Kinetic model-based factor analysis of dynamic sequences for 82-rubidium cardiac positron emission tomography

Enhanced Extraction of Blood and Tissue Time-Activity Curves in Cardiac Mouse FDG PET Imaging by Means of Constrained Nonnegative Matrix Factorization

We propose an enhanced method to accurately retrieve time-activity curves (TACs) of blood and tissue from dynamic 2-deoxy-2-[¹⁸F]fluoro-D-glucose ([¹⁸F]FDG) positron emission tomography (PET) cardiac images of mice. The method is noninvasive and consists of using a constrained nonnegative matrix factorization algorithm (CNMF) applied to the matrix (A) containing the intensity values of the voxels of the left ventricle (LV) PET image. CNMF factorizes A into nonnegative matrices H and W, respectively, representing the physiological factors (blood and tissue) and their associated weights, by minimizing an extended cost function. We verified our method on 32 C57BL/6 mice, 14 of them with acute myocardial infarction (AMI). With CNMF, we could break down the mouse LV into myocardial and blood pool images. Their corresponding TACs were used in kinetic modeling to readily determine the [¹⁸F]FDG influx constant (K_i) required to compute the myocardial metabolic rate of glucose. The calculated K_i values using CNMF for the heart of control mice were in good agreement with those published in the literature. Significant differences in K_i values for the heart of control and AMI mice were found using CNMF. The values of the elements of W agreed well with the LV structural changes induced by ligation of the left coronary artery. CNMF was compared with the recently published method based on robust unmixing of dynamic sequences using regions of interest (RUDUR). A clear improvement of signal separation was observed with CNMF compared to the RUDUR method.

Compartment model-based nonlinear unmixing for kinetic analysis of dynamic PET images

When no arterial input function is available, quantification of dynamic PET images requires a previous step devoted to the extraction of a reference time-activity curve (TAC). Factor analysis is often applied for this purpose. This paper introduces a novel approach that conducts a new kind of nonlinear factor analysis relying on a compartment model, and computes the kinetic parameters of specific binding tissues jointly. To this end, it capitalizes on data-driven parametric imaging methods to provide a physical description of the underlying PET data, directly relating the specific binding with the kinetics of the non-specific binding in the corresponding tissues. This characterization is introduced into the factor analysis formulation to yield a novel nonlinear unmixing model designed for PET image analysis. This model also explicitly introduces global kinetic parameters that allow for a direct estimation of a binding potential that represents the ratio at equilibrium of specifically bound radioligand to the concentration of nondisplaceable radioligand in each non-specific binding tissue. The performance of the method is evaluated on synthetic and real data to demonstrate its potential interest.

Quantification of myocardial blood flow (MBF) and reserve (MFR) incorporated with a novel segmentation approach: Assessments of quantitative precision and the lower limit of normal MBF and MFR in patients

BACKGROUND

Quantification of myocardial blood flow (MBF) and myocardial flow reserve (MFR) has shown diagnostic and prognostic values for the assessment of coronary artery disease (CAD). This study aimed to evaluate in patients a highly automatic Yale-MQ (myocardial blood flow quantification) software incorporated with a novel image segmentation approach for quantification of global and regional MBF and MFR from dynamic 82Rb cardiac positron emission tomography (PET).

METHODS

Global and regional MBFs and MFRs were quantified in 80 patients (18 normal and 62 CAD subjects) by two different observers using the Yale-MQ software. Lower limits of normal (LLN) values and intra- and inter-observer variabilities of MBFs and MFRs were calculated for the assessment of quantitative precision. The Yale-MQ was compared with a commercially available software (Corridor 4DM) being used as a reference.

RESULTS

The Yale-MQ method provided precise assessments of LLNs of MBF and MFR. The global and regional MBFs and MFR quantified via Yale-MQ were correlated strongly with those via Corridor4DM (R ≥ 0.867). The intra- and inter-observer variabilities of MBFs and MFRs quantified via Yale-MQ were small (≤ 7.7% for MBFs and ≤ 10.0% for MFRs) with excellent correlations (R ≥ 0.980 for MBFs and R ≥ 0.976 for MFRs).

CONCLUSIONS

The new Yale-MQ software associated with the automatic processing scheme provides a highly reproducible clinical tool for precise quantification of MBF and MFR in patients with reliable LLN values.

Consistent tracer administration profile improves test-retest repeatability of myocardial blood flow quantification with 82Rb dynamic PET imaging

OBJECTIVES

Quantification of myocardial blood flow (MBF) and stress/rest flow reserve is used increasingly to diagnose multi-vessel coronary artery disease and micro-vascular disease with PET imaging. However, variability in the measurements may limit physician confidence to direct revascularization therapies based on specific threshold values. This study evaluated the effects of rubidium-82 (82Rb) tracer injection profile using a constant-activity-rate (CA) vs a constant-flow-rate (CF) infusion to improve test-retest repeatability of MBF measurements.

METHOD

22 participants underwent single-session 82Rb dynamic PET imaging during rest and dipyridamole stress using one of 2 test-retest infusion protocols: CA-CA (n = 12) or CA-CF (n = 10). MBF was quantified using a single-tissue-compartment model (1TCM) and a simplified retention model (SRM). Non-parametric test-retest repeatability coefficients (RPCnp) were compared between groups. Myocardium-to-blood contrast and signal-to-noise ratios of the late uptake images (2 to 6 minutes) were also compared to evaluate standard myocardial perfusion image (MPI) quality.

RESULTS

MBF values in the CA-CA group were more repeatable (smaller RPCnp) than the CA-CF group using the 1TCM at rest alone, rest and stress combined, and stress/rest reserve (21% vs 36%, 16% vs 19%, and 20% vs 27%, P < 0.05, respectively), and using the SRM at Rest and Stress alone, Rest and Stress combined, and stress/rest reserve (21% vs 38%, 15% vs 25%, 22% vs 38%, and 23% vs 49%, P < 0.05, respectively). In terms of image quality, myocardium-to-blood contrast and signal-to-noise ratios were not significantly different between groups.

CONCLUSIONS

Constant-activity-rate 'square-wave' infusion of 82Rb produces more repeatable tracer injection profiles and decreases the test-retest variability of MBF measurements, when compared to a constant-flow-rate 'bolus' administration of 82Rb, especially with SRM, and without compromising standard MPI quality.

[Kinetic cluster and α-divergence-based dynamic myocardial factorial analysis of positron-emission computed tomography images]

OBJECTIVE

We purpose a novel factor analysis method based on kinetic cluster and α-divergence measure for extracting the blood input function and the time-activity curve of the regional tissue from dynamic myocardial positron emission computed tomography(PET) images.

METHODS

Dynamic PET images were decomposed into initial factors and factor images by minimizing the α-divergence between the factor model and actual image data. The kinetic clustering as a priori constraint was then incorporated into the model to solve the nonuniqueness problem, and the tissue time-activity curves and the tissue space distributions with physiological significance were generated.

RESULTS

The model was applied to the 82RbPET myocardial perfusion simulation data and compared with the traditional model-based least squares measure and the minimal spatial overlap constraint. The experimental results showed that the proposed model performed better than the traditional model in terms of both accuracy and sensitivity.

CONCLUSION

This method can select the optimal measure by α value, and incorporate the prior information of the kinetic clustering of PET image pixels to obtain the accurate time-activity curves of the tissue, which has shown good performance in visual evaluation and quantitative evaluation.

Optimally Repeatable Kinetic Model Variant for Myocardial Blood Flow Measurements with 82Rb PET

Purpose. Myocardial blood flow (MBF) quantification with ⁸²Rb positron emission tomography (PET) is gaining clinical adoption, but improvements in precision are desired. This study aims to identify analysis variants producing the most repeatable MBF measures. Methods. 12 volunteers underwent same-day test-retest rest and dipyridamole stress imaging with dynamic ⁸²Rb PET, from which MBF was quantified using 1-tissue-compartment kinetic model variants: (1) blood-pool versus uptake region sampled input function (Blood/Uptake-ROI), (2) dual spillover correction (SOC-On/Off), (3) right blood correction (RBC-On/Off), (4) arterial blood transit delay (Delay-On/Off), and (5) distribution volume (DV) constraint (Global/Regional-DV). Repeatability of MBF, stress/rest myocardial flow reserve (MFR), and stress/rest MBF difference (ΔMBF) was assessed using nonparametric reproducibility coefficients (RPC_np = 1.45 × interquartile range). Results. MBF using SOC-On, RVBC-Off, Blood-ROI, Global-DV, and Delay-Off was most repeatable for combined rest and stress: RPC_np = 0.21 mL/min/g (15.8%). Corresponding MFR and ΔMBF RPC_np were 0.42 (20.2%) and 0.24 mL/min/g (23.5%). MBF repeatability improved with SOC-On at stress (p < 0.001) and tended to improve with RBC-Off at both rest and stress (p < 0.08). DV and ROI did not significantly influence repeatability. The Delay-On model was overdetermined and did not reliably converge. Conclusion. MBF and MFR test-retest repeatability were the best with dual spillover correction, left atrium blood input function, and global DV.

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.

Absolute myocardial flow quantification with (82)Rb PET/CT: comparison of different software packages and methods

PURPOSE

In clinical cardiac (82)Rb PET, globally impaired coronary flow reserve (CFR) is a relevant marker for predicting short-term cardiovascular events. However, there are limited data on the impact of different software and methods for estimation of myocardial blood flow (MBF) and CFR. Our objective was to compare quantitative results obtained from previously validated software tools.

METHODS

We retrospectively analyzed cardiac (82)Rb PET/CT data from 25 subjects (group 1, 62 ± 11 years) with low-to-intermediate probability of coronary artery disease (CAD) and 26 patients (group 2, 57 ± 10 years; P=0.07) with known CAD. Resting and vasodilator-stress MBF and CFR were derived using three software applications: (1) Corridor4DM (4DM) based on factor analysis (FA) and kinetic modeling, (2) 4DM based on region-of-interest (ROI) and kinetic modeling, (3) MunichHeart (MH), which uses a simplified ROI-based retention model approach, and (4) FlowQuant (FQ) based on ROI and compartmental modeling with constant distribution volume.

RESULTS

Resting and stress MBF values (in milliliters per minute per gram) derived using the different methods were significantly different: using 4DM-FA, 4DM-ROI, FQ, and MH resting MBF values were 1.47 ± 0.59, 1.16 ± 0.51, 0.91 ± 0.39, and 0.90 ± 0.44, respectively (P<0.001), and stress MBF values were 3.05 ± 1.66, 2.26 ± 1.01, 1.90 ± 0.82, and 1.83 ± 0.81, respectively (P<0.001). However, there were no statistically significant differences among the CFR values (2.15 ± 1.08, 2.05 ± 0.83, 2.23 ± 0.89, and 2.21 ± 0.90, respectively; P=0.17). Regional MBF and CFR according to vascular territories showed similar results. Linear correlation coefficient for global CFR varied between 0.71 (MH vs. 4DM-ROI) and 0.90 (FQ vs. 4DM-ROI). Using a cut-off value of 2.0 for abnormal CFR, the agreement among the software programs ranged between 76 % (MH vs. FQ) and 90 % (FQ vs. 4DM-ROI). Interobserver agreement was in general excellent with all software packages.

CONCLUSION

Quantitative assessment of resting and stress MBF with (82)Rb PET is dependent on the software and methods used, whereas CFR appears to be more comparable. Follow-up and treatment assessment should be done with the same software and method.

Small-animal molecular imaging methods

The ability to trace or identify specific molecules within a specific anatomic location provides insight into metabolic pathways, tissue components, and tracing of solute transport mechanisms. With the increasing use of small animals for research, such imaging must have sufficiently high spatial resolution to allow anatomic localization as well as sufficient specificity and sensitivity to provide an accurate description of the molecular distribution and concentration.

METHODS

Imaging methods based on electromagnetic radiation, such as PET, SPECT, MRI, and CT, are increasingly applicable because of recent advances in novel scanner hardware and image reconstruction software and the availability of novel molecules that have enhanced sensitivity in these methodologies.

RESULTS

Small-animal PET has been advanced by the development of detector arrays that provide higher resolution and positron-emitting elements that allow new molecular tracers to be labeled. Micro-MRI has been improved in terms of spatial resolution and sensitivity through increased magnet field strength and the development of special-purpose coils and associated scan protocols. Of particular interest is the associated ability to image local mechanical function and solute transport processes, which can be directly related to the molecular information. This ability is further strengthened by the synergistic integration of PET with MRI. Micro-SPECT has been improved through the use of coded aperture imaging approaches as well as image reconstruction algorithms that can better deal with the photon-limited scan data. The limited spatial resolution can be partially overcome by integrating SPECT with CT. Micro-CT by itself provides exquisite spatial resolution of anatomy, but recent developments in high-spatial-resolution photon counting and spectrally sensitive imaging arrays, combined with x-ray optical devices, hold promise for actual molecular identification by virtue of the chemical bond lengths of molecules, especially biopolymers.

CONCLUSION

Given the increasing use of small animals for evaluating new clinical imaging techniques and providing more insight into pathophysiologic phenomena as well as the availability of improved detection systems, scanning protocols, and associated software, the sensitivity and specificity of molecular imaging are increasing.