Optimization algorithms and weighting factors for analysis of dynamic PET studies

PET Mapping of Receptor Occupancy Using Joint Direct Parametric Reconstruction

Receptor occupancy (RO) studies using PET neuroimaging play a critical role in the development of drugs targeting the central nervous system (CNS). The conventional approach to estimate drug receptor occupancy consists in estimation of binding potential changes between two PET scans (baseline and post-drug injection). This estimation is typically performed separately for each scan by first reconstructing dynamic PET scan data before fitting a kinetic model to time activity curves. This approach fails to properly model the noise in PET measurements, resulting in poor RO estimates, especially in low receptor density regions.

OBJECTIVE

In this work, we evaluate a novel joint direct parametric reconstruction framework to directly estimate distributions of RO and other kinetic parameters in the brain from a pair of baseline and post-drug injection dynamic PET scans.

METHODS

The proposed method combines the use of regularization on RO maps with alternating optimization to enable estimation of occupancy even in low binding regions.

RESULTS

Simulation results demonstrate the quantitative improvement of this method over conventional approaches in terms of accuracy and precision of occupancy. The proposed method is also evaluated in preclinical in-vivo experiments using 11C-MK-6884 and a muscarinic acetylcholine receptor 4 positive allosteric modulator drug, showing improved estimation of receptor occupancy as compared to traditional estimators.

CONCLUSION

The proposed joint direct estimation framework improves RO estimation compared to conventional methods, especially in intermediate to low-binding regions.

SIGNIFICANCE

This work could potentially facilitate the evaluation of new drug candidates targeting the CNS.

Pharmacokinetic analysis and simplified uptake measures for tumour lesion [18F]F-AraG PET imaging in patients with non-small cell lung cancer

INTRODUCTION

The novel positron emission tomography (PET) imaging tracer, [18F]F-AraG, targets activated T-cells, offering a potential means to improve our understanding of immune-oncological processes. The aim of this study was to determine the optimal pharmacokinetic model to quantify tumour lesion [18F]F-AraG uptake in patients with non-small cell lung cancer (NSCLC), and to validate simplified measures at different time intervals against the pharmacokinetic uptake parameter.

METHODS

Ten patients with early-stage NSCLC and three patients with advanced NSCLC underwent a dynamic PET scan of minimal 60 min. Venous and/or arterial blood sampling was obtained at maximum seven time points. Tumour lesion time activity curves and metabolite-corrected input functions were analysed using single-tissue reversible (1T2k), two-tissue irreversible (2T3k) and two-tissue reversible (2T4k) plasma input models. Simplified uptake measures, such as standardised uptake value (SUV) and tumour-to-blood (TBR) or tumour-to-plasma ratio (TPR), were evaluated for different time intervals.

RESULTS

Whole-blood and plasma radioactivity concentrations showed rapid clearance of [18F]F-AraG. Metabolite analysis revealed a low rate of metabolism, at 70 min p.i., on average, 79% (SD = 9.8%) of the total radioactivity found in blood corresponded to intact [18F]F-AraG. The time activity curves were best fitted by the 2T3k model. Strong positive correlations were found for SUV (body weight (BW), lean body mass (LBM) or body surface area (BSA) corrected), TBR and TPR for any time interval between 20 and 70 min p.i. against the 2T3k-derived Ki. The correlation of TBR at 60-70 min p.i. with 2T3K-derived Ki (r (df = 20) = 0.87, p < 0.01), was stronger than for SUVBW (r (df = 20) = 0.80, p < 0.01).

CONCLUSION

Tumour lesion [18F]F-AraG uptake in patients with NSCLC is characterised by a 2T3k model. TBR and TPR show most potential for simplified quantification of tumour lesion [18F]F-AraG uptake in patients with NSCLC.

MCP-Net: Introducing Patlak Loss Optimization to Whole-body Dynamic PET Inter-frame Motion Correction

In whole-body dynamic positron emission tomography (PET), inter-frame subject motion causes spatial misalignment and affects parametric imaging. Many of the current deep learning inter-frame motion correction techniques focus solely on the anatomy-based registration problem, neglecting the tracer kinetics that contains functional information. To directly reduce the Patlak fitting error for 18F-FDG and further improve model performance, we propose an interframe motion correction framework with Patlak loss optimization integrated into the neural network (MCP-Net). The MCP-Net consists of a multiple-frame motion estimation block, an image-warping block, and an analytical Patlak block that estimates Patlak fitting using motion-corrected frames and the input function. A novel Patlak loss penalty component utilizing mean squared percentage fitting error is added to the loss function to reinforce the motion correction. The parametric images were generated using standard Patlak analysis following motion correction. Our framework enhanced the spatial alignment in both dynamic frames and parametric images and lowered normalized fitting error when compared to both conventional and deep learning benchmarks. MCP-Net also achieved the lowest motion prediction error and showed the best generalization capability. The potential of enhancing network performance and improving the quantitative accuracy of dynamic PET by directly utilizing tracer kinetics is suggested.

Inter-pass motion correction for whole-body dynamic PET and parametric imaging

Whole-body dynamic FDG-PET imaging through continuous-bed-motion (CBM) mode multi-pass acquisition protocol is a promising metabolism measurement. However, inter-pass misalignment originating from body movement could degrade parametric quantification. We aim to apply a non-rigid registration method for inter-pass motion correction in whole-body dynamic PET. 27 subjects underwent a 90-min whole-body FDG CBM PET scan on a Biograph mCT (Siemens Healthineers), acquiring 9 over-the-heart single-bed passes and subsequently 19 CBM passes (frames). The inter-pass motion correction was executed using non-rigid image registration with multi-resolution, B-spline free-form deformations. The parametric images were then generated by Patlak analysis. The overlaid Patlak slope Ki and y-intercept Vb images were visualized to qualitatively evaluate motion impact and correction effect. The normalized weighted mean squared Patlak fitting errors (NFE) were compared in the whole body, head, and hypermetabolic regions of interest (ROI). In Ki images, ROI statistics were collected and malignancy discrimination capacity was estimated by the area under the receiver operating characteristic curve (AUC). After the inter-pass motion correction was applied, the spatial misalignment appearance between Ki and Vb images was successfully reduced. Voxel-wise normalized fitting error maps showed global error reduction after motion correction. The NFE in the whole body (p = 0.0013), head (p = 0.0021), and ROIs (p = 0.0377) significantly decreased. The visual performance of each hypermetabolic ROI in Ki images was enhanced, while 3.59% and 3.67% average absolute percentage changes were observed in mean and maximum Ki values, respectively, across all evaluated ROIs. The estimated mean Ki values had substantial changes with motion correction (p = 0.0021). The AUC of both mean Ki and maximum Ki after motion correction increased, possibly suggesting the potential of enhancing oncological discrimination capacity through inter-pass motion correction.

Total-Body PET Multiparametric Imaging of Cancer Using a Voxelwise Strategy of Compartmental Modeling

Quantitative dynamic PET with compartmental modeling has the potential to enable multiparametric imaging and more accurate quantification than static PET imaging. Conventional methods for parametric imaging commonly use a single kinetic model for all image voxels and neglect the heterogeneity of physiologic models, which can work well for single-organ parametric imaging but may significantly compromise total-body parametric imaging on a scanner with a long axial field of view. In this paper, we evaluate the necessity of voxelwise compartmental modeling strategies, including time delay correction (TDC) and model selection, for total-body multiparametric imaging. Methods: Ten subjects (5 patients with metastatic cancer and 5 healthy volunteers) were scanned on a total-body PET/CT system after injection of 370 MBq of 18F-FDG. Dynamic data were acquired for 60 min. Total-body parametric imaging was performed using 2 approaches. One was the conventional method that uses a single irreversible 2-tissue-compartment model with and without TDC. The second approach selects the best kinetic model from 3 candidate models for individual voxels. The differences between the 2 approaches were evaluated for parametric imaging of microkinetic parameters and the 18F-FDG net influx rate, KiResults: TDC had a nonnegligible effect on kinetic quantification of various organs and lesions. The effect was larger in lesions with a higher blood volume. Parametric imaging of Ki with the standard 2-tissue-compartment model introduced vascular-region artifacts, which were overcome by the voxelwise model selection strategy. Conclusion: The time delay and appropriate kinetic model vary in different organs and lesions. Modeling of the time delay of the blood input function and model selection improved total-body multiparametric imaging.

Simultaneous multifactor Bayesian analysis (SiMBA) of PET time activity curve data

Positron emission tomography (PET) is an in vivo imaging method essential for studying the neurochemical pathophysiology of psychiatric and neurological disease. However, its high cost and exposure of participants to radiation make it unfeasible to employ large sample sizes. The major shortcoming of PET imaging is therefore its lack of power for studying clinically-relevant research questions. Here, we introduce a new method for performing PET quantification and analysis called SiMBA, which helps to alleviate these issues by improving the efficiency of PET analysis by exploiting similarities between both individuals and regions within individuals. In simulated [11C]WAY100635 data, SiMBA greatly improves both statistical power and the consistency of effect size estimation without affecting the false positive rate. This approach makes use of hierarchical, multifactor, multivariate Bayesian modelling to effectively borrow strength across the whole dataset to improve stability and robustness to measurement error. In so doing, parameter identifiability and estimation are improved, without sacrificing model interpretability. This comes at the cost of increased computational overhead, however this is practically negligible relative to the time taken to collect PET data. This method has the potential to make it possible to test clinically-relevant hypotheses which could never be studied before given the practical constraints. Furthermore, because this method does not require any additional information over and above that required for traditional analysis, it makes it possible to re-examine data which has already previously been collected at great expense. In the absence of dramatic advancements in PET image data quality, radiotracer development, or data sharing, PET imaging has been fundamentally limited in the scope of research hypotheses which could be studied. This method, especially combined with the recent steps taken by the PET imaging community to embrace data sharing, will make it possible to greatly improve the research possibilities and clinical relevance of PET neuroimaging.

Validation and test–retest repeatability performance of parametric methods for [11C]UCB-J PET

[¹¹C]UCB-J is a PET radioligand that binds to the presynaptic vesicle glycoprotein 2A. Therefore, [¹¹C]UCB-J PET may serve as an in vivo marker of synaptic integrity. The main objective of this study was to evaluate the quantitative accuracy and the 28-day test–retest repeatability (TRT) of various parametric quantitative methods for dynamic [¹¹C]UCB-J studies in Alzheimer’s disease (AD) patients and healthy controls (HC). Eight HCs and seven AD patients underwent two 60-min dynamic [¹¹C]UCB-J PET scans with arterial sampling over a 28-day interval. Several plasma-input based and reference-region based parametric methods were used to generate parametric images using metabolite corrected plasma activity as input function or white matter semi-ovale as reference region. Different parametric outcomes were compared regionally with corresponding non-linear regression (NLR) estimates. Furthermore, the 28-day TRT was assessed for all parametric methods. Spectral analysis (SA) and Logan graphical analysis showed high correlations with NLR estimates. Receptor parametric mapping (RPM) and simplified reference tissue model 2 (SRTM2) BP_ND, and reference Logan (RLogan) distribution volume ratio (DVR) regional estimates correlated well with plasma-input derived DVR and SRTM BP_ND. Among the multilinear reference tissue model (MRTM) methods, MRTM1 had the best correspondence with DVR and SRTM BP_ND. Among the parametric methods evaluated, spectral analysis (SA) and SRTM2 were the best plasma-input and reference tissue methods, respectively, to obtain quantitatively accurate and repeatable parametric images for dynamic [¹¹C]UCB-J PET.

Simulated Annealing-Based Image Reconstruction for Patients With COVID-19 as a Model for Ultralow-Dose Computed Tomography

The proposed algorithm of inverse problem of computed tomography (CT), using limited views, is based on stochastic techniques, namely simulated annealing (SA). The selection of an optimal cost function for SA-based image reconstruction is of prime importance. It can reduce annealing time, and also X-ray dose rate accompanying better image quality. In this paper, effectiveness of various cost functions, namely universal image quality index (UIQI), root-mean-squared error (RMSE), structural similarity index measure (SSIM), mean absolute error (MAE), relative squared error (RSE), relative absolute error (RAE), and root-mean-squared logarithmic error (RMSLE), has been critically analyzed and evaluated for ultralow-dose X-ray CT of patients with COVID-19. For sensitivity analysis of this ill-posed problem, the stochastically estimated images of lung phantom have been reconstructed. The cost function analysis in terms of computational and spatial complexity has been performed using image quality measures, namely peak signal-to-noise ratio (PSNR), Euclidean error (EuE), and weighted peak signal-to-noise ratio (WPSNR). It has been generalized for cost functions that RMSLE exhibits WPSNR of 64.33 ± 3.98 dB and 63.41 ± 2.88 dB for 8 × 8 and 16 × 16 lung phantoms, respectively, and it has been applied for actual CT-based image reconstruction of patients with COVID-19. We successfully reconstructed chest CT images of patients with COVID-19 using RMSLE with eighteen projections, a 10-fold reduction in radiation dose exposure. This approach will be suitable for accurate diagnosis of patients with COVID-19 having less immunity and sensitive to radiation dose.

Simulating the effect of cerebral blood flow changes on regional quantification of [18F]flutemetamol and [18F]florbetaben studies

Global and regional changes in cerebral blood flow (CBF) can result in biased quantitative estimates of amyloid load by PET imaging. Therefore, the current simulation study assessed effects of these changes on amyloid quantification using a reference tissue approach for [¹⁸F]flutemetamol and [¹⁸F]florbetaben. Previously validated pharmacokinetic rate constants were used to simulate time-activity curves (TACs) corresponding to full dynamic and dual-time-window acquisition protocols. CBF changes were simulated by varying the tracer delivery (K₁) from +25 to -25%. The standardized uptake value ratio (SUVr) was computed and TACs were fitted using reference Logan (RLogan) and the simplified reference tissue model (SRTM) to obtain the relative delivery rate (R₁) and volume of distribution ratio (DVR). RLogan was least affected by CBF changes (χ² = 583 p < 0.001, χ² = 81 p < 0.001, for [¹⁸F]flutemetamol and [¹⁸F]florbetaben, respectively) and the extent of CBF sensitivity generally increased for higher levels of amyloid. Further, SRTM-derived R₁ changes correlated well with simulated CBF changes (R² > 0.95) and SUVr's sensitivity to CBF changes improved for later uptake-times, with the exception of [¹⁸F]flutemetamol cortical changes. In conclusion, RLogan is the preferred method for amyloid quantification of [¹⁸F]flutemetamol and [¹⁸F]florbetaben studies and SRTM could be additionally used for obtaining a CBF proxy.

Human Positron Emission Tomography Neuroimaging

Neuroimaging with positron emission tomography (PET) is the most powerful tool for understanding pharmacology, neurochemistry, and pathology in the living human brain. This technology combines high-resolution scanners to measure radioactivity throughout the human body with specific, targeted radioactive molecules, which allow measurements of a myriad of biological processes in vivo. While PET brain imaging has been active for almost 40 years, the pace of development for neuroimaging tools, known as radiotracers, and for quantitative analytical techniques has increased dramatically over the past decade. Accordingly, the fundamental questions that can be addressed with PET have expanded in basic neurobiology, psychiatry, neurology, and related therapeutic development. In this review, we introduce the field of human PET neuroimaging, some of its conceptual underpinnings, and motivating questions. We highlight some of the more recent advances in radiotracer development, quantitative modeling, and applications of PET to the study of the human brain.

Probabilistic Graphical Models for Dynamic PET: A Novel Approach to Direct Parametric Map Estimation and Image Reconstruction

In the context of dynamic emission tomography, the conventional processing pipeline consists of independent image reconstruction of single-time frames, followed by the application of a suitable kinetic model to time-activity curves (TACs) at the voxel or region-of-interest level. Direct 4D positron emission tomography (PET) reconstruction, by contrast, seeks to move beyond this scheme and incorporate information from multiple time frames within the reconstruction task. Established direct methods are based on a deterministic description of voxelwise TACs, captured by the chosen kinetic model, considering the photon counting process the only source of uncertainty. In this paper, we introduce a new probabilistic modeling strategy based on the key assumption that activity time course would be subject to uncertainty even if the parameters of the underlying dynamic process are known. This leads to a hierarchical model that we formulate using the formalism of probabilistic graphical modeling. The inference is addressed using a new iterative algorithm, in which kinetic modeling results are treated as prior expectation of activity time course, rather than as a deterministic match, making it possible to control the trade-off between a data-driven and a model-driven reconstruction. The proposed method is flexible to an arbitrary choice of (linear and nonlinear) kinetic models, it enables the inclusion of arbitrary (sub)differentiable priors for parametric maps, and it is simple to implement. Computer simulations and an application to a real-patient scan show how the proposed method is able to generalize over conventional indirect and direct approaches, providing a bridge between them by properly tuning the impact of the kinetic modeling step on image reconstruction.

Direct List Mode Parametric Reconstruction for Dynamic Cardiac SPECT

Recently introduced stationary dedicated cardiac SPECT scanners provide new opportunities to quantify myocardial blood flow (MBF) using dynamic SPECT. However, comparing to PET, the low sensitivity of SPECT scanners affects MBF quantification due to the high noise level, especially for 201 Thallium (201Tl) due to its typically low injected dose. The conventional indirect method for generating parametric images typically starts by reconstructing a time series of frame images followed by fitting the time-activity curve (TAC) for each voxel or segment with an appropriate kinetic model. The indirect method is simple and easy to implement; however, it usually suffers from substantial image noise that could also lead to bias. In this paper, we developed a list mode direct parametric image reconstruction algorithm to substantially reduce noise in MBF quantification using dynamic SPECT and allow for patient radiation dose reduction. GPU-based parallel computing was used to achieve more than 2000-fold acceleration. The proposed method was evaluated in both simulation and in vivo canine studies. Compared with the indirect method, the proposed direct method achieved substantially lower image noise and variability, particularly at large number of iterations and at low-count levels.

Reproducibility of compartmental modelling of 18F-FDG PET/CT to evaluate lung inflammation

Introduction

Compartmental modelling is an established method of quantifying ¹⁸F-FDG uptake; however, only recently has it been applied to evaluate pulmonary inflammation. Implementation of compartmental models remains challenging in the lung, partly due to the low signal-to-noise ratio compared to other organs and the lack of standardisation. Good reproducibility is a key requirement of an imaging biomarker which has yet to be demonstrated in pulmonary compartmental models of ¹⁸F-FDG; in this paper, we address this unmet need.

Methods

Retrospective subject data were obtained from the EVOLVE observational study: Ten COPD patients (age =66±9; 8M/2F), 10 α₁ATD patients (age =63±8; 7M/3F) and 10 healthy volunteers (age =68±8; 9M/1F) never smokers. PET and CT images were co-registered, and whole lung regions were extracted from CT using an automated algorithm; the descending aorta was defined using a manually drawn region. Subsequent stages of the compartmental analysis were performed by two independent operators using (i) a MIAKAT^TM based pipeline and (ii) an in-house developed pipeline. We evaluated the metabolic rate constant of ¹⁸F-FDG (K_im) and the fractional blood volume (V_b); Bland-Altman plots were used to compare the results. Further, we adjusted the in-house pipeline to identify the salient features in the analysis which may help improve the standardisation of this technique in the lung.

Results

The initial agreement on a subject level was poor: Bland-Altman coefficients of reproducibility for K_im and V_b were 0.0031 and 0.047 respectively. However, the effect size between the groups (i.e. COPD, α₁ATD and healthy subjects) was similar using either pipeline. We identified the key drivers of this difference using an incremental approach: ROI methodology, modelling of the IDIF and time delay estimation. Adjustment of these factors led to improved Bland-Altman coefficients of reproducibility of 0.0015 and 0.027 for K_im and V_b respectively.

Conclusions

Despite similar methodology, differences in implementation can lead to disparate results in the outcome parameters. When reporting the outcomes of lung compartmental modelling, we recommend the inclusion of the details of ROI methodology, input function fitting and time delay estimation to improve reproducibility.

Direct Estimation of Voxel-Wise Neurotransmitter Response Maps From Dynamic PET Data

Computational methods, such as the linear parametric neurotransmitter PET (lp-ntPET) method, have been developed to characterize the transient changes in radiotracer kinetics in the target tissue during endogenous neurotransmitter release. In this paper, we describe and evaluate a parametric reconstruction algorithm that uses an expectation maximization framework, along with the lp-ntPET model, to estimate the endogenous neurotransmitter response to stimuli directly from the measured PET data. Computer simulations showed that the proposed direct reconstruction method offers improved accuracy and precision for the estimated timing parameters of the neurotransmitter response at the voxel level ( t_d=1±2 min, for activation onset bias and standard deviation) compared with conventional post reconstruction modeling ( t_d=4±7 min). In addition, we applied the proposed direct parameter estimation methodology to a [¹¹C]raclopride displacement study of an awake rat and generated parametric maps illustrating the magnitude of ligand displacement from striatum. Although the estimated parametric maps of activation magnitude obtained from both direct and post reconstruction methodologies suffered from false positive activations, the proposed direct reconstruction framework offered more reliable parametric maps when the activation onset parameter was constrained.

Optimized dual-time-window protocols for quantitative [18F]flutemetamol and [18F]florbetaben PET studies

Background

A long dynamic scanning protocol may be required to accurately measure longitudinal changes in amyloid load. However, such a protocol results in a lower patient comfort and scanning efficiency compared to static scans. A compromise can be achieved by implementing dual-time-window protocols. This study aimed to optimize these protocols for quantitative [¹⁸F]flutemetamol and [¹⁸F]florbetaben studies.

Methods

Rate constants for subjects across the Alzheimer’s disease spectrum (i.e., non-displaceable binding potential (BP_ND) in the range 0.02–0.77 and 0.02–1.04 for [¹⁸F]flutemetamol and [¹⁸F]florbetaben, respectively) were established based on clinical [¹⁸F]flutemetamol (N = 6) and [¹⁸F]florbetaben (N = 20) data, and used to simulate tissue time-activity curves (TACs) of 110 min using a reference tissue and plasma input model. Next, noise was added (N = 50) and data points corresponding to different intervals were removed from the TACs, ranging from 0 (i.e., 90–90 = full-kinetic curve) to 80 (i.e., 10–90) minutes, creating a dual-time-window. Resulting TACs were fitted using the simplified reference tissue method (SRTM) to estimate the BP_ND, outliers (≥ 1.5 × BP_ND max) were removed and the bias was assessed using the distribution volume ratio (DVR = BP_ND + 1). To this end, acceptability curves, which display the fraction of data below a certain bias threshold, were generated and the area under those curves were calculated.

Results

[¹⁸F]Flutemetamol and [¹⁸F]florbetaben data demonstrated an increased bias in amyloid estimate for larger intervals and higher noise levels. An acceptable bias (≤ 3.1%) in DVR could be obtained with all except the 10–90 and 20–90-min intervals. Furthermore, a reduced fraction of acceptable data and most outliers were present for these two largest intervals (maximum percentage outliers 48 and 32 for [¹⁸F]flutemetamol and [¹⁸F]florbetaben, respectively).

Conclusions

The length of the interval inversely correlates with the accuracy of the BP_ND estimates. Consequently, a dual-time-window protocol of 0–30 and 90–110 min (=maximum of 60 min interval) allows for accurate estimation of BP_ND values for both tracers.

[¹⁸F]flutemetamol: EudraCT 2007-000784-19, registered 8 February 2007, [¹⁸F]florbetaben: EudraCT 2006-003882-15, registered 2006.

Electronic supplementary material

The online version of this article (10.1186/s13550-019-0499-4) contains supplementary material, which is available to authorized users.

Assessment of Simplified Methods for Quantification of 18F-FDHT Uptake in Patients with Metastatic Castration-Resistant Prostate Cancer

¹⁸F-fluorodihydrotestosterone (¹⁸F-FDHT) PET/CT potentially provides a noninvasive method for assessment of androgen receptor expression in patients with metastatic castration-resistant prostate cancer (mCRPC). The objective of this study was to assess simplified methods for quantifying ¹⁸F-FDHT uptake in mCRPC patients and to assess effects of tumor perfusion on these ¹⁸F-FDHT uptake metrics. Methods: Seventeen mCRPC patients were included in this prospective observational multicenter study. Test and retest 30-min dynamic ¹⁸F-FDHT PET/CT scans with venous blood sampling were performed in 14 patients. In addition, arterial blood sampling and dynamic ¹⁵O-H₂O scans were obtained in a subset of 6 patients. Several simplified methods were assessed: Patlak plots; SUV normalized to body weight (SUV_BW), lean body mass (SUV_LBM), whole blood (SUV_WB), parent plasma activity concentration (SUV_PP), area under the parent plasma curve (SUV_AUC,PP), and area under the whole-blood input curve (SUV_AUC,WB); and SUV_BW corrected for sex hormone-binding globulin levels (SUV_SHBG). Results were correlated with parameters derived from full pharmacokinetic ¹⁸F-FDHT and ¹⁵O-H₂O. Finally, the repeatability of individual quantitative uptake metrics was assessed. Results: Eighty-seven ¹⁸F-FDHT-avid lesions were evaluated. ¹⁸F-FDHT uptake was best described by an irreversible 2-tissue-compartment model. Replacing the continuous metabolite-corrected arterial plasma input function with an image-derived input function in combination with venous sample data provided similar K _i results (R ² = 0.98). Patlak K _i and SUV_AUC,PP showed an excellent correlation (R ² > 0.9). SUV_BW showed a moderate correlation to K _i (R ² = 0.70, presumably due to fast ¹⁸F-FDHT metabolism. When calculating SUV_SHBG, correlation to K _i improved (R ² = 0.88). The repeatability of full kinetic modeling parameters was inferior to that of simplified methods (repeatability coefficients > 36% vs. < 28%, respectively). ¹⁸F-FDHT uptake showed minimal blood flow dependency. Conclusion: ¹⁸F-FDHT kinetics in mCRPC patients are best described by an irreversible 2-tissue-compartment model with blood volume parameter. SUV_AUC,PP showed a near-perfect correlation with the irreversible 2-tissue-compartment model analysis and can be used for accurate quantification of ¹⁸F-FDHT uptake in whole-body PET/CT scans. In addition, SUV_SHBG could potentially be used as an even simpler method to quantify ¹⁸F-FDHT uptake when less complex scanning protocols and accuracy are required.

Semi-quantitative cerebral blood flow parameters derived from non-invasive [15O]H2O PET studies

Quantification of regional cerebral blood flow (CBF) using [¹⁵O]H₂O positron emission tomography (PET) requires the use of an arterial input function. Arterial sampling, however, is not always possible, for example in ill-conditioned or paediatric patients. Therefore, it is of interest to explore the use of non-invasive methods for the quantification of CBF. For validation of non-invasive methods, test-retest normal and hypercapnia data from 15 healthy volunteers were used. For each subject, the data consisted of up to five dynamic [¹⁵O]H₂O brain PET studies of 10 min and including arterial sampling. A measure of CBF was estimated using several non-invasive methods earlier reported in literature. In addition, various parameters were derived from the time-activity curve (TAC). Performance of these methods was assessed by comparison with full kinetic analysis using correlation and agreement analysis. The analysis was repeated with normalization to the whole brain grey matter value, providing relative CBF distributions. A reliable, absolute quantitative estimate of CBF could not be obtained with the reported non-invasive methods. Relative (normalized) CBF was best estimated using the double integration method.

Quantification of O-(2-[18F]fluoroethyl)-L-tyrosine kinetics in glioma

BACKGROUND

This study identified the optimal tracer kinetic model for quantification of dynamic O-(2-[18F]fluoroethyl)-L-tyrosine ([18F]FET) positron emission tomography (PET) studies in seven patients with diffuse glioma (four glioblastoma, three lower grade glioma). The performance of more simplified approaches was evaluated by comparison with the optimal compartment model. Additionally, the relationship with cerebral blood flow-determined by [15O]H2O PET-was investigated.

RESULTS

The optimal tracer kinetic model was the reversible two-tissue compartment model. Agreement analysis of binding potential estimates derived from reference tissue input models with the distribution volume ratio (DVR)-1 derived from the plasma input model showed no significant average difference and limits of agreement of - 0.39 and 0.37. Given the range of DVR-1 (- 0.25 to 1.5), these limits are wide. For the simplified methods, the 60-90 min tumour-to-blood ratio to parent plasma concentration yielded the highest correlation with volume of distribution VT as calculated by the plasma input model (r = 0.97). The 60-90 min standardized uptake value (SUV) showed better correlation with VT (r = 0.77) than SUV based on earlier intervals. The 60-90 min SUV ratio to contralateral healthy brain tissue showed moderate agreement with DVR with no significant average difference and limits of agreement of - 0.24 and 0.30. A significant but low correlation was found between VT and CBF in the tumour regions (r = 0.61, p = 0.007).

CONCLUSION

Uptake of [18F]FET was best modelled by a reversible two-tissue compartment model. Reference tissue input models yielded estimates of binding potential which did not correspond well with plasma input-derived DVR-1. In comparison, SUV ratio to contralateral healthy brain tissue showed slightly better performance, if measured at the 60-90 min interval. SUV showed only moderate correlation with VT. VT shows correlation with CBF in tumour.

An improved optimization algorithm of the three-compartment model with spillover and partial volume corrections for dynamic FDG PET images of small animal hearts in vivo

The three-compartment model with spillover (SP) and partial volume (PV) corrections has been widely used for noninvasive kinetic parameter studies of dynamic 2-[18F] fluoro-2deoxy-D-glucose (FDG) positron emission tomography images of small animal hearts in vivo. However, the approach still suffers from estimation uncertainty or slow convergence caused by the commonly used optimization algorithms. The aim of this study was to develop an improved optimization algorithm with better estimation performance. Femoral artery blood samples, image-derived input functions from heart ventricles and myocardial time-activity curves (TACs) were derived from data on 16 C57BL/6 mice obtained from the UCLA Mouse Quantitation Program. Parametric equations of the average myocardium and the blood pool TACs with SP and PV corrections in a three-compartment tracer kinetic model were formulated. A hybrid method integrating artificial immune-system and interior-reflective Newton methods were developed to solve the equations. Two penalty functions and one late time-point tail vein blood sample were used to constrain the objective function. The estimation accuracy of the method was validated by comparing results with experimental values using the errors in the areas under curves (AUCs) of the model corrected input function (MCIF) and the 18F-FDG influx constant K _i . Moreover, the elapsed time was used to measure the convergence speed. The overall AUC error of MCIF for the 16 mice averaged  -1.4  ±  8.2%, with correlation coefficients of 0.9706. Similar results can be seen in the overall K _i error percentage, which was 0.4  ±  5.8% with a correlation coefficient of 0.9912. The t-test P value for both showed no significant difference. The mean and standard deviation of the MCIF AUC and K _i percentage errors have lower values compared to the previously published methods. The computation time of the hybrid method is also several times lower than using just a stochastic algorithm. The proposed method significantly improved the model estimation performance in terms of the accuracy of the MCIF and K _i , as well as the convergence speed.

Quantitative PET Imaging in Drug Development: Estimation of Target Occupancy

Positron emission tomography, an imaging tool using radiolabeled tracers in humans and preclinical species, has been widely used in recent years in drug development, particularly in the central nervous system. One important goal of PET in drug development is assessing the occupancy of various molecular targets (e.g., receptors, transporters, enzymes) by exogenous drugs. The current linear mathematical approaches used to determine occupancy using PET imaging experiments are presented. These algorithms use results from multiple regions with different target content in two scans, a baseline (pre-drug) scan and a post-drug scan. New mathematical estimation approaches to determine target occupancy, using maximum likelihood, are presented. A major challenge in these methods is the proper definition of the covariance matrix of the regional binding measures, accounting for different variance of the individual regional measures and their nonzero covariance, factors that have been ignored by conventional methods. The novel methods are compared to standard methods using simulation and real human occupancy data. The simulation data showed the expected reduction in variance and bias using the proper maximum likelihood methods, when the assumptions of the estimation method matched those in simulation. Between-method differences for data from human occupancy studies were less obvious, in part due to small dataset sizes. These maximum likelihood methods form the basis for development of improved PET covariance models, in order to minimize bias and variance in PET occupancy studies.

Model selection criteria for dynamic brain PET studies

Background

 Several criteria exist to identify the optimal model for quantification of tracer kinetics. The purpose of this study was to evaluate the correspondence in kinetic model preference identification for brain PET studies among five model selection criteria: Akaike Information Criterion (AIC), AIC unbiased (AIC_C), model selection criterion (MSC), Schwartz Criterion (SC), and F-test.

Materials and Methods

Six tracers were evaluated: [¹¹C]FMZ, [¹¹C]GMOM, [¹¹C]PK11195, [¹¹C]Raclopride, [¹⁸F]FDG, and [¹¹C]PHT, including data from five subjects per tracer. Time activity curves (TACs) were analysed using six plasma input models: reversible single-tissue model (1T2k), irreversible two-tissue model (2T3k), and reversible two-tissue model (2T4k), all with and without blood volume fraction parameter (V_B). For each tracer and criterion, the percentage of TACs preferring a certain model was calculated.

Results

For all radiotracers, strong agreement was seen across the model selection criteria. The F-test was considered as the reference, as it is a frequently used hypothesis test. The F-test confirmed the AIC preferred model in 87% of all cases. The strongest (but minimal) disagreement across regional TACs was found when comparing AIC with AIC_C. Despite these regional discrepancies, same preferred kinetic model was obtained using all criteria, with an exception of one FMZ subject.

Conclusion

In conclusion, all five model selection criteria resulted in similar conclusions with only minor differences that did not affect overall model selection.

Electronic supplementary material

The online version of this article (10.1186/s40658-017-0197-0) contains supplementary material, which is available to authorized users.

Kinetic Modelling of Infection Tracers [18F]FDG, [68Ga]Ga-Citrate, [11C]Methionine, and [11C]Donepezil in a Porcine Osteomyelitis Model

Introduction

Positron emission tomography (PET) is increasingly applied for infection imaging using [¹⁸F]FDG as tracer, but uptake is unspecific. The present study compares the kinetics of [¹⁸F]FDG and three other PET tracers with relevance for infection imaging.

Methods

A juvenile porcine osteomyelitis model was used. Eleven pigs underwent PET/CT with 60-minute dynamic PET imaging of [¹⁸F]FDG, [⁶⁸Ga]Ga-citrate, [¹¹C]methionine, and/or [¹¹C]donepezil, along with blood sampling. For infectious lesions, kinetic modelling with one- and two-tissue-compartment models was conducted for each tracer.

Results

Irreversible uptake was found for [¹⁸F]FDG and [⁶⁸Ga]Ga-citrate; reversible uptake was found for [¹¹C]methionine (two-tissue model) and [¹¹C]donepezil (one-tissue model). The uptake rate for [⁶⁸Ga]Ga-citrate was slow and diffusion-limited. For the other tracers, the uptake rate was primarily determined by perfusion (flow-limited uptake). Net uptake rate for [¹⁸F]FDG and distribution volume for [¹¹C]methionine were significantly higher for infectious lesions than for correspondingly noninfected tissue. For [¹¹C]donepezil in pigs, labelled metabolite products appeared to be important for the analysis.

Conclusions

The kinetics of the four studied tracers in infection was characterized. For clinical applications, [¹⁸F]FDG remains the first-choice PET tracer. [¹¹C]methionine may have a potential for detecting soft tissue infections. [⁶⁸Ga]Ga-citrate and [¹¹C]donepezil were not found useful for imaging of osteomyelitis.

Test-retest reproducibility of quantitative binding measures of [11C]Ro15-4513, a PET ligand for GABAA receptors containing alpha5 subunits

INTRODUCTION

Alteration of γ-aminobutyric acid "A" (GABAA) receptor-mediated neurotransmission has been associated with various neurological and psychiatric disorders. [11C]Ro15-4513 is a PET ligand with high affinity for α5-subunit-containing GABAA receptors, which are highly expressed in limbic regions of the human brain (Sur et al., 1998). We quantified the test-retest reproducibility of measures of [11C]Ro15-4513 binding derived from six different quantification methods (12 variants).

METHODS

Five healthy males (median age 40 years, range 38-49 years) had a 90-min PET scan on two occasions (median interval 12 days, range 11-30 days), after injection of a median dose of 441 MegaBequerels of [11C]Ro15-4513. Metabolite-corrected arterial plasma input functions (parent plasma input functions, ppIFs) were generated for all scans. We quantified regional binding using six methods (12 variants), some of which were region-based (applied to the average time-activity curve within a region) and others were voxel-based: 1) Models requiring arterial ppIFs - regional reversible compartmental models with one and two tissue compartments (2kbv and 4kbv); 2) Regional and voxelwise Logan's graphical analyses (Logan et al., 1990), which required arterial ppIFs; 3) Model-free regional and voxelwise (exponential) spectral analyses (SA; (Cunningham and Jones, 1993)), which also required arterial ppIFs; 4) methods not requiring arterial ppIFs - voxelwise standardised uptake values (Kenney et al., 1941), and regional and voxelwise simplified reference tissue models (SRTM/SRTM2) using brainstem or alternatively cerebellum as pseudo-reference regions (Lammertsma and Hume, 1996; Gunn et al., 1997). To compare the variants, we sampled the mean values of the outcome parameters within six bilateral, non-reference grey matter regions-of-interest. Reliability was quantified in terms of median absolute percentage test-retest differences (MA-TDs; preferentially low) and between-subject coefficient of variation (BS-CV, preferentially high), both compounded by the intraclass correlation coefficient (ICC). These measures were compared between variants, with particular interest in the hippocampus.

RESULTS

Two of the six methods (5/12 variants) yielded reproducible data (i.e. MA-TD <10%): regional SRTMs and voxelwise SRTM2s, both using either the brainstem or the cerebellum; and voxelwise SA. However, the SRTMs using the brainstem yielded a lower median BS-CV (7% for regional, 7% voxelwise) than the other variants (8-11%), resulting in lower ICCs. The median ICCs across six regions were 0.89 (interquartile range 0.75-0.90) for voxelwise SA, 0.71 (0.64-0.84) for regional SRTM-cerebellum and 0.83 (0.70-0.86) for voxelwise SRTM-cerebellum. The ICCs for the hippocampus were 0.89 for voxelwise SA, 0.95 for regional SRTM-cerebellum and 0.93 for voxelwise SRTM-cerebellum.

CONCLUSION

Quantification of [11C]Ro15-4513 binding shows very good to excellent reproducibility with SRTM and with voxelwise SA which, however, requires an arterial ppIF. Quantification in the α5 subunit-rich hippocampus is particularly reliable. The very low expression of the α5 in the cerebellum (Fritschy and Mohler, 1995; Veronese et al., 2016) and the substantial α1 subunit density in this region may hamper the application of reference tissue methods.

Pharmacokinetic modeling of a novel hypoxia PET tracer [18F]HX4 in patients with non-small cell lung cancer

Background

[¹⁸F]HX4 is a promising new PET tracer developed to identify hypoxic areas in tumor tissue. This study analyzes [¹⁸F]HX4 kinetics and assesses the performance of simplified methods for quantification of [¹⁸F]HX4 uptake.

To this end, eight patients with non-small cell lung cancer received dynamic PET scans at three different time points (0, 120, and 240 min) after injection of 426 ± 72 MBq [¹⁸F]HX4, each lasting 30 min. Several compartment models were fitted to time activity curves (TAC) derived from various areas within tumor tissue using image-derived input functions.

Results

Best fits were obtained using the reversible two-tissue compartment model with blood volume parameter (2T4k+V_B). Simplified measures correlated well with V_T estimates (tumor-to-blood ratio (TBr) R² = 0.96, tumor-to-muscle ratio R² = 0.94, standardized uptake value R² = 0.89).

Conclusions

[¹⁸F]HX4 shows reversible kinetics in tumor tissue: 2T4k+V_B. TBr based on static imaging at 2 or 4 h can be used for quantification of [¹⁸F]HX4 uptake.

Electronic supplementary material

The online version of this article (doi:10.1186/s40658-016-0167-y) contains supplementary material, which is available to authorized users.

Noninvasive bi-graphical analysis for the quantification of slowly reversible radioligand binding

In this paper, we presented a novel reference-region-based (noninvasive) bi-graphical analysis for the quantification of a reversible radiotracer binding that may be too slow to reach relative equilibrium (RE) state during positron emission tomography (PET) scans. The proposed method indirectly implements the noninvasive Logan plot, through arithmetic combination of the parameters of two other noninvasive methods and the apparent tissue-to-plasma efflux rate constant for the reference region ([Formula: see text]). We investigated its validity and statistical properties, by performing a simulation study with various noise levels and [Formula: see text] values, and also evaluated its feasibility for [¹⁸F]FP-CIT PET in human brain. The results revealed that the proposed approach provides distribution volume ratio estimation comparable to the Logan plot at low noise levels while improving underestimation caused by non-RE state differently depending on [Formula: see text]. Furthermore, the proposed method was able to avoid noise-induced bias of the Logan plot, and the variability of its results was less dependent on [Formula: see text] than the Logan plot. Therefore, this approach, without issues related to arterial blood sampling given a pre-estimate of [Formula: see text] (e.g. population-based), could be useful in parametric image generation for slow kinetic tracers staying in a non-RE state within a PET scan.

Federated optimisation of kinetic analysis problems

Positron Emission Tomography (PET) data is intrinsically dynamic, and kinetic analysis of dynamic PET data can substantially augment the information provided by static PET reconstructions. Yet despite the insights into disease that kinetic analysis offers, it is not used clinically and seldom used in research beyond the preclinical stage. The utility of PET kinetic analysis is hampered by several factors including spatial inconsistency within regions of homogeneous tissue and relative computational expense when fitting complex models to individual voxels. Even with sophisticated algorithms inconsistencies can arise because local optima frequently have narrow basins of convergence, are surrounded by relatively flat (uninformative) regions, have relatively low-gradient valley floors, or combinations thereof. Based on the observation that cost functions for individual voxels frequently bear some resemblance to each-other, this paper proposes the federated optimisation of the individual kinetic analysis problems within a given image. This approach shares parameters proposed during optimisation with other, similar voxels. Federated optimisation exploits the redundancy typical of large medical images to improve the optimisation residuals, computational efficiency and, to a limited extent, image consistency. This is achieved without restricting the formulation of the kinetic model, resorting to an explicit regularisation parameter, or limiting the resolution at which parameters are computed.

Impact of image-based motion correction on dopamine D3/D2 receptor occupancy-comparison of groupwise and frame-by-frame registration approaches

BACKGROUND

Image registration algorithms are frequently used to align the reconstructed brain PET frames to remove subject head motion. However, in occupancy studies, this is a challenging task where competitive binding of a drug can further reduce the available signal for registration. The purpose of this study is to evaluate two kinds of algorithms-a conventional frame-by-frame (FBF) registration and a recently introduced groupwise image registration (GIR), for motion correction of a dopamine D3/D2 receptor occupancy study.

METHODS

The FBF method co-registers all the PET frames to a common reference based on normalised mutual information as the spatial similarity. The GIR method incorporates a pharmacokinetic model and conducts motion correction by maximising a likelihood function iteratively on tracer kinetics and subject motion. Data from eight healthy volunteers scanned with [11C]-(+)-PHNO pre- and post-administration of a range of doses of the D3 antagonist GSK618334 were used to compare the motion correction performance.

RESULTS

The groupwise registration achieved improved motion correction results, both by visual inspection of the dynamic PET data and by the reduction of the variability in the outcome measures, and required no additional steps to exclude unsuccessfully realigned PET data for occupancy modelling as compared to frame-by-frame registration. Furthermore, for the groupwise method, the resultant binding potential estimates had reduced variation and bias for individual scans and improved half maximal effective concentration (EC50) estimates were obtained for the study as a whole.

CONCLUSIONS

These results indicate that the groupwise registration approach can provide improved motion correction of dynamic brain PET data as compared to frame-by-frame registration approaches for receptor occupancy studies.

Variance Estimation for Myocardial Blood Flow by Dynamic PET

The estimation of myocardial blood flow (MBF) by (13)N-ammonia or (82)Rb dynamic PET typically relies on an empirically determined generalized Renkin-Crone equation to relate the kinetic parameter K1 to MBF. Because the Renkin-Crone equation defines MBF as an implicit function of K1, the MBF variance cannot be determined using standard error propagation techniques. To overcome this limitation, we derived novel analytical approximations that provide first- and second-order estimates of MBF variance in terms of the mean and variance of K1 and the Renkin-Crone parameters. The accuracy of the analytical expressions was validated by comparison with Monte Carlo simulations, and MBF variance was evaluated in clinical (82)Rb dynamic PET scans. For both (82)Rb and (13)N-ammonia, good agreement was observed between both (first- and second-order) analytical variance expressions and Monte Carlo simulations, with moderately better agreement for second-order estimates. The contribution of the Renkin-Crone relation to overall MBF uncertainty was found to be as high as 68% for (82)Rb and 35% for (13)N-ammonia. For clinical (82)Rb PET data, the conventional practice of neglecting the statistical uncertainty in the Renkin-Crone parameters resulted in underestimation of the coefficient of variation of global MBF and coronary flow reserve by 14-49%. Knowledge of MBF variance is essential for assessing the precision and reliability of MBF estimates. The form and statistical uncertainty in the empirical Renkin-Crone relation can make substantial contributions to the variance of MBF. The novel analytical variance expressions derived in this work enable direct estimation of MBF variance which includes this previously neglected contribution.

Precision and accuracy of clinical quantification of myocardial blood flow by dynamic PET: A technical perspective

A number of exciting advances in PET/CT technology and improvements in methodology have recently converged to enhance the feasibility of routine clinical quantification of myocardial blood flow and flow reserve. Recent promising clinical results are pointing toward an important role for myocardial blood flow in the care of patients. Absolute blood flow quantification can be a powerful clinical tool, but its utility will depend on maintaining precision and accuracy in the face of numerous potential sources of methodological errors. Here we review recent data and highlight the impact of PET instrumentation, image reconstruction, and quantification methods, and we emphasize (82)Rb cardiac PET which currently has the widest clinical application. It will be apparent that more data are needed, particularly in relation to newer PET technologies, as well as clinical standardization of PET protocols and methods. We provide recommendations for the methodological factors considered here. At present, myocardial flow reserve appears to be remarkably robust to various methodological errors; however, with greater attention to and more detailed understanding of these sources of error, the clinical benefits of stress-only blood flow measurement may eventually be more fully realized.

Quantification of 18F-fluorocholine kinetics in patients with prostate cancer

Choline kinase is upregulated in prostate cancer, resulting in increased (18)F-fluoromethylcholine uptake. This study used pharmacokinetic modeling to validate the use of simplified methods for quantification of (18)F-fluoromethylcholine uptake in a routine clinical setting.

METHODS

Forty-minute dynamic PET/CT scans were acquired after injection of 204 ± 9 MBq of (18)F-fluoromethylcholine, from 8 patients with histologically proven metastasized prostate cancer. Plasma input functions were obtained using continuous arterial blood-sampling as well as using image-derived methods. Manual arterial blood samples were used for calibration and correction for plasma-to-blood ratio and metabolites. Time-activity curves were derived from volumes of interest in all visually detectable lymph node metastases. (18)F-fluoromethylcholine kinetics were studied by nonlinear regression fitting of several single- and 2-tissue plasma input models to the time-activity curves. Model selection was based on the Akaike information criterion and measures of robustness. In addition, the performance of several simplified methods, such as standardized uptake value (SUV), was assessed.

RESULTS

Best fits were obtained using an irreversible compartment model with blood volume parameter. Parent fractions were 0.12 ± 0.4 after 20 min, necessitating individual metabolite corrections. Correspondence between venous and arterial parent fractions was low as determined by the intraclass correlation coefficient (0.61). Results for image-derived input functions that were obtained from volumes of interest in blood-pool structures distant from tissues of high (18)F-fluoromethylcholine uptake yielded good correlation to those for the blood-sampling input functions (R(2) = 0.83). SUV showed poor correlation to parameters derived from full quantitative kinetic analysis (R(2) < 0.34). In contrast, lesion activity concentration normalized to the integral of the blood activity concentration over time (SUVAUC) showed good correlation (R(2) = 0.92 for metabolite-corrected plasma; 0.65 for whole-blood activity concentrations).

CONCLUSION

SUV cannot be used to quantify (18)F-fluoromethylcholine uptake. A clinical compromise could be SUVAUC derived from 2 consecutive static PET scans, one centered on a large blood-pool structure during 0-30 min after injection to obtain the blood activity concentrations and the other a whole-body scan at 30 min after injection to obtain lymph node activity concentrations.

Molecular imaging of targeted therapies with positron emission tomography: the visualization of personalized cancer care

INTRODUCTION

Molecular imaging has been defined as the visualization, characterization and measurement of biological processes at the molecular and cellular level in humans and other living systems. In oncology it enables to visualize (part of) the functional behaviour of tumour cells, in contrast to anatomical imaging that focuses on the size and location of malignant lesions. Available molecular imaging techniques include single photon emission computed tomography (SPECT), positron emission tomography (PET) and optical imaging. In PET, a radiotracer consisting of a positron emitting radionuclide attached to the biologically active molecule of interest is administrated to the patient. Several approaches have been undertaken to use PET for the improvement of personalized cancer care. For example, a variety of radiolabelled ligands have been investigated for intratumoural target identification and radiolabelled drugs have been developed for direct visualization of the biodistibution in vivo, including intratumoural therapy uptake. First indications of the clinical value of PET for target identification and response prediction in oncology have been reported. This new imaging approach is rapidly developing, but uniformity of scanning processes, standardized methods for outcome evaluation and implementation in daily clinical practice are still in progress. In this review we discuss the available literature on molecular imaging with PET for personalized targeted treatment strategies.

CONCLUSION

Molecular imaging with radiolabelled targeted anticancer drugs has great potential for the improvement of personalized cancer care. The non-invasive quantification of drug accumulation in tumours and normal tissues provides understanding of the biodistribution in relation to therapeutic and toxic effects.

Methodological considerations in quantification of 3'-deoxy-3'-[18F]fluorothymidine uptake measured with positron emission tomography in patients with non-small cell lung cancer

PURPOSE

To investigate the effect of image-derived input functions (IDIF), input function corrections and volume of interest (VOI) size in quantification of [(18)F]FLT uptake in non-small cell lung cancer (NSCLC) patients.

PROCEDURES

Twenty-three NSCLC patients were scanned on a HR+ scanner. IDIFs were defined over the aorta and left ventricle. Activity concentration and metabolite fraction were measured in venous blood samples. Venous blood samples at 30, 40 and 60 min after injection were used to calibrate the IDIF time-activity curves. Adaptive thresholds were used for VOI definition. Full kinetic analysis and simplified measures were performed.

RESULTS

Non-linear regression analysis showed better fits for the irreversible model compared to the reversible model in the majority. Calibrated and metabolite corrected plus plasma-to-blood ratio corrected input function resulted in high correlations between SUV and Patlak K i (Pearson correlation coefficients 0.86-0.96, p value < 0.001). No significant differences in correlation between SUV and Patlak K i were observed with variation of IDIF structure or VOI size.

CONCLUSIONS

Plasma-to-blood ratio correction, metabolite correction and calibration improved the correlation between SUV and Patlak K i significantly, indicating the need for these corrections when K i is used to validate semi-quantitative measures, such as SUV.

Assessment of simplified methods to measure 18F-FLT uptake changes in EGFR-mutated non-small cell lung cancer patients undergoing EGFR tyrosine kinase inhibitor treatment

3'-deoxy-3'-(18)F-fluorothymidine ((18)F-FLT) PET/CT provides a noninvasive assessment of proliferation and, as such, could be a valuable imaging biomarker in oncology. The aim of the present study was to assess the validity of simplified quantitative parameters of (18)F-FLT uptake in non-small cell lung cancer (NSCLC) patients before and after the start of treatment with a tyrosine kinase inhibitor (TKI).

METHODS

Ten patients with metastatic NSCLC harboring an activating epidermal growth factor receptor mutation were included in this prospective observational study. Patients underwent (15)O-H2O and (18)F-FLT PET/CT scanning on 3 separate occasions: within 7 d before treatment, and 7 and 28 d after the first therapeutic dose of a TKI (gefitinib or erlotinib). Dynamic scans were acquired and venous blood samples were collected during the (18)F-FLT scan to measure parent fraction and plasma and whole-blood radioactivity concentrations. Simplified measures (standardized uptake value [SUV] and tumor-to-blood ratio [TBR]) were correlated with fully quantitative measures derived from kinetic modeling.

RESULTS

Twenty-nine of thirty (18)F-FLT PET/CT scans were evaluable. According to the Akaike criterion, a reversible 2-tissue model with 4 rate constants and blood volume parameter was preferred in 84% of cases. Relative therapy-induced changes in SUV and TBR correlated with those derived from kinetic analyses (r(2) = 0.83-0.97, P < 0.001, slope = 0.72-1.12). (18)F-FLT uptake significantly decreased at 7 and 28 d after the start of treatment compared with baseline (P < 0.01). Changes in (18)F-FLT uptake were not correlated with changes in perfusion, as measured using (15)O-H2O.

CONCLUSION

SUV and TBR could both be used as surrogate simplified measures to assess changes in (18)F-FLT uptake in NSCLC patients treated with a TKI, at the cost of a small underestimation in uptake changes or the need for a blood sample and metabolite measurement, respectively.

Evaluation of analytical reconstruction with a new gap-filling method in comparison to iterative reconstruction in [¹¹C]-raclopride PET studies

AIM

The aim of the study was (1) to evaluate the reconstruction strategies with dynamic [¹¹C]-raclopride human positron emission tomography (PET) studies acquired from ECAT high-resolution research tomograph (HRRT) scanner and (2) to justify for the selected gap-filling method for analytical reconstruction with simulated phantom data.

METHODS

A new transradial bicubic interpolation method has been implemented to enable faster analytical 3D-reprojection (3DRP) reconstructions for the ECAT HRRT PET scanner data. The transradial bicubic interpolation method was compared to the other gap-filling methods visually and quantitatively using the numerical Shepp-Logan phantom. The performance of the analytical 3DRP reconstruction method with this new gap-filling method was evaluated in comparison with the iterative statistical methods: ordinary Poisson ordered subsets expectation maximization (OPOSEM) and resolution modeled OPOSEM methods. The image reconstruction strategies were evaluated using human data at different count statistics and consequently at different noise levels. In the assessments, 14 [¹¹C]-raclopride dynamic PET studies (test-retest studies of 7 healthy subjects) acquired from the HRRT PET scanner were used. Besides the visual comparisons of the methods, we performed regional quantitative evaluations over the cerebellum, caudate and putamen structures. We compared the regional time-activity curves (TACs), areas under the TACs and binding potential (BPND) values.

RESULTS AND CONCLUSIONS

The results showed that the new gap-filling method preserves the linearity of the 3DRP method. Results with the 3DRP after gap-filling method exhibited hardly any dependency on the count statistics (noise levels) in the sinograms while we observed changes in the quantitative results with the EM-based methods for different noise contamination in the data. With this study, we showed that 3DRP with transradial bicubic gap-filling method is feasible for the reconstruction of high-resolution PET data with missing sinogram bins. The calculated intraclass correlation coefficient (ICC) values were similar for all tested methods and validated the test-retest study. The gap-filling and 3DRP method can be used for quantitative PET studies in which high temporal information is crucial and can serve as a reference method for comparison studies of the other reconstruction methods.

Quality and precision of parametric images created from PET sinogram data by direct reconstruction: proof of concept

We have previously implemented the direct reconstruction of dense kinetic model parameter images ("parametric images") from sinogram data, and compared it to conventional image domain kinetic parameter estimation methods . Although it has been shown that the direct reconstruction algorithm estimates the kinetic model parameters with lower root mean squared error than the conventional image domain techniques, some theoretical obstacles remain. These obstacles include the difficulty of evaluating the accuracy and precision of the estimated parameters. In image domain techniques, the reconstructed time activity curve (TAC) and the model predicted TAC are compared, and the goodness-of-fit is evaluated as a measure of the accuracy and precision of the estimated parameters. This approach cannot be applied to the direct reconstruction technique as there are no reconstructed TACs. In this paper, we propose ways of evaluating the precision and goodness-of-fit of the kinetic model parameters estimated by the direct reconstruction algorithm. Specifically, precision of the estimates requires the calculation of variance images for each parameter, and goodness-of-fit is addressed by reconstructing the difference between the measured and the fitted sinograms. We demonstrate that backprojecting the difference from sinogram space to image space creates error images that can be examined for goodness-of-fit and model selection purposes. The presence of nonrandom structures in the error images may indicate an inadequacy of the kinetic model that has been incorporated into the direct reconstruction algorithm. We introduce three types of goodness-of-fit images. We propose and demonstrate a number-of-runs image as a means of quantifying the adequacy or deficiency of the model. We further propose and demonstrate images of the F statistic and the change in the Akaike Information Criterion as devices for identifying the statistical advantage of one model over another at each voxel. As direct reconstruction to parametric images proliferates, it will be essential for imagers to adopt methods such as those proposed herein to assess the accuracy and precision of their parametric images.

Generalized separable parameter space techniques for fitting 1K-5K serial compartment models

PURPOSE

Kinetic modeling is widely used to analyze dynamic imaging data, estimating kinetic parameters that quantify functional or physiologic processes in vivo. Typical kinetic models give rise to nonlinear solution equations in multiple dimensions, presenting a complex fitting environment. This work generalizes previously described separable nonlinear least-squares techniques for fitting serial compartment models with up to three tissue compartments and five rate parameters.

METHODS

The approach maximally separates the linear and nonlinear aspects of the modeling equations, using a formulation modified from previous basis function methods to avoid a potential mathematical degeneracy. A fast and robust algorithm for solving the linear subproblem with full user-defined constraints is also presented. The generalized separable parameter space technique effectively reduces the dimensionality of the nonlinear fitting problem to one dimension for 2K-3K compartment models, and to two dimensions for 4K-5K models.

RESULTS

Exhaustive search fits, which guarantee identification of the true global minimum fit, required approximately 10 ms for 2K-3K and 1.1 s for 4K-5K models, respectively. The technique is also amenable to fast gradient-descent iterative fitting algorithms, where the reduced dimensionality offers improved convergence properties. The objective function for the separable parameter space nonlinear subproblem was characterized and found to be generally well-behaved with a well-defined global minimum. Separable parameter space fits with the Levenberg-Marquardt algorithm required fewer iterations than comparable fits for conventional model formulations, averaging 1 and 7 ms for 2K-3K and 4K-5K models, respectively. Sensitivity to initial conditions was likewise reduced.

CONCLUSIONS

The separable parameter space techniques described herein generalize previously described techniques to encompass 1K-5K compartment models, enable robust solution of the linear subproblem with full user-defined constraints, and are amenable to rapid and robust fitting using iterative gradient-descent type algorithms.

Spatio-temporal pharmacokinetic model based registration of 4D PET neuroimaging data

In dynamic positron emission tomography (PET) neuroimaging studies, where scan durations often exceed 1h, registration of motion-corrupted dynamic PET images is necessary in order to maintain the integrity of the physiological, pharmacological, or biochemical information derived from the tracer kinetic analysis of the scan. In this work, we incorporate a pharmacokinetic model, which is traditionally used to analyse PET data following any registration, into the registration process itself in order to allow for a groupwise registration of the temporal time frames. The new method is shown to achieve smaller registration errors and improved kinetic parameter estimates on validation data sets when compared with image similarity based registration approaches. When applied to measured clinical data from 10 healthy subjects scanned with [(11)C]-(+)-PHNO (a dopamine D3/D2 receptor tracer), it reduces the intra-class variability on the receptor binding outcome measure, further supporting the improvements in registration accuracy. Our method incorporates a generic tracer kinetic model which makes it applicable to different PET radiotracers to remove motion artefacts and increase the integrity of dynamic PET studies.

Longitudinal amyloid imaging using 11C-PiB: methodologic considerations

Several methods are in use for analyzing (11)C-Pittsburgh compound-B ((11)C-PiB) data. The objective of this study was to identify the method of choice for measuring longitudinal changes in specific (11)C-PiB binding.

METHODS

Dynamic 90-min (11)C-PiB baseline and follow-up scans (interval, 30 ± 5 mo) were obtained for 7 Alzheimer disease (AD) patients, 11 patients with mild cognitive impairment (MCI), and 11 healthy controls. Parametric images were generated using reference parametric mapping (RPM2), reference Logan values, and standardized uptake value volume ratios (SUVr), the latter for intervals between 60 and 90 (SUVr(60-90)) and 40 and 60 (SUVr(40-60)) minutes after injection. In all analyses, cerebellar gray matter was used as a reference region. A global cortical volume of interest was defined using a probability map-based template. Percentage change between baseline and follow-up was derived for all analytic methods.

RESULTS

SUVr(60-90) and SUVr(40-60) overestimated binding with 13% and 10%, respectively, compared with RPM2. Reference Logan values were on average 6% lower than RPM2. Both SUVr measures showed high intersubject variability. Over time, R1, the delivery of tracer to the cortex relative to that to the cerebellum, decreased in AD patients (P < 0.05) but not in MCI patients and controls. Simulations showed that SUVr, but not RPM2 and reference Logan values, was highly dependent on uptake period and that changes in SUVr over time were sensitive to changes in flow.

CONCLUSION

To reliably assess amyloid binding over time--for example, in drug intervention studies--it is essential to use fully quantitative methods for data acquisition and analysis.

Pharmacokinetic analysis of [18F]FAZA in non-small cell lung cancer patients

PURPOSE

[(18)F]Fluoroazomycin arabinoside (FAZA) is a positron emission tomography (PET) tracer developed to enable identification of hypoxic regions within a tumour. The aims of this study were to determine the optimal kinetic model along with validation of using alternatives to arterial blood sampling for analysing [(18)F]FAZA studies and to assess the validity of simplified analytical methods.

METHODS

Dynamic 70-min [(18)F]FAZA PET/CT scans were obtained from nine non-small cell lung cancer patients. Continuous arterial blood sampling, together with manual arterial and venous sampling, was performed to derive metabolite-corrected plasma input functions. Volumes of interest (VOIs) were defined for tumour, healthy lung muscle and adipose tissue generating [(18)F]FAZA time-activity curves (TACs). TACs were analysed using one- and two-tissue compartment models using both metabolite-corrected blood sampler plasma input functions (BSIF) and image-derived plasma input functions (IDIF).

RESULTS

The reversible two-tissue compartment model with blood volume parameter (2T4k+VB) best described kinetics of [(18)F]FAZA in tumours. Volumes of distribution (VT) obtained using IDIF correlated well with those derived using BSIF (R(2) = 0.82). Venous samples yielded the same radioactivity concentrations as arterial samples for times >50 min post-injection (p.i.). In addition, both plasma to whole blood ratios and parent fractions were essentially the same for venous and arterial samples. Both standardised uptake value (SUV), normalised to lean body mass, and tumour to blood ratio correlated well with VT (R(2) = 0.77 and R(2) = 0.87, respectively, at 50-60 min p.i.), although a bias was observed at low VT.

CONCLUSION

The 2T4k+VB model provided the best fit to the dynamic [(18)F]FAZA data. IDIF with venous blood samples can be used as input function. Further data are needed to validate the use of simplified methods.

Mapping human brain fatty acid amide hydrolase activity with PET

Endocannabinoid tone has recently been implicated in a number of prevalent neuropsychiatric conditions. [(11)C]CURB is the first available positron emission tomography (PET) radiotracer for imaging fatty acid amide hydrolase (FAAH), the enzyme which metabolizes the prominent endocannabinoid anandamide. Here, we sought to determine the most suitable kinetic modeling approach for quantifying [(11)C]CURB that binds selectively to FAAH. Six healthy volunteers were scanned with arterial blood sampling for 90 minutes. Kinetic parameters were estimated regionally using a one-tissue compartment model (TCM), a 2-TCM with and without irreversible trapping, and an irreversible 3-TCM. The 2-TCM with irreversible trapping provided the best identifiability of PET outcome measures among the approaches studied (coefficient of variation (COV) of the net influx constant K(i) and the composite parameter λk(3) (λ=K(1)/k(2)) <5%, and COV(k(3))<10%). Reducing scan time to 60 minutes did not compromise the identifiability of rate constants. Arterial spin labeling measures of regional cerebral blood flow were only slightly correlated with K(i), but not with k(3) or λk(3). Our data suggest that λk(3) is sensitive to changes in FAAH activity, therefore, optimal for PET quantification of FAAH activities with [(11)C]CURB. Simulations showed that [(11)C]CURB binding in healthy subjects is far from a flow-limited uptake.

Modeling cyclosporine A inhibition of the distribution of a P-glycoprotein PET ligand, 11C-verapamil, into the maternal brain and fetal liver of the pregnant nonhuman primate: impact of tissue blood flow and site of inhibition

Through PET imaging, our laboratory has studied the dynamic biodistribution of (11)C-verapamil, a P-gp substrate, in the nonhuman primate Macaca nemestrina. To gain detailed insight into the kinetics of verapamil transport across the blood-brain barrier (BBB) and the blood-placental barrier (BPB), we analyzed these dynamic biodistribution data by compartmental modeling.

METHODS

Thirteen pregnant macaques (gestational age, 71-159 d; term, ∼172 d) underwent PET imaging with (11)C-verapamil before and during infusion (6, 12, or 24 mg/kg/h) of cyclosporine A (CsA, a P-glycoprotein [P-gp] inhibitor). Dynamic (11)C-verapamil brain or fetal liver (reporter of placental P-gp function) activity was assessed by a 1- or 2-tissue-compartment model.

RESULTS

The 1-tissue-compartment model best explained the observed brain and fetal liver distribution of (11)C-radioactivity. When P-gp was completely inhibited, the brain and fetal liver distribution clearance (K1) approximated tissue blood flow (Q); that is, extraction ratio (K1/Q) was approximately 1, indicating that in the absence of P-gp function, the distribution of (11)C-verapamil radioactivity into these compartments is limited by blood flow. The potency of CsA to inhibit P-gp was tissue-independent (maternal BBB half-maximal inhibitory concentration [IC50], 5.67 ± 1.07 μM, vs. BPB IC50, 7.63 ± 3.16 μM).

CONCLUSION

We propose that on deliberate or inadvertent P-gp inhibition, the upper boundary of increase in human brain (or fetal) distribution of lipophilic drugs such as verapamil will be limited by tissue blood flow. This finding provides a means to predict the magnitude of P-gp-based drug interactions at the BBB and BPB when only the baseline distribution of the drug (i.e., in the absence of P-gp inhibition) across these barriers is available through PET. Our data suggest that P-gp-based drug interactions at the human BBB and BPB can be clinically significant, particularly for those P-gp substrate drugs for which P-gp plays a significant role in excluding the drug from these privileged compartments.

3.5D dynamic PET image reconstruction incorporating kinetics-based clusters

Standard 3D dynamic positron emission tomographic (PET) imaging consists of independent image reconstructions of individual frames followed by application of appropriate kinetic model to the time activity curves at the voxel or region-of-interest (ROI). The emerging field of 4D PET reconstruction, by contrast, seeks to move beyond this scheme and incorporate information from multiple frames within the image reconstruction task. Here we propose a novel reconstruction framework aiming to enhance quantitative accuracy of parametric images via introduction of priors based on voxel kinetics, as generated via clustering of preliminary reconstructed dynamic images to define clustered neighborhoods of voxels with similar kinetics. This is then followed by straightforward maximum a posteriori (MAP) 3D PET reconstruction as applied to individual frames; and as such the method is labeled '3.5D' image reconstruction. The use of cluster-based priors has the advantage of further enhancing quantitative performance in dynamic PET imaging, because: (a) there are typically more voxels in clusters than in conventional local neighborhoods, and (b) neighboring voxels with distinct kinetics are less likely to be clustered together. Using realistic simulated (11)C-raclopride dynamic PET data, the quantitative performance of the proposed method was investigated. Parametric distribution-volume (DV) and DV ratio (DVR) images were estimated from dynamic image reconstructions using (a) maximum-likelihood expectation maximization (MLEM), and MAP reconstructions using (b) the quadratic prior (QP-MAP), (c) the Green prior (GP-MAP) and (d, e) two proposed cluster-based priors (CP-U-MAP and CP-W-MAP), followed by graphical modeling, and were qualitatively and quantitatively compared for 11 ROIs. Overall, the proposed dynamic PET reconstruction methodology resulted in substantial visual as well as quantitative accuracy improvements (in terms of noise versus bias performance) for parametric DV and DVR images. The method was also tested on a 90 min (11)C-raclopride patient study performed on the high-resolution research tomography. The proposed method was shown to outperform the conventional method in visual as well as quantitative accuracy improvements (in terms of noise versus regional DVR value performance).

Fourier domain closed-form formulas for estimation of kinetic parameters in reversible multi-compartment models

Background

Compared with static imaging, dynamic emission computed tomographic imaging with compartment modeling can quantify in vivo physiologic processes, providing useful information about molecular disease processes. Dynamic imaging involves estimation of kinetic rate parameters. For multi-compartment models, kinetic parameter estimation can be computationally demanding and problematic with local minima.

Methods

This paper offers a new perspective to the compartment model fitting problem where Fourier linear system theory is applied to derive closed-form formulas for estimating kinetic parameters for the two-compartment model. The proposed Fourier domain estimation method provides a unique solution, and offers very different noise response as compared to traditional non-linear chi-squared minimization techniques.

Results

The unique feature of the proposed Fourier domain method is that only low frequency components are used for kinetic parameter estimation, where the DC (i.e., the zero frequency) component in the data is treated as the most important information, and high frequency components that tend to be corrupted by statistical noise are discarded. Computer simulations show that the proposed method is robust without having to specify the initial condition. The resultant solution can be fine tuned using the traditional iterative method.

Conclusions

The proposed Fourier-domain estimation method has closed-form formulas. The proposed Fourier-domain curve-fitting method does not require an initial condition, it minimizes a quadratic objective function and a closed-form solution can be obtained. The noise is easier to control, simply by discarding the high frequency components, and emphasizing the DC component.

Kinetic parameter estimation using a closed-form expression via integration by parts

Dynamic emission computed tomographic imaging with compartment modeling can quantify in vivo physiologic processes, eliciting more information regarding underlying molecular disease processes than is obtained from static imaging. However, estimation of kinetic rate parameters for multi-compartment models can be computationally demanding and problematic due to local minima. A number of techniques for kinetic parameter estimation have been studied and are in use today, generally offering a tradeoff between computation time, robustness of fit and flexibility with differing sets of assumptions. This paper presents a means to eliminate all differential operations by using the integration-by-parts method to provide closed-form formulas, so that the mathematical model is less sensitive to data sampling and noise. A family of closed-form formulas are obtained. Computer simulations show that the proposed method is robust without having to specify the initial condition.

Computation of variance in compartment model parameter estimates from dynamic PET data

PURPOSE

This paper investigates the validity of the analytical framework for variance ("analytical variance") in kinetic parameter and macroparameter estimations. Analytical variance is compared against the variance obtained from Monte Carlo simulations ("MC variance") for two different compartment models at different noise levels.

METHODS

Kinetic parameters for one-tissue (1T) and two-tissue (2T) compartment models are used to generate time-activity curves (TAC). Gaussian noise is added to the noiseless TAC to generate noise realizations for each noise level. The kinetic parameters are then estimated by minimizing the weighted squared error between the noisy TAC and the model output. Standard deviation is computed statistically from the estimated parameters and computed analytically using the framework at each noise level. The ratio of standard deviation to true parameter value obtained from Monte Carlo simulations and analytical computations is compared.

RESULTS

Difference between the analytical and MC variance increases with the level of noise and complexity of the compartment model. The standard deviation of the analytical variance also increases with the noise-level and model complexity. The difference between the analytical and MC variance is less than 3% for 1T compartment model and less than 10% for 2T compartment model at all noise levels. In addition, the standard deviation in the analytical variance is less than 15% for 1T and 2T compartment models at all noise levels.

CONCLUSIONS

These results indicate that the proposed framework for the variance in the kinetic parameter estimations can be used for 1T and 2T compartment models even in the existence of high noise.