Deep learning-aided respiratory motion compensation in PET/CT: addressing motion induced resolution loss, attenuation correction artifacts and PET-CT misalignment.

PURPOSE: Respiratory motion (RM) significantly impacts image quality in thoracoabdominal PET/CT imaging. This study introduces a unified data-driven respiratory motion correction (uRMC) method, utilizing deep learning neural networks, to solve all the major issues caused by RM, i.e., PET resolution loss, attenuation correction artifacts, and PET-CT misalignment. METHODS: In a retrospective study, 737 patients underwent [18F]FDG PET/CT scans using the uMI Panorama PET/CT scanner. Ninety-nine patients, who also had respiration monitoring device (VSM), formed the validation set. The remaining data of the 638 patients were used to train neural networks used in the uRMC. The uRMC primarily consists of three key components: (1) data-driven respiratory signal extraction, (2) attenuation map generation, and (3) PET-CT alignment. SUV metrics were calculated within 906 lesions for three approaches, i.e., data-driven uRMC (proposed), VSM-based uRMC, and OSEM without motion correction (NMC). RM magnitude of major organs were estimated. RESULTS: uRMC enhanced diagnostic capabilities by revealing previously undetected lesions, sharpening lesion contours, increasing SUV values, and improving PET-CT alignment. Compared to NMC, uRMC showed increases of 10% and 17% in SUVmax and SUVmean across 906 lesions. Sub-group analysis showed significant SUV increases in small and medium-sized lesions with uRMC. Minor differences were found between VSM-based and data-driven uRMC methods, with the SUVmax was found statistically marginal significant or insignificant between the two methods. The study observed varied motion amplitudes in major organs, typically ranging from 10 to 20 mm. CONCLUSION: A data-driven solution for respiratory motion in PET/CT has been developed, validated and evaluated. To the best of our knowledge, this is the first unified solution that compensates for the motion blur within PET, the attenuation mismatch artifacts caused by PET-CT misalignment, and the misalignment between PET and CT.

Validation and Evaluation of a Vendor-Provided Head Motion Correction Algorithm on the uMI Panorama PET/CT System.

Brain PET imaging often faces challenges from head motion (HM), which can introduce artifacts and reduce image resolution, crucial in clinical settings for accurate treatment planning, diagnosis, and monitoring. United Imaging Healthcare has developed NeuroFocus, an HM correction (HMC) algorithm for the uMI Panorama PET/CT system, using a data-driven, statistics-based approach. The HMC algorithm automatically detects HM using a centroid-of-distribution technique, requiring no parameter adjustments. This study aimed to validate NeuroFocus and assess the prevalence of HM in clinical short-duration 18F-FDG scans. Methods: The study involved 317 patients undergoing brain PET scans, divided into 2 groups: 15 for HMC validation and 302 for evaluation. Validation involved patients undergoing 2 consecutive 3-min single-bed-position brain 18F-FDG scans-one with instructions to remain still and another with instructions to move substantially. The evaluation examined 302 clinical single-bed-position brain scans for patients with various neurologic diagnoses. Motion was categorized as small or large on the basis of a 5% SUV change in the frontal lobe after HMC. Percentage differences in SUVmean were reported across 11 brain regions. Results: The validation group displayed a large negative difference (-10.1%), with variation of 5.2% between no-HM and HM scans. After HMC, this difference decreased dramatically (-0.8%), with less variation (3.2%), indicating effective HMC application. In the evaluation group, 38 of 302 patients experienced large HM, showing a 10.9% ± 8.9% SUV increase after HMC, whereas most exhibited minimal uptake changes (0.1% ± 1.3%). The HMC algorithm not only enhanced the image resolution and contrast but also aided in disease identification and reduced the need for repeat scans, potentially optimizing clinical workflows. Conclusion: The study confirmed the effectiveness of NeuroFocus in managing HM in short clinical 18F-FDG studies on the uMI Panorama PET/CT system. It found that approximately 12% of scans required HMC, establishing HMC as a reliable tool for clinical brain 18F-FDG studies.

The regional pattern of age-related synaptic loss in the human brain differs from gray matter volume loss: in vivo PET measurement with [11C]UCB-J.

PURPOSE: Aging is a major societal concern due to age-related functional losses. Synapses are crucial components of neural circuits, and synaptic density could be a sensitive biomarker to evaluate brain function. [11C]UCB-J is a positron emission tomography (PET) ligand targeting synaptic vesicle glycoprotein 2A (SV2A), which can be used to evaluate brain synaptic density in vivo. METHODS: We evaluated age-related changes in gray matter synaptic density, volume, and blood flow using [11C]UCB-J PET and magnetic resonance imaging (MRI) in a wide age range of 80 cognitive normal subjects (21-83 years old). Partial volume correction was applied to the PET data. RESULTS: Significant age-related decreases were found in 13, two, and nine brain regions for volume, synaptic density, and blood flow, respectively. The prefrontal cortex showed the largest volume decline (4.9% reduction per decade: RPD), while the synaptic density loss was largest in the caudate (3.6% RPD) and medial occipital cortex (3.4% RPD). The reductions in caudate are consistent with previous SV2A PET studies and likely reflect that caudate is the site of nerve terminals for multiple major tracts that undergo substantial age-related neurodegeneration. There was a non-significant negative relationship between volume and synaptic density reductions in 16 gray matter regions. CONCLUSION: MRI and [11]C-UCB-J PET showed age-related decreases of gray matter volume, synaptic density, and blood flow; however, the regional patterns of the reductions in volume and SV2A binding were different. Those patterns suggest that MR-based measures of GM volume may not be directly representative of synaptic density.

EEG microstates analysis after TMS in patients with subacute stroke during the resting state.

To investigate whether intermittent theta burst stimulation over the cerebellum induces changes in resting-state electroencephalography microstates in patients with subacute stroke and its correlation with cognitive and emotional function. Twenty-four stroke patients and 17 healthy controls were included in this study. Patients and healthy controls were assessed at baseline, including resting-state electroencephalography and neuropsychological scales. Fifteen patients received lateral cerebellar intermittent theta burst stimulation as well as routine rehabilitation training (intermittent theta burst stimulation-RRT group), whereas 9 patients received only conventional rehabilitation training (routine rehabilitation training group). After 2 wk, baseline data were recorded again in both groups. Stroke patients exhibited reduced parameters in microstate D and increased parameters in microstate C compared with healthy controls. However, after the administration of intermittent theta burst stimulation over the lateral cerebellum, significant alterations were observed in the majority of metrics for both microstates D and C. Lateral cerebellar intermittent theta burst stimulation combined with conventional rehabilitation has a stronger tendency to improve emotional and cognitive function in patients with subacute stroke than conventional rehabilitation. The improvement of mood and cognitive function was significantly associated with microstates C and D. We identified electroencephalography microstate spatiotemporal dynamics associated with clinical improvement following a course of intermittent theta burst stimulation therapy.

Nomograms for the prediction of decannulation in patients with neurological injury: a study based on clinical practice.

BACKGROUND: Rational prediction of the probability of decannulation in tracheotomy patients is of great importance to clinicians and patients' families. This study aimed to develop a prediction model for decannulation in tracheotomized patients with neurological injury using routine clinical data and blood tests. METHODS: We developed a prediction model based on 186 tracheotomized patients, and data were collected from January 2018 to March 2021. The least absolute shrinkage and selection operator (LASSO) regression model was used to optimize feature selection for the decannulation risk model. The performance of the prediction model was evaluated in terms of discrimination, calibration, and clinical utility using measures such as C-index, calibration plot, and decision curve analysis (DCA). Internal validation was performed through bootstrapping validation. RESULTS: A total of 66.13% (123/186) of patients were decannulated. Predictors included in the prediction nomogram were age, gender, subtype of neurological injury, Glasgow Coma Scale (GCS) score, swallowing function, duration of tracheotomy, procalcitonin (PCT) level, white blood cell (WBC) count, and serum albumin (ALB) level. The predictive model showed good discrimination, with a C-index of 0.755 (95% confidence interval: 0.68-0.83). Internal validation also confirmed a satisfactory C-index of 0.690. The DCA indicated that the nomogram added substantial value in predicting decannulation risk for patients with threshold probabilities falling between >21% and <98% compared to the existing scheme. CONCLUSIONS: This predictive model serves as a valuable instrument for clinicians to quantitatively assess the probability of decannulation in patients with neurological injury, aiding in informed decision-making and patient management.

MULTI-TASK DEEP LEARNING AND UNCERTAINTY ESTIMATION FOR PET HEAD MOTION CORRECTION.

Head motion occurring during brain positron emission tomography images acquisition leads to a decrease in image quality and induces quantification errors. We have previously introduced a Deep Learning Head Motion Correction (DL-HMC) method based on supervised learning of gold-standard Polaris Vicra motion tracking device and showed the potential of this method. In this study, we upgrade our network to a multi-task architecture in order to include image appearance prediction in the learning process. This multi-task Deep Learning Head Motion Correction (mtDL-HMC) model was trained on 21 subjects and showed enhanced motion prediction performance compared to our previous DL-HMC method on both quantitative and qualitative results for 5 testing subjects. We also evaluate the trustworthiness of network predictions by performing Monte Carlo Dropout at inference on testing subjects. We discard the data associated with a great motion prediction uncertainty and show that this does not harm the quality of reconstructed images, and can even improve it.

Markerless head motion tracking and event-by-event correction in brain PET.

Objective.Head motion correction (MC) is an essential process in brain positron emission tomography (PET) imaging. We have used the Polaris Vicra, an optical hardware-based motion tracking (HMT) device, for PET head MC. However, this requires attachment of a marker to the subject's head. Markerless HMT (MLMT) methods are more convenient for clinical translation than HMT with external markers. In this study, we validated the United Imaging Healthcare motion tracking (UMT) MLMT system using phantom and human point source studies, and tested its effectiveness on eight18F-FPEB and four11C-LSN3172176 human studies, with frame-based region of interest (ROI) analysis. We also proposed an evaluation metric, registration quality (RQ), and compared it to a data-driven evaluation method, motion-corrected centroid-of-distribution (MCCOD).Approach.UMT utilized a stereovision camera with infrared structured light to capture the subject's real-time 3D facial surface. Each point cloud, acquired at up to 30 Hz, was registered to the reference cloud using a rigid-body iterative closest point registration algorithm.Main results.In the phantom point source study, UMT exhibited superior reconstruction results than the Vicra with higher spatial resolution (0.35 ± 0.27 mm) and smaller residual displacements (0.12 ± 0.10 mm). In the human point source study, UMT achieved comparable performance as Vicra on spatial resolution with lower noise. Moreover, UMT achieved comparable ROI values as Vicra for all the human studies, with negligible mean standard uptake value differences, while no MC results showed significant negative bias. TheRQevaluation metric demonstrated the effectiveness of UMT and yielded comparable results to MCCOD.Significance.We performed an initial validation of a commercial MLMT system against the Vicra. Generally, UMT achieved comparable motion-tracking results in all studies and the effectiveness of UMT-based MC was demonstrated.

Framing protocol optimization in oncological Patlak parametric imaging with uKinetics.

PURPOSE: Total-body PET imaging with ultra-high sensitivity makes high-temporal-resolution framing protocols possible for the first time, which allows to capture rapid tracer dynamic changes. However, whether protocols with higher number of temporal frames can justify the efficacy with substantially added computation burden for clinical application remains unclear. We have developed a kinetic modeling software package (uKinetics) with the advantage of practical, fast, and automatic workflow for dynamic total-body studies. The aim of this work is to verify the uKinetics with PMOD and to perform framing protocol optimization for the oncological Patlak parametric imaging. METHODS: Six different protocols with 100, 61, 48, 29, 19 and 12 temporal frames were applied to analyze 60-min dynamic 18F-FDG PET scans of 10 patients, respectively. Voxel-based Patlak analysis coupled with automatically extracted image-derived input function was applied to generate parametric images. Normal tissues and lesions were segmented manually or automatically to perform correlation analysis and Bland-Altman plots. Different protocols were compared with the protocol of 100 frames as reference. RESULTS: Minor differences were found between uKinetics and PMOD in the Patlak parametric imaging. Compared with the protocol with 100 frames, the relative difference of the input function and quantitative kinetic parameters remained low for protocols with at least 29 frames, but increased for the protocols with 19 and 12 frames. Significant difference of lesion Ki values was found between the protocols with 100 frames and 12 frames. CONCLUSION: uKinetics was proved providing equivalent oncological Patlak parametric imaging comparing to PMOD. Minor differences were found between protocols with 100 and 29 frames, which indicated that 29-frame protocol is sufficient and efficient for the oncological 18F-FDG Patlak applications, and the protocols with more frames are not needed. The protocol with 19 frames yielded acceptable results, while that with 12 frames is not recommended.

Principal component analysis of synaptic density measured with [11C]UCB-J PET in early Alzheimer's disease.

BACKGROUND: Synaptic loss is considered an early pathological event and major structural correlate of cognitive impairment in Alzheimer's disease (AD). We used principal component analysis (PCA) to identify regional patterns of covariance in synaptic density using [11C]UCB-J PET and assessed the association between principal components (PC) subject scores with cognitive performance. METHODS: [11C]UCB-J binding was measured in 45 amyloid + participants with AD and 19 amyloid- cognitively normal participants aged 55-85. A validated neuropsychological battery assessed performance across five cognitive domains. PCA was applied to the pooled sample using distribution volume ratios (DVR) standardized (z-scored) by region from 42 bilateral regions of interest (ROI). RESULTS: Parallel analysis determined three significant PCs explaining 70.2% of the total variance. PC1 was characterized by positive loadings with similar contributions across the majority of ROIs. PC2 was characterized by positive and negative loadings with strongest contributions from subcortical and parietooccipital cortical regions, respectively, while PC3 was characterized by positive and negative loadings with strongest contributions from rostral and caudal cortical regions, respectively. Within the AD group, PC1 subject scores were positively correlated with performance across all cognitive domains (Pearson r = 0.24-0.40, P = 0.06-0.006), PC2 subject scores were inversely correlated with age (Pearson r = -0.45, P = 0.002) and PC3 subject scores were significantly correlated with CDR-sb (Pearson r = 0.46, P = 0.04). No significant correlations were observed between cognitive performance and PC subject scores in CN participants. CONCLUSIONS: This data-driven approach defined specific spatial patterns of synaptic density correlated with unique participant characteristics within the AD group. Our findings reinforce synaptic density as a robust biomarker of disease presence and severity in the early stages of AD.

Deep learning-based attenuation map generation with simultaneously reconstructed PET activity and attenuation and low-dose application.

Objective. In PET/CT imaging, CT is used for positron emission tomography (PET) attenuation correction (AC). CT artifacts or misalignment between PET and CT can cause AC artifacts and quantification errors in PET. Simultaneous reconstruction (MLAA) of PET activity (λ-MLAA) and attenuation (µ-MLAA) maps was proposed to solve those issues using the time-of-flight PET raw data only. However,λ-MLAA still suffers from quantification error as compared to reconstruction using the gold-standard CT-based attenuation map (µ-CT). Recently, a deep learning (DL)-based framework was proposed to improve MLAA by predictingµ-DL fromλ-MLAA andµ-MLAA using an image domain loss function (IM-loss). However, IM-loss does not directly measure the AC errors according to the PET attenuation physics. Our preliminary studies showed that an additional physics-based loss function can lead to more accurate PET AC. The main objective of this study is to optimize the attenuation map generation framework for clinical full-dose18F-FDG studies. We also investigate the effectiveness of the optimized network on predicting attenuation maps for synthetic low-dose oncological PET studies.Approach. We optimized the proposed DL framework by applying different preprocessing steps and hyperparameter optimization, including patch size, weights of the loss terms and number of angles in the projection-domain loss term. The optimization was performed based on 100 skull-to-toe18F-FDG PET/CT scans with minimal misalignment. The optimized framework was further evaluated on 85 clinical full-dose neck-to-thigh18F-FDG cancer datasets as well as synthetic low-dose studies with only 10% of the full-dose raw data.Main results. Clinical evaluation of tumor quantification as well as physics-based figure-of-merit metric evaluation validated the promising performance of our proposed method. For both full-dose and low-dose studies, the proposed framework achieved <1% error in tumor standardized uptake value measures.Significance. It is of great clinical interest to achieve CT-less PET reconstruction, especially for low-dose PET studies.

Fast Reconstruction for Deep Learning PET Head Motion Correction.

Head motion correction is an essential component of brain PET imaging, in which even motion of small magnitude can greatly degrade image quality and introduce artifacts. Building upon previous work, we propose a new head motion correction framework taking fast reconstructions as input. The main characteristics of the proposed method are: (i) the adoption of a high-resolution short-frame fast reconstruction workflow; (ii) the development of a novel encoder for PET data representation extraction; and (iii) the implementation of data augmentation techniques. Ablation studies are conducted to assess the individual contributions of each of these design choices. Furthermore, multi-subject studies are conducted on an 18F-FPEB dataset, and the method performance is qualitatively and quantitatively evaluated by MOLAR reconstruction study and corresponding brain Region of Interest (ROI) Standard Uptake Values (SUV) evaluation. Additionally, we also compared our method with a conventional intensity-based registration method. Our results demonstrate that the proposed method outperforms other methods on all subjects, and can accurately estimate motion for subjects out of the training set. All code is publicly available on GitHub: https://github.com/OnofreyLab/dl-hmc_fast_recon_miccai2023.

Cross-Attention for Improved Motion Correction in Brain PET.

Head movement during long scan sessions degrades the quality of reconstruction in positron emission tomography (PET) and introduces artifacts, which limits clinical diagnosis and treatment. Recent deep learning-based motion correction work utilized raw PET list-mode data and hardware motion tracking (HMT) to learn head motion in a supervised manner. However, motion prediction results were not robust to testing subjects outside the training data domain. In this paper, we integrate a cross-attention mechanism into the supervised deep learning network to improve motion correction across test subjects. Specifically, cross-attention learns the spatial correspondence between the reference images and moving images to explicitly focus the model on the most correlative inherent information - the head region the motion correction. We validate our approach on brain PET data from two different scanners: HRRT without time of flight (ToF) and mCT with ToF. Compared with traditional and deep learning benchmarks, our network improved the performance of motion correction by 58% and 26% in translation and rotation, respectively, in multi-subject testing in HRRT studies. In mCT studies, our approach improved performance by 66% and 64% for translation and rotation, respectively. Our results demonstrate that cross-attention has the potential to improve the quality of brain PET image reconstruction without the dependence on HMT. All code will be released on GitHub: https://github.com/OnofreyLab/dl_hmc_attention_mlcn2023.

An objective evaluation method for head motion estimation in PET-Motion corrected centroid-of-distribution.

Head motion presents a continuing problem in brain PET studies. A wealth of motion correction (MC) algorithms had been proposed in the past, including both hardware-based methods and data-driven methods. However, in most real brain PET studies, in the absence of ground truth or gold standard of motion information, it is challenging to objectively evaluate MC quality. For MC evaluation, image-domain metrics, e.g., standardized uptake value (SUV) change before and after MC are commonly used, but this measure lacks objectivity because 1) other factors, e.g., attenuation correction, scatter correction and parameters used in the reconstruction, will confound MC effectiveness; 2) SUV only reflects final image quality, and it cannot precisely inform when an MC method performed well or poorly during the scan time period; 3) SUV is tracer-dependent and head motion may cause increases or decreases in SUV for different tracers, so evaluating MC effectiveness is complicated. Here, we present a new algorithm, i.e., motion corrected centroid-of-distribution (MCCOD) to perform objective quality control for measured or estimated rigid motion information. MCCOD is a three-dimensional surrogate trace of the center of tracer distribution after performing rigid MC using the existing motion information. MCCOD is used to inform whether the motion information is accurate, using the PET raw data only, i.e., without PET image reconstruction, where inaccurate motion information typically leads to abrupt changes in the MCCOD trace. MCCOD was validated using simulation studies and was tested on real studies acquired from both time-of-flight (TOF) and non-TOF scanners. A deep learning-based brain mask segmentation was implemented, which is shown to be necessary for non-TOF MCCOD generation. MCCOD is shown to be effective in detecting abrupt translation motion errors in slowly varying tracer distribution caused by the motion tracking hardware and can be used to compare different motion estimation methods as well as to improve existing motion information.

Adaptive data-driven motion detection and optimized correction for brain PET.

Head motion during PET scans causes image quality degradation, decreased concentration in regions with high uptake and incorrect outcome measures from kinetic analysis of dynamic datasets. Previously, we proposed a data-driven method, center of tracer distribution (COD), to detect head motion without an external motion tracking device. There, motion was detected using one dimension of the COD trace with a semiautomatic detection algorithm, requiring multiple user defined parameters and manual intervention. In this study, we developed a new data-driven motion detection algorithm, which is automatic, self-adaptive to noise level, does not require user-defined parameters and uses all three dimensions of the COD trace (3DCOD). 3DCOD was first validated and tested using 30 simulation studies (18F-FDG, N = 15; 11C-raclopride (RAC), N = 15) with large motion. The proposed motion correction method was tested on 22 real human datasets, with 20 acquired from a high resolution research tomograph (HRRT) scanner (18F-FDG, N = 10; 11C-RAC, N = 10) and 2 acquired from the Siemens Biograph mCT scanner. Real-time hardware-based motion tracking information (Vicra) was available for all real studies and was used as the gold standard. 3DCOD was compared to Vicra, no motion correction (NMC), one-direction COD (our previous method called 1DCOD) and two conventional frame-based image registration (FIR) algorithms, i.e., FIR1 (based on predefined frames reconstructed with attenuation correction) and FIR2 (without attenuation correction) for both simulation and real studies. For the simulation studies, 3DCOD yielded -2.3 ± 1.4% (mean ± standard deviation across all subjects and 11 brain regions) error in region of interest (ROI) uptake for 18F-FDG (-3.4 ± 1.7% for 11C-RAC across all subjects and 2 regions) as compared to Vicra (perfect correction) while NMC, FIR1, FIR2 and 1DCOD yielded -25.4 ± 11.1% (-34.5 ± 16.1% for 11C- RAC), -13.4 ± 3.5% (-16.1 ± 4.6%), -5.7 ± 3.6% (-8.0 ± 4.5%) and -2.6 ± 1.5% (-5.1 ± 2.7%), respectively. For real HRRT studies, 3DCOD yielded -0.3 ± 2.8% difference for 18F-FDG (-0.4 ± 3.2% for 11C-RAC) as compared to Vicra while NMC, FIR1, FIR2 and 1DCOD yielded -14.9 ± 9.0% (-24.5 ± 14.6%), -3.6 ± 4.9% (-13.4 ± 14.3%), -0.6 ± 3.4% (-6.7 ± 5.3%) and -1.5 ± 4.2% (-2.2 ± 4.1%), respectively. In summary, the proposed motion correction method yielded comparable performance to the hardware-based motion tracking method for multiple tracers, including very challenging cases with large frequent head motion, in studies performed on a non-TOF scanner.

Deep learning-based attenuation correction for whole-body PET - a multi-tracer study with 18F-FDG, 68 Ga-DOTATATE, and 18F-Fluciclovine.

A novel deep learning (DL)-based attenuation correction (AC) framework was applied to clinical whole-body oncology studies using 18F-FDG, 68 Ga-DOTATATE, and 18F-Fluciclovine. The framework used activity (λ-MLAA) and attenuation (µ-MLAA) maps estimated by the maximum likelihood reconstruction of activity and attenuation (MLAA) algorithm as inputs to a modified U-net neural network with a novel imaging physics-based loss function to learn a CT-derived attenuation map (µ-CT). METHODS: Clinical whole-body PET/CT datasets of 18F-FDG (N = 113), 68 Ga-DOTATATE (N = 76), and 18F-Fluciclovine (N = 90) were used to train and test tracer-specific neural networks. For each tracer, forty subjects were used to train the neural network to predict attenuation maps (µ-DL). µ-DL and µ-MLAA were compared to the gold-standard µ-CT. PET images reconstructed using the OSEM algorithm with µ-DL (OSEMDL) and µ-MLAA (OSEMMLAA) were compared to the CT-based reconstruction (OSEMCT). Tumor regions of interest were segmented by two radiologists and tumor SUV and volume measures were reported, as well as evaluation using conventional image analysis metrics. RESULTS: µ-DL yielded high resolution and fine detail recovery of the attenuation map, which was superior in quality as compared to µ-MLAA in all metrics for all tracers. Using OSEMCT as the gold-standard, OSEMDL provided more accurate tumor quantification than OSEMMLAA for all three tracers, e.g., error in SUVmax for OSEMMLAA vs. OSEMDL: - 3.6 ± 4.4% vs. - 1.7 ± 4.5% for 18F-FDG (N = 152), - 4.3 ± 5.1% vs. 0.4 ± 2.8% for 68 Ga-DOTATATE (N = 70), and - 7.3 ± 2.9% vs. - 2.8 ± 2.3% for 18F-Fluciclovine (N = 44). OSEMDL also yielded more accurate tumor volume measures than OSEMMLAA, i.e., - 8.4 ± 14.5% (OSEMMLAA) vs. - 3.0 ± 15.0% for 18F-FDG, - 14.1 ± 19.7% vs. 1.8 ± 11.6% for 68 Ga-DOTATATE, and - 15.9 ± 9.1% vs. - 6.4 ± 6.4% for 18F-Fluciclovine. CONCLUSIONS: The proposed framework provides accurate and robust attenuation correction for whole-body 18F-FDG, 68 Ga-DOTATATE and 18F-Fluciclovine in tumor SUV measures as well as tumor volume estimation. The proposed method provides clinically equivalent quality as compared to CT in attenuation correction for the three tracers.

Supervised Deep Learning for Head Motion Correction in PET.

Head movement is a major limitation in brain positron emission tomography (PET) imaging, which results in image artifacts and quantification errors. Head motion correction plays a critical role in quantitative image analysis and diagnosis of nervous system diseases. However, to date, there is no approach that can track head motion continuously without using an external device. Here, we develop a deep learning-based algorithm to predict rigid motion for brain PET by lever-aging existing dynamic PET scans with gold-standard motion measurements from external Polaris Vicra tracking. We propose a novel Deep Learning for Head Motion Correction (DL-HMC) methodology that consists of three components: (i) PET input data encoder layers; (ii) regression layers to estimate the six rigid motion transformation parameters; and (iii) feature-wise transformation (FWT) layers to condition the network to tracer time-activity. The input of DL-HMC is sampled pairs of one-second 3D cloud representations of the PET data and the output is the prediction of six rigid transformation motion parameters. We trained this network in a supervised manner using the Vicra motion tracking information as gold-standard. We quantitatively evaluate DL-HMC by comparing to gold-standard Vicra measurements and qualitatively evaluate the reconstructed images as well as perform region of interest standard uptake value (SUV) measurements. An algorithm ablation study was performed to determine the contributions of each of our DL-HMC design choices to network performance. Our results demonstrate accurate motion prediction performance for brain PET using a data-driven registration approach without external motion tracking hardware. All code is publicly available on GitHub: https://github.com/OnofreyLab/dl-hmc_miccai2022.

Assessment of transient dopamine responses to smoked cannabis.

BACKGROUND: Dopaminergic mechanisms that may underlie cannabis' reinforcing effects are not well elucidated in humans. This positron emission tomography (PET) imaging study used the dopamine D2/3 receptor antagonist [11C]raclopride and kinetic modelling testing for transient changes in radiotracer uptake to assess the striatal dopamine response to smoked cannabis in a preliminary sample. METHODS: PET emission data were acquired from regular cannabis users (n = 14; 7 M/7 F; 19-32 years old) over 90 min immediately after [11C]raclopride administration (584 ± 95 MBq) as bolus followed by constant infusion (Kbol = 105 min). Participants smoked a cannabis cigarette, using a paced puff protocol, 35 min after scan start. Plasma concentrations of Δ9-THC and metabolites and ratings of subjective "high" were collected during imaging. Striatal dopamine responses were assessed voxelwise with a kinetic model testing for transient reductions in [11C]raclopride binding, linear-parametric neurotransmitter PET (lp-ntPET) (cerebellum as a reference region). RESULTS: Cannabis smoking increased plasma Δ9-THC levels (peak: 0-10 min) and subjective high (peak: 0-30 min). Significant clusters (>16 voxels) modeled by transient reductions in [11C]raclopride binding were identified for all 12 analyzed scans. In total, 26 clusters of significant responses to cannabis were detected, of which 16 were located in the ventral striatum, including at least one ventral striatum cluster in 11 of the 12 analyzed scans. CONCLUSIONS: These preliminary data support the sensitivity of [11C]raclopride PET with analysis of transient changes in radiotracer uptake to detect cannabis smoking-induced dopamine responses. This approach shows future promise to further elucidate roles of mesolimbic dopaminergic signaling in chronic cannabis use. ClinicalTrials.gov Identifier: NCT02817698.

Generation of parametric Ki images for FDG PET using two 5-min scans.

PURPOSE: The net uptake rate constant (Ki ) derived from dynamic imaging is considered the gold standard quantification index for FDG PET. In this study, we investigated the feasibility and assessed the clinical usefulness of generating Ki images for FDG PET using only two 5-min scans with population-based input function (PBIF). METHODS: Using a Siemens Biograph mCT, 10 subjects with solid lung nodules underwent a single-bed dynamic FDG PET scan and 13 subjects (five healthy and eight cancer patients) underwent a whole-body dynamic FDG PET scan in continuous-bed-motion mode. For each subject, a standard Ki image was generated using the complete 0-90 min dynamic data with Patlak analysis (t* = 20 min) and individual patient's input function, while a dual-time-point Ki image was generated from two 5-min scans based on the Patlak equations at early and late scans with the PBIF. Different start times for the early (ranging from 20 to 55 min with an increment of 5 min) and late (ranging from 50 to 85 min with an increment of 5 min) scans were investigated with the interval between scans being at least 30 min (36 protocols in total). The optimal dual-time-point protocols were then identified. Regions of interest (ROI) were drawn on nodules for the lung nodule subjects, and on tumors, cerebellum, and bone marrow for the whole-body-imaging subjects. Quantification accuracy was compared using the mean value of each ROI between standard Ki (gold standard) and dual-time-point Ki , as well as between standard Ki and relative standardized uptake value (SUV) change that is currently used in clinical practice. Correlation coefficients and least squares fits were calculated for each dual-time-point protocol and for each ROI. Then, the predefined criteria for identifying a reliable dual-time-point Ki estimation for each ROI were empirically determined as: (1) the squared correlation coefficient (R2 ) between standard Ki and dual-time-point Ki is larger than 0.9; (2) the absolute difference between the slope of the equality line (1.0) and that of the fitted line when plotting standard Ki versus dual-time-point Ki is smaller than 0.1; (3) the absolute value of the intercept of the fitted line when plotting standard Ki versus dual-time-point Ki normalized by the mean of the standard Ki across all subjects for each ROI is smaller than 10%. Using Williams' one-tailed t test, the correlation coefficient (R) between standard Ki and dual-time-point Ki was further compared with that between standard Ki and relative SUV change, for each dual-time-point protocol and for each ROI. RESULTS: Reliable dual-time-point Ki images were obtained for all the subjects using our proposed method. The percentage error introduced by the PBIF on the dual-time-point Ki estimation was smaller than 1% for all 36 protocols. Using the predefined criteria, reliable dual-time-point Ki estimation could be obtained in 25 of 36 protocols for nodules and in 34 of 36 protocols for tumors. A longer time interval between scans provided a more accurate Ki estimation in general. Using the protocol of 20-25 min plus 80-85 or 85-90 min, very high correlations were obtained between standard Ki and dual-time-point Ki (R2  = 0.994, 0.980, 0.971 and 0.925 for nodule, tumor, cerebellum, and bone marrow), with all the slope values with differences ≤0.033 from 1 and all the intercept values with differences ≤0.0006 mL/min/cm3 from 0. The corresponding correlations were much lower between standard Ki and relative SUV change (R2  = 0.673, 0.684, 0.065, 0.246). Dual-time-point Ki showed a significantly higher quantification accuracy with respect to standard Ki than relative SUV change for all the 36 protocols (p < 0.05 using Williams' one-tailed t test). CONCLUSIONS: Our proposed approach can obtain reliable Ki images and accurate Ki quantification from dual-time-point scans (5-min per scan), and provide significantly higher quantification accuracy than relative SUV change that is currently used in clinical practice.

Partial volume correction analysis for 11C-UCB-J PET studies of Alzheimer's disease.

PURPOSE: 11C-UCB-J PET imaging, targeting synaptic vesicle glycoprotein 2A (SV2A), has been shown to be a useful indicator of synaptic density in Alzheimer's disease (AD). For SV2A imaging, a decrease in apparent tracer uptake is often due to the combination of gray-matter (GM) atrophy and SV2A decrease in the remaining tissue. Our aim is to reveal the true SV2A change by performing partial volume correction (PVC). METHODS: We performed two PVC algorithms, Müller-Gärtner (MG) and 'iterative Yang' (IY), on 17 AD participants and 11 cognitive normal (CN) participants using the brain-dedicated HRRT scanner. Distribution volume VT, the rate constant K1, binding potential BPND (centrum semiovale as reference region), and tissue volume were compared. RESULTS: In most regions, both PVC algorithms reduced the between-group differences. Alternatively, in hippocampus, IY increased the significance of between-group differences while MG reduced it (VT, BPND and K1 group differences: uncorrected: 20%, 27%, 17%; MG: 18%, 22%, 14%; IY: 22%, 28%, 17%). The group difference in hippocampal volume (10%) was substantially smaller than any PET measures. MG increased GM binding values to a greater extent than IY due to differences in algorithm assumptions. CONCLUSION: 11C-UCB-J binding is significantly reduced in AD hippocampus, but PVC is important to adjust for significant volume reduction. After correction, PET measures are substantially more sensitive to group differences than volumetric MRI measures. Assumptions of each PVC algorithm are important and should be carefully examined and validated. For 11C-UCB-J, the less stringent assumptions of IY support its use as a PVC algorithm over MG.

Automatic Inter-Frame Patient Motion Correction for Dynamic Cardiac PET Using Deep Learning.

Patient motion during dynamic PET imaging can induce errors in myocardial blood flow (MBF) estimation. Motion correction for dynamic cardiac PET is challenging because the rapid tracer kinetics of 82Rb leads to substantial tracer distribution change across different dynamic frames over time, which can cause difficulties for image registration-based motion correction, particularly for early dynamic frames. In this paper, we developed an automatic deep learning-based motion correction (DeepMC) method for dynamic cardiac PET. In this study we focused on the detection and correction of inter-frame rigid translational motion caused by voluntary body movement and pattern change of respiratory motion. A bidirectional-3D LSTM network was developed to fully utilize both local and nonlocal temporal information in the 4D dynamic image data for motion detection. The network was trained and evaluated over motion-free patient scans with simulated motion so that the motion ground-truths are available, where one million samples based on 65 patient scans were used in training, and 600 samples based on 20 patient scans were used in evaluation. The proposed method was also evaluated using additional 10 patient datasets with real motion. We demonstrated that the proposed DeepMC obtained superior performance compared to conventional registration-based methods and other convolutional neural networks (CNN), in terms of motion estimation and MBF quantification accuracy. Once trained, DeepMC is much faster than the registration-based methods and can be easily integrated into the clinical workflow. In the future work, additional investigation is needed to evaluate this approach in a clinical context with realistic patient motion.