50, 76, 80, 86, and 92. Each frame of the dynamic ^68Ga -FAPI-04 total-body PET images was subsequently input into the deep learning model to obtain dynamic ^68Ga -FAPI-04 total-body PET parametric images. By comparing and analyzing the quality of dynamic PET parametric images, we attempted to determine the optimal scan time for static PET. The experimental results revealed that the dynamic ^68Ga -FAPI-04 total-body PET parametric image generated from the 58th frame of the dynamic PET image via the deep learning model exhibited the poorest performance. This may be because the diffusion of the radioactive tracer was not stable at the time of PET/CT scanning in frame 58, but it was relatively stable from frames 70–92. Static PET images acquired at different scanning times do indeed affect the quality of the dynamic ^68Ga -FAPI-04 total-body PET parametric images generated via deep learning. Specifically, when the radioactive tracer is unstable during the early stage of scanning, the quality of the dynamic ^68Ga -FAPI-04 total-body PET parametric images generated from static PET images via deep learning appears to be inferior.

Abstract Purpose To date, some studies have employed deep learning techniques to directly generate dynamic positron emission tomography (PET) parametric images from static PET. Compared with traditional methods, this approach requires only a single PET/computed tomography (CT) scan. However, current methods tend to employ static PET images captured at fixed scanning times without considering whether static PET images acquired at different scanning times impact the quality of the dynamic PET parametric images generated via deep learning. Methods A single-frame image from dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET can actually be regarded as a static PET image acquired during that specific period of the PET/CT scan. We extracted 5 frames of dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET at equal intervals, specifically frames 50, 76, 80, 86, and 92. Each frame of the dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET images was subsequently input into the deep learning model to obtain dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET parametric images. By comparing and analyzing the quality of dynamic PET parametric images, we attempted to determine the optimal scan time for static PET. Results The experimental results revealed that the dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET parametric image generated from the 58th frame of the dynamic PET image via the deep learning model exhibited the poorest performance. This may be because the diffusion of the radioactive tracer was not stable at the time of PET/CT scanning in frame 58, but it was relatively stable from frames 70–92. Conclusion Static PET images acquired at different scanning times do indeed affect the quality of the dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET parametric images generated via deep learning. Specifically, when the radioactive tracer is unstable during the early stage of scanning, the quality of the dynamic $$^{68}{\textrm{Ga}}$$ -FAPI-04 total-body PET parametric images generated from static PET images via deep learning appears to be inferior.

Dynamic positron emission tomography (PET) parametric imaging typically requires a 60-min acquisition period, causing patient discomfort and reducing clinical efficiency. This study explores the feasibility of generating parametric K-i images from 10-min dynamic PET images acquired in the early or late scanning phases employing a multichannel feature fusion cold sampling (MCFFCoS) framework. PET data from 103 patients are acquired using the uEXPLORER total-body PET/CT scanner during 60-min scans. This study conducts deep learning experiments, taking early-phase or late-phase PET images as input, respectively. The generated K-i images are evaluated by visual quality and quantitative metrics, including root-mean-squared error (RMSE), structural similarity index (SSIM), and peak signal-to-noise ratio (PSNR). Volumes of interest (VOIs) analysis is performed using linear regression and Bland-Altman plots. In the quantitative evaluation of total-body data, the parametric K-i images generated from late-phase PET data generally outperform those derived from early-phase data. The analysis of VOIs indicates that the appropriate scanning protocol for PET parametric imaging may vary for different body regions. The deep learning approach is able to generate high-quality parametric K-i images from 10-min dynamic PET scans, bypassing the requirements of long acquisition time for the estimation of blood input function in kinetic modeling.

Dynamic Positron Emission Tomography (PET) parametric imaging plays a crucial role in the diagnosis of and research on tumors and neurological disorders. However, it requires long-term continuous PET/computed tomography (CT) scans, which significantly increase the complexity of imaging and have become one of the main limitations hindering its development and clinical application. To address this issue, we propose a novel approach, namely, a dynamic Ga-68-PSMA total-body PET dual-parametric imaging model based on W-Net with an improved diffusion model. We construct W-Net as the backbone network of the model. The differences in the shared downsampling module Phi(s) and middle layer networks, can be divided into three categories: W-Net 1, W-Net 2, and W-Net 3. Furthermore, we extend the cold diffusion model to generate single-class images to simultaneously produce dynamic Ga-68 -PSMA total-body PET K-1 parametric images. Compared to other methods, the K-1 parametric images achieved peak signal-to-noise ratio (PSNR) values improvement of 0.247-4.335 dB and mean-squared error (MSE) error reduction of 0.00026-0.01288; and for K-1 parametric images, PSNR and structural similarity index measure (SSIM) metrics were enhanced by 0.106-1.590 dB and 0.002-0.004, respectively, while MSE errors decreased by 0.00003-0.00078. The Pearson correlation coefficient (PCC) value between the generated and original images indicates that they have a strong positive correlation.

Multiorgan segmentation in total-body positron emission tomography (PET) images is crucial for accurately locating abnormalities and assisting in the observation of corresponding metabolic regions in the human body. Despite the emergence of numerous advanced methods in the field of multiorgan segmentation in recent years, available PET image segmentation techniques remain relatively limited. The complexity and variability of textures in PET images, the varying visibility and contrast of organs due to different metabolic activities, and the challenges posed by blurred organ boundaries in PET images all contribute to the increased difficulty of multiorgan segmentation. In this article, we propose the dual-prompt enhanced multiorgan segmentation model (DPESeg) for total-body PET image segmentation. Our approach focuses on enhancing the model's ability to perceive organ thresholds and shapes by introducing textual and disentangled organ features, thereby improving segmentation accuracy. We validate our model on a dataset of total-body PET images obtained from 110 patients. Both visual and quantitative results demonstrate that DPESeg performs well in the multiorgan segmentation task, with a 2.02% improvement in the Dice coefficient and a 1.90% improvement in the Jaccard index compared to the best-performing comparison algorithm.

Fast PET imaging is clinically important for reducing motion artifacts and improving patient comfort. While recent diffusion-based deep learning methods have shown promise, they often fail to capture the true PET degradation process, suffer from accumulated inference errors, introduce artifacts, and require extensive reconstruction iterations. To address these challenges, we propose a novel multistage diffusion framework tailored for fast PET imaging. At the coarse level, we design a multistage structure to approximate the temporal non-linear PET degradation process in a data-driven manner, using paired PET images collected under different acquisition duration. A Phase Error Correction Network (PECNet) ensures consistency across stages by correcting accumulated deviations. At the fine level, we introduce a deterministic cold diffusion mechanism, which simulates intra-stage degradation through interpolation between known acquisition durations—significantly reducing reconstruction iterations to as few as 10. Evaluations on [68Ga]FAPI and [18F]FDG PET datasets demonstrate the superiority of our approach, achieving peak PSNRs of 36.2 dB and 39.0 dB, respectively, with average SSIMs over 0.97. Our framework offers high-fidelity PET imaging with fewer iterations, making it practical for accelerated clinical imaging.

Dynamic total-body imaging enables new perspectives to investigate the potential relationship between the central and peripheral regions. Employing uEXPLORER dynamic [11C]CFT PET/CT imaging with voxel-wise simplified reference tissue model (SRTM) kinetic modeling and semi-quantitative measures, we explored how the correlation pattern between nigrostriatal and digestive regions differed between the healthy participants as controls (HC) and patients with Parkinson’s disease (PD). Eleven participants (six HCs and five PDs) underwent 75-min dynamic [11C]CFT scans on a total-body PET/CT scanner (uEXPLORER, United Imaging Healthcare) were retrospectively enrolled. Time activity curves for four nigrostriatal nuclei (caudate, putamen, pallidum, and substantia nigra) and three digestive organs (pancreas, stomach, and duodenum) were obtained. Total-body parametric images of relative transporter rate constant (R1) and distribution volume ratio (DVR) were generated using the SRTM with occipital lobe as the reference tissue and a linear regression with spatial-constraint algorithm. Standardized uptake value ratio (SUVR) at early (1–3 min, SUVREP) and late (60–75 min, SUVRLP) phases were calculated as the semi-quantitative substitutes for R1 and DVR, respectively. Significant differences in estimates between the HC and PD groups were identified in DVR and SUVRLP of putamen (DVR: 4.82 ± 1.58 vs. 2.58 ± 0.53; SUVRLP: 4.65 ± 1.36 vs. 2.84 ± 0.67; for HC and PD, respectively, both p < 0.05) and SUVREP of stomach (1.12 ± 0.27 vs. 2.27 ± 0.65 for HC and PD, respectively; p < 0.01). In the HC group, negative correlations were observed between stomach and substantia nigra in both the R1 and SUVREP values (r=-0.83, p < 0.05 for R1; r=-0.94, p < 0.01 for SUVREP). Positive correlations were identified between pancreas and putamen in both DVR and SUVRLP values (r = 0.94, p < 0.01 for DVR; r = 1.00, p < 0.001 for SUVRLP). By contrast, in the PD group, no correlations were found between the aforementioned target nigrostriatal and digestive areas. The parametric images of R1 and DVR generated from the SRTM model, along with SUVREP and SUVRLP, were proposed to quantify dynamic total-body [11C]CFT PET/CT in HC and PD groups. The distinction in correlation patterns of nigrostriatal and digestive regions between HC and PD groups identified by R1 and DVR, or SUVRs, may provide new insights into the disease mechanism.

While [68 Ga]Ga-FAPI-04 PET/CT is widely used in various malignant tumors diagnosis, its specificity is challenged by high uptake in benign lesions such as inflammatory lymphadenopathy, bone fractures, and degenerative changes. This study aimed to quantitatively assess and characterize the metabolic heterogeneity of [68 Ga]Ga-FAPI-04 uptake in benign and malignant lesions using dynamic total-body PET/CT. Dynamic total-body [68 Ga]Ga-FAPI-04 PET/CT scans (0–60 min post-injection) were performed on 17 oncology patients. Time-activity curves (TACs) were generated for benign and malignant lesions with high [68 Ga]Ga-FAPI-04 uptake. The reversible two-tissue compartment model (2T4k) was used to derive kinetic metrics, including K1, k2, k3, k4, vB and VT. Receiver operating characteristic (ROC) curve analysis was used to determine the cut-off values for differentiating benign and malignant lesions. The study included 58 malignant and 55 inflammatory lesions with high [68 Ga]Ga-FAPI-04 uptake. Malignant lesions exhibited higher K1 (0.277 ± 0.217 ml/ccm/min vs. 0.221 ± 0.216 ml/ccm/min, P = 0.011), vB (0.042 ± 0.007 vs. 0.013 ± 0.004, P < 0.001), and lower k3 (0.267 ± 0.041 1/min vs. 0.481 ± 0.085 1/min, P = 0.008) compared to benign lesions. Lesions were classified into low, medium, and high-probability groups for being malignant based on K1, k3 and vB values, with probabilities of 0

Background Conventional PET/CT imaging reconstruction is typically performed using voxel size of 3.0–4.0 mm in three axes. It is hypothesized that a smaller voxel sizes could improve the accuracy of small lesion detection. This study aims to explore the advantages and conditions of small voxel imaging on clinical application. Methods Both NEMA IQ phantom and 30 patients with an injected dose of 3.7 MBq/kg were scanned using a total-body PET/CT (uEXPLORER). Images were reconstructed using matrices of 192 × 192, 512 × 512, and 1024 × 1024 with scanning duration of 3 min, 5 min, 8 min, and 10 min, respectively. Results In the phantom study, the contrast recovery coefficient reached the maximum in matrix group of 512 × 512, and background variability increased as voxel size decreased. In the clinical study, SUV max , SD, and TLR increased, while SNR decreased as the voxel size decreased. When the scanning duration increased, SNR increased, while SUV max , SD, and TLR decreased. The SUV mean was more reluctant to the changes in imaging matrix and scanning duration. The mean subjective scores for all 512 × 512 groups and 1024 × 1024 groups (scanning duration ≥ 8 min) were over three points. One false-positive lesion was found in groups of 512 × 512 with scanning duration of 3 min, 1024 × 1024 with 3 min and 5 min, respectively. Meanwhile, the false-negative lesions found in group of 192 × 192 with duration of 3 min and 5 min, 512 × 512 with 3 min and 1024 × 1024 with 3 min and 5 min were 5, 4, 1, 4, and 1, respectively. The reconstruction time and storage space occupation were significantly increased as the imaging matrix increased. Conclusions PET/CT imaging with smaller voxel can improve SUV max and TLR of lesions, which is advantageous for the diagnosis of small or hypometabolic lesions if with sufficient counts. With an 18 F-FDG injection dose of 3.7 MBq/kg, uEXPLORER PET/CT imaging using matrix of 512 × 512 with 5 min or 1024 × 1024 with 8 min can meet the image requirements for clinical use.

Total-body dynamic positron emission tomography (PET) imaging with total-body coverage and ultrahigh sensitivity has played an important role in accurate tracer kinetic analyses in physiology, biochemistry, and pharmacology. However, dynamic PET scans typically entail prolonged durations ( ≥ 60 minutes), potentially causing patient discomfort and resulting in artifacts in the final images. Therefore, we propose a dynamic frame prediction method for total-body PET imaging via deep learning technology to reduce the required scanning time. On the basis of total-body dynamic PET data acquired from 13 subjects who received [68Ga]Ga-FAPI-04 (68Ga-FAPI) and 24 subjects who received [68Ga]Ga-PSMA-11 (68Ga-PSMA), we propose a bidirectional dynamic frame prediction network that uses the initial and final 10 min of PET imaging data (frames 1–6 and frames 25–30, respectively) as inputs. The peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) were employed as evaluation metrics for an image quality assessment. Moreover, we calculated parametric images (68Ga-FAPI: K_1 , 68Ga-PSMA: K_i ) based on the supplemented sequence data to observe the quantitative accuracy of our approach. Regions of interest (ROIs) and statistical analyses were utilized to evaluate the performance of the model. Both the visual and quantitative results illustrate the effectiveness of our approach. The generated dynamic PET images yielded PSNRs of 36.056 ± 0.709 dB for the 68Ga-PSMA group and 33.779 ± 0.760 dB for the 68Ga-FAPI group. Additionally, the SSIM reached 0.935 ± 0.006 for the 68Ga-FAPI group and 0.922 ± 0.009 for the 68Ga-PSMA group. By conducting a quantitative analysis on the parametric images, we obtained PSNRs of 36.155 ± 4.813 dB (68Ga-PSMA, K_1 ) and 43.150 ± 4.102 dB (68Ga-FAPI, K_i ). The obtained SSIM values were 0.932 ± 0.041 (68Ga-PSMA) and 0.980 ± 0.011 (68Ga-FAPI). The ROI analysis conducted on our generated dynamic PET sequences also revealed that our method can accurately predict temporal voxel intensity changes, maintaining overall visual consistency with the ground truth. In this work, we propose a bidirectional dynamic frame prediction network for total-body 68Ga-PSMA and 68Ga-FAPI PET imaging with a reduced scan duration. Visual and quantitative analyses demonstrated that our approach performed well when it was used to predict one-hour dynamic PET images. https://github.com/OPMZZZ/BDF-NET .

Purpose: Parkinson’s disease (PD) is a neurodegenerative disease characterized by progressive loss of dopaminergic neurons in the brain. To achieve better explorations of dopamine changes both centrally and peripherally, we employed uEXPLORER dynamic [11C]CFT PET/CT imaging combined with voxel-wise kinetic modeling. Methods: Eleven participants (five patients, PD and six healthy volunteers, HC) underwent 75-min dynamic scans were enrolled. Volumes of interest for four nigrostriatal nuclei (caudate, putamen, pallidum and substantial nigra) and three digestive organs (pancreas, stomach and duodenum) were delineated. Total-body parametric images of relative transporter rate constant (R1) and distribution volume ratio (DVR) using the simplified reference tissue model (SRTM2) were quantitatively generated by a linear regression with spatial-constraint algorithm. Standardized uptake value ratio (SUVR) at early and late phase were calculated as the semi-quantitative substitutes. Results: Significant differences between the two groups were identified in DVR and SUVRLP of putamen (P < 0.05) and SUVREP of stomach (P < 0.01). For HC group, negative correlations of R1 were achieved between stomach and both putamen and substantial nigra (all P < 0.05); positive correlations of DVR were identified between pancreas and all four brain nuclei (all P < 0.05). Yet in PD group, correlations of R1 or DVR between the targeted digestive and brain areas were considerably diminished. Similar trends in correlations were also found in SUVR analysis. Conclusions: We introduced a pioneering approach using dynamic total-body [11C]CFT PET/CT imaging to investigate distinctive patterns of potential “brain-GI” interplays, which may provide new insights towards the understanding of PD.

Fibroblast activation protein inhibitor (FAPI) is an ideal diagnostic and therapeutic target in malignant tumors. However, the knowledge of kinetic modeling and parametric imaging of 68Ga-FAPI is limited. Purpose: The purpose of this study was to explore the pharmacokinetics of 68Ga-FAPI-04 PET/CT in pancreatic cancer and gastric cancer and to conduct parametric imaging of dynamic total-body data compared with SUV imaging. Methods: Dynamic total-body 68Ga-FAPI-04 PET/CT was performed on 13 patients. The lesion time-activity curves were fitted by 3-compartment models and multigraphical models. The kinetics parameters derived from the 2-tissue reversible compartment model (2T4K) and multigraphical models were analyzed. Parametric [Formula: see text] imaging was generated using the 2T4K and Logan models, and their performances were evaluated compared with SUV images. Results: 2T4K had the lowest Akaike information criterion value, and its fitting curves matched excellently with the origin time-activity curves. Visual assessment revealed that the [Formula: see text](2T4K) images and [Formula: see text](Logan with spatial constraint [SC]) images both showed less image noise and higher lesion conspicuity compared with SUV images. Objective image quality assessment demonstrated that parametric [Formula: see text](2T4K) images and parametric [Formula: see text](Logan with SC) images had a 5.0-fold and 5.0-fold higher average signal-to-noise ratio and 3.6-fold and 4.1-fold higher average contrast-to-noise ratio compared with conventional SUV images, respectively. In addition, no significant differences in signal-to-noise ratio and contrast-to-noise of pathologic lesions were observed between parametric [Formula: see text](2T4K) images and parametric [Formula: see text](Logan with SC) images (all P > 0.05). Conclusions: The 2T4K model was the preferred compartment model. Total-body parametric imaging of 68Ga-FAPI-04 PET yielded superior quantification beyond SUV with enhanced lesion contrast, which may serve as a promising imaging method to make an early diagnosis, to better reflect tumor characterization, or to allow evaluation of treatment response. [Formula: see text](2T4K) images are comparable in image quality and consistent to [Formula: see text](Logan with SC) images in lesions conspicuity; however, [Formula: see text](Logan with SC) images presented an appealing alternative to [Formula: see text](2T4K) images because of their simplicity.

The neural mechanisms underlying the neuropsychiatric disorders suffered by Long-Covid-19 patients remain unknown. Non-invasive [18F]-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) brain imaging is an effective way to study this issue. In the present study, based on 18F-FDG PET brain imaging and analysis of the brain metabolic network, we found significant reductions in metabolic levels in the frontal, temporal, and parietal lobes in patients with Long-Covid-19. A significant reduction in metabolic connectivity between temporal and frontal networks in patients with Long-Covid-19 may be the underlying neural mechanism underlying cognitive impairment.