Shoko Kinouchi

Development of a small prototype for a proof-of-concept of OpenPET imaging

The OpenPET geometry is our new idea to visualize a physically opened space between two detector rings. In this paper, we developed the first small prototype to show a proof-of-concept of OpenPET imaging. Two detector rings of 110 mm diameter and 42 mm axial length were placed with a gap of 42 mm. The basic imaging performance was confirmed through phantom studies; the open imaging was realized at the cost of slight loss of axial resolution and 24% loss of sensitivity. For a proof-of-concept of PET image-guided radiation therapy, we carried out the in-beam tests with (11)C radioactive beam irradiation in the heavy ion medical accelerator in Chiba to visualize in situ distribution of primary particles stopped in a phantom. We showed that PET images corresponding to dose distribution were obtained. For an initial proof-of-concept of real-time multimodal imaging, we measured a tumor-inoculated mouse with (18)F-FDG, and an optical image of the mouse body surface was taken during the PET measurement by inserting a digital camera in the ring gap. We confirmed that the tumor in the gap was clearly visualized. The result also showed the extension effect of an axial field-of-view (FOV); a large axial FOV of 126 mm was obtained with the detectors that originally covered only an 84 mm axial FOV. In conclusion, our initial imaging studies showed promising performance of the OpenPET.

A single-ring OpenPET enabling PET imaging during radiotherapy

We develop an OpenPET system which can provide an accessible open space to the patient during PET scanning. Our first-generation OpenPET geometry which we called dual-ring OpenPET consisted of two separated detector rings and it could extend its axial field of view (FOV) therefore enabling imaging the gap region in addition to the in-ring region. However, applications such as dose verification by in-beam PET measurement during particle therapy and real-time tumor tracking by PET require sensitivity focused onto the gap rather than on the wide FOV. In this paper, we propose a second-generation OpenPET geometry, single-ring OpenPET, which can provide an accessible and observable open space with higher sensitivity and a reduced number of detectors than the earlier one. The proposed geometry has a cylinder shape cut at a slant angle, in which the shape of each cut end becomes an ellipse. We provided a theoretical analysis for sensitivity of the proposed geometry, compared with the dual-ring OpenPET and a geometry where the conventional PET was positioned at a slant angle against the patient bed to form an accessible open space, which we called a slant PET. The central sensitivity depends on the solid angle of these geometries. As a result, we found that the single-ring OpenPET has a sensitivity 1.2 times higher than the dual-ring OpenPET and 1.3 times higher than the slant PET when designed for a 600 mm bed width with 300 mm accessible open space and about 200 detector blocks, each with a front area of 2500 mm². In addition, numerical simulation was carried out to show the imaging property of the proposed geometry realized with the ellipsoidal rings and these results indicate that the depth-of-interaction detector can provide uniform resolution even when the detectors are arranged in an ellipsoidal ring.

Real-Time Imaging System for the OpenPET

The OpenPET and its real-time imaging capability have great potential for real-time tumor tracking in medical procedures such as biopsy and radiation therapy. For the real-time imaging system, we intend to use the one-pass list-mode dynamic row-action maximum likelihood algorithm (DRAMA) and implement it using general-purpose computing on graphics processing units (GPGPU) techniques. However, it is difficult to make consistent reconstructions in real-time because the amount of list-mode data acquired in PET scans may be large depending on the level of radioactivity, and the reconstruction speed depends on the amount of the list-mode data. In this study, we developed a system to control the data used in the reconstruction step while retaining quantitative performance. In the proposed system, the data transfer control system limits the event counts to be used in the reconstruction step according to the reconstruction speed, and the reconstructed images are properly intensified by using the ratio of the used counts to the total counts. We implemented the system on a small OpenPET prototype system and evaluated the performance in terms of the real-time tracking ability by displaying reconstructed images in which the intensity was compensated. The intensity of the displayed images correlated properly with the original count rate and a frame rate of 2 frames per second was achieved with average delay time of 2.1 s.

Development of a small single-ring OpenPET prototype with a novel transformable architecture.

The single-ring OpenPET (SROP), for which the detector arrangement has a cylinder shape cut by two parallel planes at a slant angle to form an open space, is our original proposal for in-beam PET. In this study, we developed a small prototype of an axial-shift type SROP (AS-SROP) with a novel transformable architecture for a proof-of-concept. In the AS-SROP, detectors originally forming a cylindrical PET are axially shifted little by little. We designed the small AS-SROP prototype for 4-layer depth-of-interaction detectors arranged in a ring diameter of 250 mm. The prototype had two modes: open and closed. The open mode formed the SROP with the open space of 139 mm and the closed mode formed a conventional cylindrical PET. The detectors were simultaneously moved by a rotation handle allowing them to be transformed between the two modes. We evaluated the basic performance of the developed prototype and carried out in-beam imaging tests in the HIMAC using (11)C radioactive beam irradiation. As a result, we found the open mode enabled in-beam PET imaging at a slight cost of imaging performance; the spatial resolution and sensitivity were 2.6 mm and 5.1% for the open mode and 2.1 mm and 7.3% for the closed mode. We concluded that the AS-SROP can minimize the decrease of resolution and sensitivity, for example, by transforming into the closed mode immediately after the irradiation while maintaining the open space only for the in-beam PET measurement.

Compartmental analysis of washout effect in rat brain: in-beam OpenPET measurement using a 11C beam

In-beam positron emission tomography (PET) is expected to enable visualization of a dose verification using positron emitters (β+ decay). For accurate dose verification, correction of the washout of the positron emitters should be made. In addition, the quantitative washout rate has a potential usefulness as a diagnostic index, but modeling for this has not been studied yet. In this paper, therefore, we applied compartment analyses to in-beam PET data acquired by our small OpenPET prototype, which has a physically opened field-of-view (FOV) between two detector rings. A rat brain was located at the FOV and was irradiated by a (11)C beam. Time activity curves of the irradiated field were measured immediately after the irradiations, and the washout rate was obtained based on two models: the two-washout model (medium decay, k2m; slow decay, k2s) developed in a study of rabbit irradiation; and the two-compartment model used in nuclear medicine, where efflux from tissue to blood (k2), influx (k3) and efflux (k4) from the first to second compartments in tissue were evaluated. The observed k2m and k2s were 0.34 and 0.005 min(-1), respectively, which was consistent with the rabbit study. Also k2m was close to the washout rate in cerebral blood flow (CBF) measurements by dynamic PET with (15)O-water, while, k2, k3, and k4 were 0.16, 0.15 and 0.007 min(-1). Our present work suggested the dynamics of (11)C might be relevant to CBF or permeability of a molecule containing (11)C atoms might be regulated by a transporter because the k2 was relatively low compared with a simple diffusion tracer.

GPU-Based PET Image Reconstruction Using an Accurate Geometrical System Model

In positron emission tomography (PET), 3D iterative image reconstruction methods have a huge computational burden. In this paper, we developed a list-mode image reconstruction method using graphics processing units (GPUs). Efficiency of acceleration for GPU implementation largely depends on the method chosen, where a reduced number of conditional statements and a reduced memory size are required. On the other hand, accurate system models are required to improve the quality of reconstructed images. Various accurate system models for conventional CPU implementation have been proposed, but these models basically require many conditional statements and huge memory size. Therefore, we developed a new system model which matches GPU implementation better. In this model, the detector response functions, which vary depending on each line of response (LOR), are pre-computed in CPUs and modeled by sixth-order polynomial functions in order to reduce the memory size occupied in GPUs. Each element of a system matrix is obtained on-the-fly in GPUs by calculating the distance between an LOR and a voxel. Therefore the developed system model enables efficient GPU implementation of the accurate system modeling with a reduced number of conditional statements and a reduced memory size. We applied the developed method to a small OpenPET prototype, in which 4-layered depth-of-interaction (DOI) detectors were used. For image reconstruction, we used the dynamic row-action maximum likelihood algorithm (DRAMA). Compared with a conventional model for GPU implementation, in which DRFs are given as a Gaussian function of fixed width, we saw no remarkable difference for DOI data, but for non-DOI data the proposed model outperformed the conventional at the peripheral region of the field-of-view. The proposed model had almost the same calculation time as the conventional model did. For further acceleration, we tried parallel GPU implementation, and we obtained 3.8-fold acceleration by using 4 GPUs.

A small prototype of a single-ring OpenPET

The OpenPET geometry is our original idea to visualize a physically opened space even with a full ring geometry. One of our targets is in-beam PET, which is a method for in situ monitoring of charged particle therapy. In our initial idea, the OpenPET had a physically opened field-of-view (FOV) between two detector rings separated by a gap. Originally, the OpenPET was proposed to provide a stress-less brain imaging device. For a dedicated in-beam PET scanner, this dual-ring OpenPET is a good candidate, but it is not necessarily the most efficient geometry because it has a wide FOV (i.e., a gap FOV plus two in-ring FOVs) while only a limited FOV around the irradiation field is required in actual use of in-beam PET. At the last conference, therefore, we proposed a single-ring OpenPET (SROP) dedicated for in-beam PET as our 2nd generation geometry. The detector ring of the SROP geometry was the cylinder both ends of which were cut by parallel aslant planes. In this paper, we developed a small prototype of the SROP for a proof-of-concept. It consisted of 2 ellipse-shaped detector rings, each of which had 16 detectors. Each ellipse-shaped detector ring had a major axis of 281.6 mm and a minor axis of 207.5 mm. The rings were slanted by 45 deg and staggered to obtain an open space of 74.5 mm width. We carried out initial in-beam imaging tests in the Heavy Ion Medical Accelerator in Chiba (HIMAC) using a 11C beam as well as a 12C beam. PET measurement started at the beginning of the irradiation, and continued for 20 min after the irradiation. For about 3Gy irradiation, a 6 mm range difference was clearly detected with the 11C beam irradiation. Our initial imaging studies showed promising performance of the SROP prototype.

Simulation design of a single-ring OpenPET for in-beam PET

One of the challenging applications of PET is for in-beam PET, which is an in situ monitoring method for charged particle therapy. For this purpose, we have previously proposed an open-type PET scanner, OpenPET. The original OpenPET has a physically opened field-of-view (FOV) between two detector rings which irradiation beams pass through. This dual-ring OpenPET has a wide axial FOV including the gap. Therefore this geometry is not necessarily the most efficient when it is applied to in-beam PET in which only a limited FOV around the irradiation field is required. In this paper, we proposed new single-ring OpenPET geometry as more efficient geometry dedicated to in-beam PET. The detector ring of the proposed geometry is a cylinder both ends of which are cut by parallel aslant planes. The proposed geometry can be made compact so that the beam port can be placed close to the patient.

GPU implementation of list-mode DRAMA for real-time OpenPET image reconstruction

The OpenPET, which has a physically opened space between two detector rings, is our new geometry to enable PET imaging during radiation therapy. Especially, tracking a moving target such as a tumor in the lung will become possible if the real-time imaging system is realized. In this paper, we developed a list-mode image reconstruction method using general-purpose computing on graphics processing units (GPGPUs) toward real-time image reconstruction. We used the dynamic row-action maximum likelihood algorithm (DRAMA). For GPU implementation, efficiency of acceleration depends on implementation methods; reduced conditional statements and efficient memory accesses are required. On the other hand, accurate system model is required to improve quality of reconstructed images. Therefore, we developed a new system model which was suited for the GPU implementation. In the new system model, the detector response functions (DRFs), which were calculated analytically to represent the probability distribution of each LOR, were modeled by sixth-order polynomial functions. The system model enabled us to calculate each element of the system matrix with reduced conditional statements. We applied the developed method to a small OpenPET prototype, which was developed for a proof-of-concept. We compared the proposed system model to the sub-LOR model, a geometrically-defined accurate system model which we had previously proposed. The difference between the reconstructed images with the new system model using GPU and the sub-LOR model using CPU was very small. Our new system model on GPU was 46.3 times faster than the sub-LOR model on CPU.

Total variation minimization for in-beam PET image reconstruction

One of the challenging applications of positron emission tomography (PET) is in-beam PET, which is an in situ dose monitoring method for charged particle therapy. It is known that the activity of positron emitters produced through fragmentation reaction is generally low. Image reconstruction from low count data would require alternative algorithms to suppress noise in images, such as the maximum a posteriori (MAP) reconstruction algorithm, which is often used to improve an image reconstruction problem by introducing penalty functions. In particular, the total variation (TV) norm has been proposed as a penalty function to suppress noise while preserving edges. In this study, we applied a MAP-TV image reconstruction algorithm to in-beam PET imaging, and we evaluated the effect of TV constraint for the in-beam PET image in terms of detection performance of distal edge a of positron emitter distribution along beam irradiation. We applied the one-step-Iate algorithm combined with the TV norm to a new small prototype of the single-ring OpenPET (SROP), which is our second generation OpenPET geometry. The small SROP prototype consisted of two oval detector rings which were slanted by 45 degree and stacked. We carried out initial in-beam experiments in the Heavy Ion Medical Accelerator in Chiba (HIMAC). In the experiment, we used a 12C beam. The target was a rectangular parallelepiped phantom (40 × 40 mm2 and 100 mm long) made of polymethyl methacrylate (PMMA). We calculated averaged peak position of profiles obtained in reconstructed images. In the experimental results, reconstructed images became smoother and less noisy with stronger constraint. It was shown that the MAP-TV algorithm enables smoothness and less noisy in reconstructed images even from the low count.