Ravindra Mohan Manjeshwar

Quantitative comparison of OSEM and penalized likelihood image reconstruction using relative difference penalties for clinical PET.

Ordered subset expectation maximization (OSEM) is the most widely used algorithm for clinical PET image reconstruction. OSEM is usually stopped early and post-filtered to control image noise and does not necessarily achieve optimal quantitation accuracy. As an alternative to OSEM, we have recently implemented a penalized likelihood (PL) image reconstruction algorithm for clinical PET using the relative difference penalty with the aim of improving quantitation accuracy without compromising visual image quality. Preliminary clinical studies have demonstrated visual image quality including lesion conspicuity in images reconstructed by the PL algorithm is better than or at least as good as that in OSEM images. In this paper we evaluate lesion quantitation accuracy of the PL algorithm with the relative difference penalty compared to OSEM by using various data sets including phantom data acquired with an anthropomorphic torso phantom, an extended oval phantom and the NEMA image quality phantom; clinical data; and hybrid clinical data generated by adding simulated lesion data to clinical data. We focus on mean standardized uptake values and compare them for PL and OSEM using both time-of-flight (TOF) and non-TOF data. The results demonstrate improvements of PL in lesion quantitation accuracy compared to OSEM with a particular improvement in cold background regions such as lungs.

Theoretical Comparison of Motion Correction Techniques for PET Image Reconstruction

We theoretically compare the bias and variance properties of two motion correction techniques for regularized PET image reconstruction from motion-gated datasets. The first method (RRA) consists of independent reconstructions of all gates followed by registration to a reference gate and averaging. The second method (MBMC) uses the entire dataset simultaneously to reconstruct the motion corrected reference gate image by including the motion information in the system model as a warping matrix. Both methods are capable of correcting for non-rigid motion and assume the presence of accurate gate-to-gate motion information, which can be obtained by co-registering gated anatomical images. In order to compare the two techniques, we simulated a motion phantom where a circular phantom warps into an ellipse and a modified version of the NCAT phantom with respiratory motion. We determined differences in voxelwise variances between the two methods as a function of the smoothing parameter when biases are matched. We showed that the MBMC approach results in lower variance for maximum-likelihood (ML) image reconstruction with or without linear post-smoothing. In penalized-likelihood (PL) image reconstruction with quadratic penalties, the MBMC approach outperforms the RRA approach up to a certain degree of smoothing beyond which both methods have approximately equal variance.

Motion compensated image reconstruction of respiratory gated PET/CT

Detectability of lung tumors in PET images is severely compromised owing to respiratory motion. A model based motion compensation method for image reconstruction from respiratory-gated PET/CT data is presented. The object at each gate is modeled as a deformation of the object at a reference gate. Gated CT images are co-registered to the image at the reference gate to estimate object motion. This motion information is in turn used as part of the system model in the forward and back projection steps of the reconstruction algorithm. We simulated the respiratory-gated PET/CT acquisition of the mathematical NCAT phantom with realistic respiratory motion. The model-based approach was compared with post-reconstruction registration of independently reconstructed gates. Simulation results show that the model-based approach achieves better tumor detectability than methods based on post-reconstruction registration

3D implementation of Scatter Estimation in 3D PET

Successful 3D imaging requires accurate and robust methods for scatter estimation and correction. We developed a computationally efficient fully 3D approach modeling both the axial and trans-axial scatter components. Simulation results showed good agreement with the Monte Carlo scatter and improved image quality (IQ). We tested the proposed algorithm on clinical data with similar IQ improvements.

Fully 3D PET Iterative Reconstruction Using Distance-Driven Projectors and Native Scanner Geometry

Incorporating all data corrections into the system model optimizes image quality in statistical iterative PET image reconstruction. We have previously shown that including attenuation, randoms and scatter in the forward 3D iterative model results in faster convergence and improved image quality for ML-OSEM. This paper extends this work to allow the accurate modeling of crystal efficiency, detector deadtime, and the native block-based detector geometry. In order to model these effects, it is necessary to perform forward and back-projections directly from image space to the projection geometry of the PET scanner, rather than to an idealized, equally spaced projection space. We have modified the distance-driven projectors to accurately model both the uneven spacing of the sinogram due to the ring curvature as well as the gaps resulting from the block structure of the scanner. This results in a reconstruction method, which can incorporate the crystal efficiency and block deadtime effects into the forward system model while maintaining the fast reconstruction times enabled by the distance driven projector design. Results on the GE Discovery STEtrade scanner show improvements in image resolution consistent with removing the interpolative smoothing of the data into the equally spaced projection space.

A framework for automated tumor detection in thoracic FDG pet images using texture-based features

This paper proposes a novel framework for tumor detection in Positron Emission Tomography (PET) images. A set of 8 second-order texture features obtained from the gray level co-occurrence matrix (GLCM) across 26 offsets, together with uptake value was used to construct a feature vector at each voxel in the data. Volume of Interest (VOI) samples from 42 images (7 patients with 6 gates each), marked by a radiologist, representing 5 distinct anatomy types and pathology were used to train a logit boost classifier. A ten-fold cross-validation showed a true positive rate of 96%and a false positive rate of 8% for tumor classification. The test dataset consisted of 50 × 50 × 40 representative VOIs from gated PET images of 3 patients. The classifier was run on the test data, followed by an SUV-based thresholding and elimination of noise using connected component analysis. The method detected 10/12 (83%) tumors while detecting an average of 20 false positive structures.

Comparative evaluation of scatter correction in 3D PET using different scatter-level approximations

ObjectiveIn 3D PET, scatter of the gamma photons is one of the most significant physical factors which degrades not only image quality but also quantification. The currently most used scatter estimation method is the analytic single scatter simulation (SSS) which usually accommodates for multiple scattering by scaling the single scatter estimation. However, it has not been clear yet how accurate this approximation is for cases where multiple scatter is significant, raising the question: “How important is correction for multiple scattered photons, and how accurately do we need to simulate all scattered events by appropriate scaling?” This study answers these questions and evaluates the accuracy of SSS implementation in the open-source library STIR.MethodsDifferent scatter orders approximations are evaluated including different levels of scattering and different scaling approaches using Monte Carlo (i.e. SimSET) data. SimSET simulations of a large anthropomorphic phantom were reconstructed with iterative reconstruction algorithms. Images reconstructed with 3D filtered back-projection reprojection algorithm have been compared quantitatively in order to clarify the errors due to different scatter order approximations.ResultsQuantification in regions has improved by scatter correction. For example, in the heart the ideal value was 3, whereas before scatter correction the standard uptake value (SUV) was 4.0, after single scatter correction was 3.3 and after single and double scatter correction was 3.0. After correction by scaling single scatter with tail-fit, the SUV was 3.1, whereas with total-fit it was 3.0. Similarly, for the SSS correction methodology implemented in STIR using tail-fit the heart SUV was 3.1 whereas using total-fit it was 3.0.ConclusionsThe results demonstrate that correction for double scatter improves image contrast and therefore it is required for the accurate estimation of activity distribution in PET imaging. However, it has been also shown that scaling the single scatter distribution is a reasonable approximation to compensate for total scatter. Finally, scatter correction with STIR has shown excellent agreement with Monte Carlo simulations.

Impact of PSF modelling on the convergence rate and edge behaviour of EM images in PET

EM reconstructions with point-spread-function (PSF) modelling is performed to increase the spatial resolution in PET images. These images exhibit slower initial convergence compared to reconstructions without PSF modelling. Furthermore, they exhibit more pronounced ringing around the edges of sharp features. We investigate the effect of different objects and PSF modelling on the convergence rate and edge behaviour of the EM algorithm in two stages: (i) at the initial iterations where the updates are large and (ii) at the later iterations where the updates are small. For the initial iterations, we compare the sharpness of the EM updates with and without PSF modelling. We show via simulations that the PSF modelling during the backprojection step causes smoother updates and consequently smoother images in the early stages of the EM algorithm. For the later iterations, we approximate the image as the ML image plus a perturbation term and develop an approximate update equation for the perturbation, which depends on the Hessian (H) of the log-likelihood. Based on this equation and the spectral analysis of H, we demonstrate how edges with ringing are preserved in the later stages of the algorithm and eliminated only for the case of noiseless data reconstruction with an unrealistically high number of iterations. In addition, we provide an intuitive explanation for the creation of the edge artefacts in terms of the PSF modelling during the backprojection step.

A Computational Positron Emission Tomography Simulation Model for Imaging β-Amyloid in Mice

PurposeWe aimed to develop a computational simulation model for β-amyloid (Aβ) positron emission tomography (PET) imaging.ProceduresModel parameters were set to reproduce levels of Aβ within the PDAPP mouse. Pharmacokinetic curves of virtual tracers were computed and a PET detector simulator was configured for a commercially available preclinical PET-imaging system.ResultsWe modeled the effects of Aβ therapy and tracer affinity on the ability to differentiate Aβ levels by PET. Varying affinity had a significant effect on the ability to quantitate Aβ. Further, PET tracers for Aβ monomers were more sensitive to the therapeutic reduction in Aβ levels than total brain amyloid. Following therapy, the decrease in total brain Aβ corresponded to the slow rate of change in total amyloid load as expected.ConclusionsWe have developed a first proof-of-concept Aβ-PET simulation model that will be a useful tool in the interpretation of preclinical Aβ imaging data and tracer development.

Comparison of different methods for data-driven respiratory gating of PET data

Respiratory movement degrades image quality in PET/CT. The first step in correcting for movement is to gate the data into different motion states. In current practice, the gating is based on information from external devices that measure physical parameters such as the chest position. Various groups have proposed methods to extract a gating signal out of the PET data. Here we compare methods using PCA, Laplacian Eigenmaps, Spectral Analysis and sensitivity. We test the methods on clinical PET list mode data for different tracers. We evaluate correlation with the chest position as measured by the Varian RPM system, and stability under increased noise in the PET data, both by reducing counts and reducing total duration. We also compare SUVmax and lesion displacement when gating the PET data based on the signals extracted by the different methods.