Maximum Likelihood Estimation Maximization Research Articles

Machine learning-based event recognition in SiFi Compton camera imaging for proton therapy monitoring

Objective. Online monitoring of dose distribution in proton therapy is currently being investigated with the detection of prompt gamma (PG) radiation emitted from a patient during irradiation. The SiPM and scintillation Fiber based Compton Camera (SiFi-CC) setup is being developed for this aim. Approach. A machine learning approach to recognize Compton events is proposed, reconstructing the PG emission profile during proton therapy. The proposed method was verified on pseudo-data generated by a Geant4 simulation for a single proton beam impinging on a polymethyl methacrylate (PMMA) phantom. Three different models including the boosted decision tree (BDT), multilayer perception (MLP) neural network, and k-nearest neighbour (k-NN) were trained using 10-fold cross-validation and then their performances were assessed using the receiver operating characteristic (ROI) curves. Subsequently, after event selection by the most robust model, a software based on the List-Mode Maximum Likelihood Estimation Maximization (LM-MLEM) algorithm was applied for the reconstruction of the PG emission distribution profile. Main results. It was demonstrated that the BDT model excels in signal/background separation compared to the other two. Furthermore, the reconstructed PG vertex distribution after event selection showed a significant improvement in distal falloff position determination. Significance. A highly satisfactory agreement between the reconstructed distal edge position and that of the simulated Compton events was achieved. It was also shown that a position resolution of 3.5 mm full width at half maximum (FWHM) in distal edge position determination is feasible with the proposed setup.

Reducing Respiratory Motion-Induced Errors in PET and CT Using an Iterative Image-to-Image Reconstruction Algorithm

Respiratory motion may change SUVmax and SUVpeak in PET images and compromise target definition and treatment response assessment in radiation therapy. While 4D imaging techniques may help address these issues, the time required for 4D data acquisition, image reconstruction and administrative costs keep 3D PET imaging staying as the mainstream in clinical practice. The purpose of this study is to explore the potential of using 4DCT information to remove motion-induced errors in 3D PET/CT images. A novel iterative image-to-image (III) reconstruction algorithm was developed. Specifically, a motion-blurred PET image was considered as a weighted average of PET images acquired at different respiration phases. PET images at each phase were assumed to be deformed from a virtual reference image. Image-intensity-based B-Spline registrations were performed on 4DCT from the end-inhale to the other phases. Resultant deformation maps were used to establish a set of composition equations between the virtual reference image and the blurred PET image. A maximum likelihood estimation maximization (MLEM) algorithm was used to solve these equations. The solution of these equations was smoothed by minimizing total gradient variations before generating the virtual reference image. A 10-phase 4D dataset with the resolution of 3.64×3.64×3.24 for PET and1.37×1.37×2.15 for CT was acquired from a lung cancer patient and used to benchmark the computed results. Different from previous studies, the III-reconstruction was performed in the image space, so there is no any tracking device or sorting operation required during the PET image acquisition. For the blurred PET, III-reconstructed PET and PET0 (which is the 4D-PET at the end-inhale phase), their relative SUVpeak are 10.2, 12.7 and 12.8, and SUVmax are 9.2, 12.5 and 13.1. With PET0 taken as the baseline, the III-reconstruction reduced the motion-induced error from 20.3% to 0.8% for SUVpeak, and from 29.8% to 4.6% for SUVmax. The mean SUV difference between the III-reconstructed PET and PET0 was 1.8% in the tumor region, and 13.2% over the whole image domain. Large differences between the two PET images were located mainly in the chest wall where ribs were mis-registered about 6 mm due to discontinuous motion between the ribs and lungs. The III-reconstruction algorithm was also applied to the 4DCT image. Image contrast ratios for the blurred 3DCT, III-reconstructed CT, and CT0 are 1.8, 3.7 and 4.0, respectively. The III-reconstruction method can be used to correct motion-induced errors in SUVmax and SUVpeak and improve image contrasts in free-breathing PET and CT images. This method does not require sorting projection data, thereby reducing financing cost and clinical burden. Further investigation of this reconstruction method is warranted.

An automatic algorithm to exploit the symmetries of the system response matrix in positron emission tomography iterative reconstruction

Positron emission tomography (PET) iterative 3D reconstruction is a very computational demanding task. One of the main issues of the iterative reconstruction concerns the management of the system response matrix (SRM). The SRM models the relationship between the projection and the voxel space and its memory footprint can easily exceed hundreds of GB. Moreover, in order to make the reconstruction fast enough not to hinder its practical application, the SRM must be stored in the random access memory of the workstation used for the reconstruction. This issue is normally solved by implementing efficient storage schemes and by reducing the number of redundant patterns in the SRM through symmetries. However, finding a sufficient number of symmetries is often non-trivial and is typically performed using dedicated solutions that cannot be exported to different detectors and geometries. In this paper, an automatic approach to reduce the memory footprint of a pre-computed SRM is described. The proposed approach was named symmetry search algorithm (SSA) and consists in an algorithm that searches for some of the redundant patterns present in the SRM, leading to its lossy compression. This approach was built to detect translations, reflections and coordinates swap in voxel space. Therefore, it is particularly well suited for those scanners where some of the rotational symmetries are broken, e.g. small animal scanner where the modules are arranged in a polygonal ring made of few elements, and dual head planar PET systems. In order to validate this approach, the SSA is applied to the SRM of a preclinical scanner (the IRIS PET/CT). The data acquired by the scanner were reconstructed with a dedicated maximum likelihood estimation maximization algorithm with both the uncompressed and the compressed SRMs. The results achieved show that the information lost due to the SSA compression is negligible. Compression factors up to 52 when using the SSA together with manually inserted symmetries and up to 204 when using the SSA alone, can be obtained for the IRIS SRM. These results come without significant differences in the values and in the main quality metrics of the reconstructed images, i.e. spatial resolution and noise. Although the compression factors depend on the system considered, the SSA is applicable to any SRM and therefore it can be considered a general tool to reduce the footprint of a pre-computed SRM.

X-Ray Scatter Correction on Soft Tissue Images for Portable Cone Beam CT.

Soft tissue images from portable cone beam computed tomography (CBCT) scanners can be used for diagnosis and detection of tumor, cancer, intracerebral hemorrhage, and so forth. Due to large field of view, X-ray scattering which is the main cause of artifacts degrades image quality, such as cupping artifacts, CT number inaccuracy, and low contrast, especially on soft tissue images. In this work, we propose the X-ray scatter correction method for improving soft tissue images. The X-ray scatter correction scheme to estimate X-ray scatter signals is based on the deconvolution technique using the maximum likelihood estimation maximization (MLEM) method. The scatter kernels are obtained by simulating the PMMA sheet on the Monte Carlo simulation (MCS) software. In the experiment, we used the QRM phantom to quantitatively compare with fan-beam CT (FBCT) data in terms of CT number values, contrast to noise ratio, cupping artifacts, and low contrast detectability. Moreover, the PH3 angiography phantom was also used to mimic human soft tissues in the brain. The reconstructed images with our proposed scatter correction show significant improvement on image quality. Thus the proposed scatter correction technique has high potential to detect soft tissues in the brain.

MO‐F‐CAMPUS‐J‐05: Verification for Prompt Gamma Ray Imaging During Proton Boron Fusion Therapy

Purpose:The purpose of this study is to verify the acquisition of the three dimensional single photon emission computed tomography (SPECT) image using prompt gamma ray originated from proton boron fusion therapy (PBFT).Methods:The real‐time imaging system during the PBFT was simulated to acquire the tomographic image of the prompt gamma ray from treated tumor by the proton boron reaction, using Monte Carlo simulation (MCNPX). We acquired percentage depth dose (PDD) of the proton beam in the water phantom including the boron uptake region (BUR), the energy spectra of the prompt gamma ray and tomographic image which can show treated tumor regions. The prompt gamma ray image was reconstructed using maximum likelihood estimation maximization (MLEM) reconstruction algorithm with 64 projections. In addition, in order to evaluate the reconstructed image, the image profiles between BURs were extracted from the image.Results:In the PDD results, the Bragg‐peak was amplified definitely when the proton's maximum dose level was located at the BUR. This amplification is based on the generation of alpha particles. In addition, the prompt gamma ray peak of 719 keV was observed from the energy spectrum. Through the previous process, the tomographic image of prompt gamma ray from the BUR was reconstructed. The line profile was extracted from image including BURs, and it shows high signal‐to‐noise ratio.Conclusion:We confirmed that the real‐time prompt gamma ray image during the PBFT was successfully deducted, and results of quantitative image analysis show good agreement with the original pattern of the BUR. This study first verified the imaging capability of the PBFT‐SPECT system. In conclusion, the PBFT‐SPECT system can realize the treated tumor monitoring during the PBFT.

Acceleration of Fully 3D Monte Carlo Based System Matrix Computation for Image Reconstruction in Small Animal SPECT

It has already been proved that Fully Three Dimensional Monte Carlo (F3DMC) is a robust image reconstruction algorithm that can be applied in Single Photon Emission Computed Tomgraphy (SPECT) and small animal Positron Emission Tomography (PET). F3DMC has still not yet been validated on real data in small animal SPECT application. The advantage of such image reconstruction technique is that all the physical processes occuring within the detector and its geometrical parameters can be precisely modelled within the system matrix thanks to powerful Monte Carlo Simulation toolkit. Once the system matrix is computed, it can be integrated within an iterative reconstruction algorithm such as Maximum Likelihood Estimation Maximization (MLEM) in order to resolve the inverse image reconstruction problem. However, such reconstruction technique is penalized by the huge time consumption required for the computation of the system matrix since the accuracy of this latter requires the simulation of large number of photons tracks from the imaged subject to the detector. In this study, we proposed two main solutions to tackle the problem of time consumption. The first has already been proposed in anterior works and consists in parallelizing the Monte Carlo simulations performed with the Geant4 toolkit on a Computing Grid (CG) and the second suggests to apply a Forced Detection (FD) technique in order to accelerate the convergence of the system matrix elements. Results show that an accelerated version of a F3DMC technique is feasible in a reasonable delay and leads to reconstructed images with good spatial resolution and a good capability of restoring relative quantification. Hence, it has been proven that F3DMC is an applicable reconstruction technique in small animal SPECT.

Motion correction of PET brain images through deconvolution: I. Theoretical development and analysis in software simulations

Image quality is significantly degraded even by small amounts of patient motion in very high-resolution PET scanners. Existing correction methods that use known patient motion obtained from tracking devices either require multi-frame acquisitions, detailed knowledge of the scanner, or specialized reconstruction algorithms. A deconvolution algorithm has been developed that alleviates these drawbacks by using the reconstructed image to estimate the original non-blurred image using maximum likelihood estimation maximization (MLEM) techniques. A high-resolution digital phantom was created by shape-based interpolation of the digital Hoffman brain phantom. Three different sets of 20 movements were applied to the phantom. For each frame of the motion, sinograms with attenuation and three levels of noise were simulated and then reconstructed using filtered backprojection. The average of the 20 frames was considered the motion blurred image, which was restored with the deconvolution algorithm. After correction, contrast increased from a mean of 2.0, 1.8 and 1.4 in the motion blurred images, for the three increasing amounts of movement, to a mean of 2.5, 2.4 and 2.2. Mean error was reduced by an average of 55% with motion correction. In conclusion, deconvolution can be used for correction of motion blur when subject motion is known.

Reconstruction methods for quantitative brain SPECT

The effects of various reconstruction algorithms and compensation techniques for brain SPECT (single-photon-emission computed tomography) reconstruction are investigated. The conventional FB (filtered backprojection) and iterative ML-EM (maximum-likelihood estimation-maximization) algorithms are used. Projection data were acquired via dual energy windows for use in scatter compensation. Assuming uniform attenuation, the Chang attenuation compensation algorithm is applied. Restoration filters are used with the FB algorithms to compensate for the average effects of collimator and scatter response functions. Two-dimensional (2-D) and three-dimensional (3-D) compensation methods are compared. Compensation for the one-dimensional spatially variant collimator response, using an iterative reconstruction method, is shown. The results show that conventional reconstruction and compensation methods provide substantial improvement in reconstructed image quality and quantitative accuracy compared with no compensation. 3-D processing gives better results than 2-D processing. Iterative reconstruction provided more accurate quantitation with a substantial increase in computational cost.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>