Artificial intelligence-generated content (AIGC) has shown remarkable performance in nuclear medicine imaging (NMI), offering cost-effective software solutions for tasks such as image enhancement, motion correction, and attenuation correction. However, these advancements come with the risk of hallucinations, generating realistic yet factually incorrect content. Hallucinations can misrepresent anatomic and functional information, compromising diagnostic accuracy and clinical trust. This paper presents a comprehensive perspective on hallucination-related challenges in AIGC for NMI, introducing the DREAM report, which covers recommendations for definition, representative examples, detection and evaluation metrics, and attributions and mitigation strategies. This position statement paper aims to initiate a common understanding for discussions and future research toward enhancing AIGC applications in NMI, thereby supporting their safe and effective deployment in clinical practice.

Total-body PET systems enable reliable non-invasive extraction of image-derived input functions (IDIFs) by simultaneously imaging major blood pools and target organs. In brain imaging, a time delay in tracer arrival between the IDIF site (e.g., ascending aorta) and the brain tissue may affect time activity curve (TAC) fitting and bias kinetic parameter estimates if neglected. While time delay correction (TDC) has been used in total-body dynamic PET, its specific effect on brain kinetic quantification and the impact on studying the brain in systemic disease populations, such as cancer patients with exposure to chemotherapy, has not been well demonstrated. In this study, we evaluate the effect of TDC on brain TAC fitting and examine its influence on brain kinetic quantification of 18F-fluorodeoxyglucose (FDG) and group differentiation in comparing cancer patients with healthy subjects. A total of 20 individuals (13 healthy and 7 lung cancer) were included in the study and underwent dynamic FDG imaging using the uEXPLORER total-body PET/CT system. A two-tissue irreversible compartmental model with TDC was implemented to analyze FDG kinetics, including the fractional blood volume vb, delivery rate K1, and net influx rate Ki in different brain regions. Results show that TDC improves early phase brain TAC fitting, increases vb estimates closer to an expected physiological range, and reduces K1 overestimation. Importantly, TDC enhances the detection of group differences in K1. These findings highlight the importance of TDC for improving the accuracy of brain kinetic quantification, particularly for measuring tracer transport, contributing to a tool for studying the central nervous system in cancer patients and other clinical populations.

Imaging the blood-brain barrier (BBB) permeability of molecular PET tracers may allow pathway-specific assessment of the diverse transport mechanisms expressed at the BBB. However, PET quantification of BBB permeability typically requires dual-tracer protocols involving a flow tracer, increasing methodological complexity and clinical burden. We evaluated a single-tracer dynamic PET method for quantifying the BBB permeability-surface area (PS) of [18F]fluorodeoxyglucose (FDG) against the conventional dual-tracer method. Our method uses high-temporal resolution imaging (1 s/frame), an image-derived arterial input function, and distributed kinetic modelling to simultaneously estimate cerebral blood flow (CBF) and transvascular transport rate K1, from which BBB PS is calculated via the Renkin-Crone equation. Single-tracer and conventional dual-tracer PS estimates were compared in 18 volunteers scanned with the flow-tracer [11C]butanol and [18F]FDG PET. Single-tracer [18F]FDG PS estimates had 3.7% mean bias and 4.2% standard deviation of differences compared to dual-tracer estimates. This was enabled by strong agreement between [18F]FDG and [11C]butanol CBF estimates (Pearson R = 0.85, p < 0.001; mean difference: -0.057 ± 0.078 mL/min/cm3). These results demonstrate that high-temporal resolution dynamic PET enables single-tracer quantification of both CBF and BBB PS without a dedicated flow tracer, expanding opportunities for quantitative studies of molecular BBB transport across a broad range of tracers and disorders.

Cancer remains a critical global health challenge, with the World Health Organization (WHO) projecting 35 million new cases by 2050, necessitating advanced diagnostic tools such as whole-body FDG-PET imaging, which detects metabolic activity in pathologies. While 3D PET scans are powerful, their computational demands become excessive when dealing with deep learning, motivating the use of efficient 2D representations, moreover, this representation would assist radiologists in analyzing exams, since there are too few specialized professionals to interpret all the scans. This work proposes a Swin Transformer-based method to classify 2D Maximum Intensity Projection (MIP) images from FDG-PET scans into binary categories (cancerous vs. healthy), addressing challenges of multi-cancer detection (melanoma, lymphoma, lung cancer) and variability in image coverage. The approach achieved results of 82.08 ± 4.3 ± 2.9 ± 4.4 ± 7.1

Reducing scan times, radiation dose, and enhancing image quality, especially for lower-performance scanners, are critical in low-count/low-dose PET imaging. Deep learning (DL) techniques have been investigated for PET image denoising. However, existing models have often resulted in compromised image quality when achieving low-count/low-dose PET and have limited generalizability to different image noise levels, acquisition protocols, and patient populations. Recently, diffusion models have emerged as a state-of-the-art generative model to generate high-quality samples and have demonstrated strong potential for medical imaging tasks. However, for low-dose PET imaging, existing diffusion models fail to generate consistent 3D reconstructions (i.e., adjacent slices exhibit noticeable discontinuities or ”flickering” along the z-axis), struggle to generalize across varying noise levels, and often produce visually appealing but distorted details and biased tracer uptake. Here, we develop DDPET-3D, a dose-aware diffusion model for 3D low-dose PET imaging to address these challenges. In this work, ”3D” denotes 3D-consistent reconstruction achieved via a 2.5D conditioning backbone, rather than a fully 3D diffusion network. Collected from 4 medical centers globally with different scanners and clinical protocols, we extensively evaluated the proposed model using a total of 9,783 18F-FDG studies (1,596 patients) with low-dose/low-count levels ranging from 1% to 50%. With a cross-center, cross-scanner validation, the proposed DDPET-3D demonstrated its potential to generalize to different low-dose levels, different scanners, and different clinical protocols. As confirmed by reader studies conducted by board-certified nuclear medicine physicians, the readers rated the denoised images as comparable to—or better than—the full-dose images and prior DL baselines based on qualitative visual assessment. We also evaluated the lesion-level quantitative accuracy using a Monte Carlo simulation study and a lesion segmentation network. The presented results show the potential to achieve low-dose PET while maintaining image quality. Lastly, a group of real low-dose scans was also included for evaluation to demonstrate the clinical potential of DDPET-3D. Code and trained models are publicly available at https://github.com/HuidongXie/DDPET-3D

Purpose To develop and validate LION (Lesion Identification in Oncological Nuclear imaging), an open-source PET-only tumor segmentation pipeline for [ 18 F]FDG and PSMA-targeted PET/CT, and to investigate how training data characteristics influence segmentation performance. Materials and Methods In this retrospective multicenter study, 5,209 [ 18 F]FDG PET/CT scans spanning 19 disease types and 2,046 PSMA-targeted PET/CT scans were used to train PET-only segmentation models. Tumor segmentation incorporated organs with physiological uptake as auxiliary classes to enable PET-only inference. Tumor Occurrence Maps (TOMs) quantified tumor spatial diversity across the training data. For [ 18 F]FDG, disease-specific and mixed-disease models trained on progressively larger subsets were compared to test whether increasing spatial diversity improves generalization. Scanner-related domain shift was analyzed using DINOv2 embeddings. Models were evaluated on multicenter holdout cohorts (616 [ 18 F]FDG across 4 diseases; 443 PSMA-targeted prostate cancer scans) and compared with three open-source tools. Results Organ context improved median Dice from 0.62 to 0.71 for [ 18 F]FDG and from 0.75 to 0.83 for PSMA. Spatial diversity measured by TOMs was strongly associated with Dice (Spearman ρ = 0.87, P < 0.001). A mixed-disease model trained on 500 patients matched the performance of a lymphoma specialist model trained on 3,031 cases. DINOv2 embeddings revealed scanner-induced domain shift, between same-disease cohorts. LION achieved median Dice scores of 0.71 ([ 18 F]FDG) and 0.85 (PSMA) and outperformed other open-source approaches on common holdout patients. Conclusion LION enables PET-only automated segmentation for [ 18 F]FDG and PSMA-targeted PET. Training data composition, particularly spatial diversity quantified by TOMs, was strongly associated with segmentation performance.

Objective.This work presents and evaluates a Monte Carlo (MC) -based scatter correction (SC) method developed for the Jagiellonian positron emission tomography (J-PET) scanner, a modular PET system based on plastic scintillators.Approach.The algorithm employs SimSET-based simulations integrated into a time-of-flight ordered-subsets expectation maximization reconstruction framework to estimate scatter contributions. Phantom studies using the NEMA image quality (IQ) phantom and a proof-of-principle human subject scan with the J-PET scanner were analyzed. To accelerate computation, lutetium-yttrium oxyorthosilicate (LYSO) crystals were also assessed in simulations as a surrogate for the native plastic material BC-404.Main results.In phantom experiments, SC improved contrast recovery coefficients by over 20% and reduced background variability by 8.5%, without introducing significant noise. Residual activity in cold regions was also considerably reduced. Substituting LYSO in the simulations decreased runtime by nearly one order of magnitude, while maintaining deviations below 6% in IQ metrics compared to BC-404. Human subject data demonstrated qualitatively reduced residual scatter and improved organ delineation. Quantitative comparison with the commercial PET/CT scanner by General Electric HealthCare (GE) discovery MI Gen 2 showed consistent activity concentration ratios across organs, although higher noise and residual scatter between organs were observed in J-PET, which most likely originates from lower count density.Significance.The proposed MC-based SC method provides robust scatter removal for J-PET, improving quantitative performance and establishing a foundation for advanced correction and reconstruction techniques. These results bring the plastic scintillator-based J-PET scanner closer to enabling clinically relevant quantitative PET imaging.

Deep learning (DL)-based denoisers have shown promise for improving PET image quality, particularly in low-dose or short-duration acquisitions. However, their impact on lesion detectability - a key metric for oncologic diagnosis - remains insufficiently understood. This study investigates the limits of deep learning denoising on lesion contrast, quantification and detectability across varying noise levels and lesion to background ratios using a total-body PET dataset. PET data from a [18F]-FDG scan on the EXPLORER total-body scanner were randomly down-sampled to simulate reduced count levels. Synthetic 6-mm liver lesions with lesion-to-background activity concentration ratios (LBR) from 1.5:1 to 3:1 were inserted using the Monte Carlo-based tool DIANA. Images were reconstructed with TOF-OSEM and evaluated with and without an anatomically and metabolically informed diffusion model as the DL-based denoiser. The variance in each image voxel was quantified using the coefficient of variation across all image representations, the lesion contrast was calculated using the contrast to noise ratio and the lesion uptake was estimated using SUV-mean and -max, respectively. Lesion detectability was computed using a three-channel Channelized Hotelling Observer ( CHO) and summarized by the area under the ROC curve (AUC). The denoiser improved the overall image appearance and reliably reduced voxel variance independent of the input noise level. However, the denoiser's positive impact on the lesion contrast decreased at lower LBR and the lesion uptake was improved only for the highest LBR settings. In a comparison of denoised images with their noisy originals, CHO analysis showed slightly increased lesion detectability at lower noise levels for both noisy and denoised images, however, the differences do not rise to statistical significance. Detectability generally increased with higher lesion contrast. Denoised images outperformed noisy originals at the highest contrast. However, at the lowest end, denoised images approached chance performance (AUC = 0.58), while noisy images retained slightly higher detectability (AUC similar to 0.67). While DL- based denoisers enhance visual image quality and work well with high-contrast features, they can impair measured contrast, quantification, and detectability of low-contrast lesions. In addition, the probability for artificial intelligence-induced hallucinations and illusions under certain conditions may impact detection tasks and warrants further investigations. Rigorous validation is critical to ensure safe clinical translation without compromising diagnostic confidence.

We present a large whole-body and total-body curated dataset of dual-modality 2-deoxy-2-[18F]fluoro-D-glucose (FDG)-Positron Emission Tomography/Computed Tomography (PET/CT) studies, consisting of 1,683 PET/CT images and the corresponding CT-derived segmentations of 130 target regions. This multi-center dataset includes images from individuals without overt disease and patients with a range of malignant and inflammatory pathologies, including arthritis, lymphoma, and melanoma, as well as cancers of the lung, head-neck, and genito-urinary tract. Target regions were first automatically segmented from CT images using an in-house software and subsequently verified and corrected by physicians-in-training. In total, the segmented regions encompass 130 volumes, including abdominal organs, muscles, bones, cardiac subregions, vessels, adipose tissue, and skeletal muscle around the third lumbar vertebra. PET/CT images and corresponding CT-derived segmentations are provided in anonymized NIfTI format. The dataset can be used for deep learning training, validation, or multi-modality image analysis and thus fills an important gap in available resources to advance the use of PET/CT data in clinical management.

Objectives To test the hypothesis that recently-developed total body-positron emission tomography (TB-PET) imaging with integrated computed tomography (CT) will enable low-dose, quantitative, domain-specific evaluation of the total inflammatory burden of psoriatic arthritis (PsA) and associate with established outcome measures of the clinical domains of PsA.Methods Seventy-one adult participants (40 with PsA, 16 with rheumatoid arthritis (RA), and 15 with osteoarthritis (OA)) underwent 20-min TB-PET/CT scans using [18F]FDG, a glucose analogue radiotracer. [18F]FDG uptake was assessed qualitatively and quantitatively. Rheumatological examinations were performed prior to the scan. For both evaluations, domain-specific assessments included 68 joints, 6 entheses, 20 nails, axial disease and dactylitis.Results [18F]FDG PET uptake consistent with joint involvement and enthesitis was noted in 100% of participants with PsA. Other features included nail matrix pathology (53%), spinal involvement (60%), active sacroiliitis (13%) and dactylitis (10%). Patterns of [18F]FDG uptake in PsA differed from those in participants with RA or OA. There was a high concordance between TB-PET measures and the domain-specific assessments of the joint (75%), entheseal (79%) and nail (65%) pathology. TB-PET was positive for an additional 15% of joints, 20% of entheses and 13% of nails that were negative on clinical assessments.Conclusion TB-PET/CT identified inflammatory pathologies characteristic to all clinical domains of PsA and thus provided an in vivo evaluation of systemic PsA inflammatory burden. This promising tool may further contribute to identifying pathologies that may be occult, provide biomarkers to diagnose and differentiate PsA at an early stage, and to monitor early treatment response.

Our aim was to characterize the diagnostic accuracy indices for nodal (N)-staging with [18F]FDG Total-Body (TB) and short-axial field-of-view (SAFOV) PET/CT in non-small cell lung cancer (NSCLC) patients referred for staging or restaging. In this prospective single center cross-over head-to-head comparative study 48 patients underwent [18F]FDG TB and SAFOV PET/CT on the same day. In total 700 lymph node levels (1R/L, 2R/L, 3a/p, 4R/L, 5, 6, 7, 8R/L, 9R/L, 10-14R/L) of 28 patients could be correlated to a composite reference standard (histopathological correlation, imaging after localized or systemic treatment), which allowed determination of true positive (TP), false positive (FP), true negative (TN) and false negative (FN) lesions. Lymph nodes were characterized semi-quantitatively by maximum standardized uptake value (SUVmax), tumor-to-background ratio (TBR), metabolic tumor volume (MTV) and total lesion glycolysis (TLG) leading to threshold for each scanner. TB and SAFOV PET/CT showed high diagnostic accuracy indices for patient-based N-staging. Sensitivity and specificity were 86.0

Positron emission tomography (PET) image denoising, along with lesion and organ segmentation, are critical steps in PET-aided diagnosis. However, existing methods typically treat these tasks independently, overlooking inherent synergies between them as correlated steps in the analysis pipeline. In this work, we present the anatomically and metabolically informed diffusion (AMDiff) model, a unified framework for denoising and lesion/organ segmentation in low-count PET imaging. By integrating multi-task functionality and exploiting the mutual benefits of these tasks, AMDiff enables direct quantification of clinical metrics, such as total lesion glycolysis (TLG), from low-count inputs. The AMDiff model incorporates a semantic-informed denoiser based on diffusion strategy and a denoising-informed segmenter utilizing nnMamba architecture. The segmenter constrains denoised outputs via a lesion-organ-specific regularizer, while the denoiser enhances the segmenter by providing enriched image information through a denoising revision module. These components are connected via a warming-up mechanism to optimize multi-task interactions. Experiments on multi-vendor, multi-center, and multi-noise-level datasets demonstrate the superior performance of AMDiff. For test cases below 20% of the clinical count levels from participating sites, AMDiff achieves TLG quantification biases of -21.60±47.26%, outperforming its ablated versions which yield biases of -30.83±59.11% (without the lesion-organ-specific regularizer) and -35.63±54.08% (without the denoising revision module). By leveraging its internal multi-task synergies, AMDiff surpasses standalone PET denoising and segmentation methods. Compared to the benchmark denoising diffusion model, AMDiff reduces the normalized root-mean-square error for lesion/liver by 22.92/17.27% on average. Compared to the benchmark nnMamba segmentation model, AMDiff improves lesion/liver Dice coefficients by 10.17/2.02% on average.

Bone marrow (BM) metabolic quantification with 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) is of broad clinical significance for accurate assessment of BM at staging and follow-up, especially when immunotherapy is involved. However, current methods of quantifying BM may be inaccurate because the volume defined to measure bone marrow may also consist of a fraction of trabecular bone in which 18F-FDG activity is negligible, resulting in a potential underestimation of true BM uptake. In this study, we demonstrate this bone-led tissue composition effect and propose a bone fraction correction (BFC) method using X-ray dual-energy computed tomography (DECT) material decomposition. This study included ten scans from five cancer patients who underwent baseline and follow-up dynamic 18F-FDG PET and DECT scans using the uEXPLORER total-body PET/CT system. The voxel-wise bone volume fraction was estimated from DECT and then incorporated into the PET measurement formulas for BFC. The standardized uptake value (SUV), 18F-FDG delivery rate K1, and net influx rate Ki values in BM regions were estimated with and without BFC and compared using the statistical analysis. The results first demonstrated the feasibility of performing voxel-wise material decomposition using DECT for metabolic BM imaging. With BFC, the SUV, K1, and Ki values significantly increased by an average of 13.28% in BM regions compared to those without BFC (all P<0.0001), indicating the impact of BFC for BM quantification. Parametric imaging with BFC further confirmed regional analysis. Our study using DECT suggests current SUV and kinetic quantification of BM are likely underestimated in PET due to the presence of a significant bone volume fraction. Incorporating tissue composition information through BFC may improve BM metabolic quantification.

Artificial intelligence-generated content (AIGC) has shown remarkable performance in nuclear medicine imaging (NMI), offering cost-effective software solutions for tasks such as image enhancement, motion correction, and attenuation correction. However, these advancements come with the risk of hallucinations, generating realistic yet factually incorrect content. Hallucinations can misrepresent anatomical and functional information, compromising diagnostic accuracy and clinical trust. This paper presents a comprehensive perspective of hallucination-related challenges in AIGC for NMI, introducing the DREAM report, which covers recommendations for definition, representative examples, detection and evaluation metrics, underlying causes, and mitigation strategies. This position statement paper aims to initiate a common understanding for discussions and future research toward enhancing AIGC applications in NMI, thereby supporting their safe and effective deployment in clinical practice.

Neuroimaging of blood-brain barrier permeability has been instrumental in identifying its broad involvement in neurological and systemic diseases. However, current methods evaluate the blood-brain barrier mainly as a structural barrier. Here we developed a non-invasive positron emission tomography method in humans to measure the blood-brain barrier permeability of molecular radiotracers that cross the blood-brain barrier through its molecule-specific transport mechanism. Our method uses high-temporal resolution dynamic imaging and kinetic modeling for multiparametric imaging and quantification of the blood-brain barrier permeability-surface area product of molecular radiotracers. We show, in humans, our method can resolve blood-brain barrier permeability across three radiotracers and demonstrate its utility in studying brain aging and brain-body interactions in metabolic dysfunction-associated steatotic liver inflammation. Our method opens new directions to effectively study the molecular permeability of the human blood-brain barrier in vivo using the large catalogue of available molecular positron emission tomography tracers.

Background & aims: Brown adipose tissue (BAT) has been mainly investigated as a potential target against cardiometabolic disease, but it has also been linked to cancer-related outcomes. Although pre- clinical data support that BAT and the thermogenic adipocytes in white adipose tissue may play an adverse role in the pathogenesis of cancer cachexia, results from studies in patients have reported inconsistent results. The purpose of this study was to examine the interrelationship between presence of detectable BAT, changes in body weight, and cachexia in patients with cancer. We hypothesized that evidence of BAT at cancer diagnosis would be associated with greater weight loss and risk of cancer cachexia up to a year after cancer diagnosis. Methods: We conducted a retrospective cohort study in treatment-na & iuml;ve patients with detectable BAT (BAT+, n = 57) and without evidence of BAT (BAT-, n = 73) on 2-deoxy-2-[18F]fluoro-D-glucose positron emission tomography-computed tomography (18F-FDG-PET-CT) imaging performed for cancer staging (2004-2020). Patients' clinical, demographic, and anthropometric characteristics were extracted from their electronic medical record for up to a year after diagnosis. The two groups were a priori matched for demographic, anthropometric, and disease-related characteristics at diagnosis, as well as for season and outdoor temperature on the day of the PET-CT scan. Cancer cachexia was defined as weight loss greater than 5 % or 2 % if body mass index was lower than 20 kg/m2. Poisson regression models were fitted to estimate the relative risk (RR) for developing cancer cachexia over the 1-year follow-up among BAT+ compared to BAT- patients. Results: The BAT+ group experienced a lower magnitude of weight loss compared with the BAT- group during the 1-year follow-up (p = 0.014 for interaction between BAT status and time). The risk for cancer cachexia was 44 % lower in the BAT+ than the BAT- group, adjusted for age, sex, outdoor temperature on the day of the 18F-FDG-PET-CT imaging, cancer site and stage (RR: 0.56, 95 % CI: 0.32 to 0.97). Conclusion: Contrary to our original hypothesis, evidence of BAT assessed by 18F-FDG-PET-CT imaging at cancer diagnosis was associated with greater body weight maintenance and lower risk for developing cancer cachexia up to one year after diagnosis. Larger, prospective studies and mechanistic experiments are needed to expand and identify the causal factors of our observations. (c) 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

The standard Patlak plot, a simple yet efficient model, is widely used to describe irreversible tracer kinetics for dynamic PET imaging. Its widespread application to whole-body parametric imaging remains constrained because of the need for a full-time-course input function (e.g., 1 h). In this paper, we demonstrate the relative Patlak (RP) plot, which eliminates the need for the early-time input function, for total-body parametric imaging and its application to 20-min clinical scans acquired in list mode. Methods: We conducted a theoretic analysis to indicate that the RP intercept b' is equivalent to a ratio of the SUV relative to the plasma concentration, whereas the RP slope Ki' is equal to the standard Patlak Ki (net influx rate) multiplied by a global scaling factor for each subject. One challenge in applying RP to a short scan duration (e.g., 20 min) is the resulting high noise in the parametric images. We applied a self-supervised deep-kernel method for noise reduction. Using the standard Patlak plot as the reference, the RP method was evaluated for lesion quantification, lesion-to-background contrast, and myocardial visualization in total-body parametric imaging in 22 human subjects (12 healthy subjects and 10 cancer patients) who underwent a 1-h dynamic 18F-FDG scan. The RP method was also applied to the dynamic data reconstructed from a clinical standard 20-min list-mode scan either at 1 or 2 h after injection for 2 cancer patients. Results: We demonstrated that it is feasible to obtain high-quality parametric images from 20-min scans using RP parametric imaging with a self-supervised deep-kernel noise-reduction strategy. The RP slope Ki' was highly correlated with the standard Patlak Ki in lesions and major organs, demonstrating its quantitative potential across subjects. Compared with conventional SUVs, the Ki' images significantly improved lesion contrast and enabled visualization of the myocardium for potential cardiac assessment. The application of the RP parametric imaging to the 2 clinical scans also showed similar benefits. Conclusion: Using total-body PET with the RP approach, it is feasible to generate parametric images using data from a 20-min clinical list-mode scan.