Feasibility study of a point-of-care positron emission tomography system with interactive imaging capability

Abstract

Purpose: We investigated the feasibility of a novel positron emission tomography (PET) system that provides near real-time feedback to an operator who can interactively scan a patient to optimize image quality. The system should be compact and mobile to support point-of-care (POC) molecular imaging applications. In this study, we present the key technologies required and discuss the potential benefits of such new capability.

Methods: The core of this novel PET technology includes trackable PET detectors and a fully three-dimensional, fast image reconstruction engine implemented on multiple graphics processing units (GPUs) to support dynamically changing geometry by calculating the system matrix on-the-fly using a tube-of-response approach. With near real-time image reconstruction capability, a POC-PET system may comprise a maneuverable front PET detector and a second detector panel which can be stationary or moved synchronously with the front detector such that both panels face the region-of-interest (ROI) with the detector trajectory contoured around a patient's body. We built a proof-of-concept prototype using two planar detectors each consisting of a photomultiplier tube (PMT) optically coupled to an array of 48 × 48 lutetium-yttrium oxyorthosilicate (LYSO) crystals (1.0 × 1.0 × 10.0 mm³ each). Only 38 × 38 crystals in each arrays can be clearly re-solved and used for coincidence detection. One detector was mounted to a robotic arm which can position it at arbitrary locations, and the other detector was mounted on a rotational stage. A cylindrical phantom (102 mm in diameter, 150 mm long) with nine spherical lesions (8:1 tumor-to-background activity concentration ratio) was imaged from 27 sampling angles. List-mode events were reconstructed to form images without or with time-of-flight (TOF) information. We conducted two Monte Carlo simulations using two POC-PET systems. The first one uses the same phantom and detector setup as our experiment, with the detector coincidence re-solving time (CRT) ranging from 100 to 700 ps full-width-at-half-maximum (FWHM). The second study simulates a body-size phantom (316 × 228 × 160 mm³ ) imaged by a larger POC-PET system that has 4 × 6 modules (32 × 32 LYSO crystals/module, four in axial and six in transaxial directions) in the front panel and 3 × 8 modules (16 × 16 LYSO crystals/module, three in axial and eight in transaxial directions) in the back panel. We also evaluated an interactive scanning strategy by progressively increasing the number of data sets used for image reconstruction. The updated images were analyzed based on the number of data sets and the detector CRT.

Results: The proof-of-concept prototype re-solves most of the spherical lesions despite a limited number of coincidence events and incomplete sampling. TOF information reduces artifacts in the reconstructed images. Systems with better timing resolution exhibit improved image quality and reduced artifacts. We observed a reconstruction speed of 0.96 × 10⁶ events/s/iteration for 600 × 600 × 224 voxel rectilinear space using four GPUs. A POC-PET system with significantly higher sensitivity can interactively image a body-size object from four angles in less than 7 min.

Conclusions: We have developed GPU-based fast image reconstruction capability to support a PET system with arbitrary and dynamically changing geometry. Using TOF PET detectors, we demonstrated the feasibility of a PET system that can provide timely visual feedback to an operator who can scan a patient interactively to support POC imaging applications.

Keywords: GPU reconstruction; interactive PET Imaging; point-of- care application; time-of-flight PET.

Figure 1

(a) Structure of the fully 3‐D GPU‐based list‐mode image reconstruction workflow. The PET detector trajectories are modeled as a series of static positions during the scan. Current system geometry is computed from the initial detector geometry and the current locations of detector panels. Sensitivity image of current system geometry is computed and added to the previous sensitivity image for image reconstruction using all list‐mode events acquired so far. Using the arrival time of each event and the time frame information of each panel location, the crystal positions of each coincidence event are calculated and sent to GPU memory for list‐mode image reconstruction. The image reconstruction will stop after a pre‐defined number of iterations (N = 10 in this work). (b) Flowchart showing list‐mode GPU reconstruction kernel.

Figure 2

(a) Schematic drawing of the prototype POC‐PET system; (b) Illustration of the sizes and distribution of the spherical tumor inserts in the cylindrical phantom. The 6.3 mm tumor highlighted in a blue box is the one that was accidentally filled with air bubbles. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3

The DAQ architecture of the prototype system. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4

(a) An illustration showing the locations of the two PET detectors and the sampling pattern used to image the cylindrical phantom. The blue block denotes the back detector while the green block denotes the front detector. The red circle represents the cylindrical phantom. The coincidence events were acquired at 27 angles by placing both the front and back detectors at seven different positions, as illustrated in dashed line (of all colors); (b) The distribution of nine spherical tumors simulated in the Monte Carlo study. The tumors were each placed at the center slice of the cylindrical phantom. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5

Scanning strategy when imaging a body‐sized phantom using a large POC‐PET system. (a) scanning angle 1; (b) scanning angle 2; (c) scanning angle 3; (d) scanning angle 4. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 6

(a) Flood image of the front detector; (b) Flood image of the back detector; (c) Timing spectrum of the central crystals (4‐by‐4) of the front detector against the entire back detector. The measured CRT was around 740 ps FWHM. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7

(a) Illustration of the sizes and distribution of the spherical tumor inserts in the cylindrical phantom; (b) Sensitivity image of the prototype POC‐PET system; (c) Image of a cylindrical phantom containing nine spherical tumors with 8:1 tumor‐to‐background radioactivity concentration, reconstructed without TOF information; (d) Image reconstructed with TOF information. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 8

Images of the cylindrical phantom measured by the simulated POC‐PET systems. (a) The distribution of nine spherical tumors in the phantom simulated in the Monte Carlo study Images were reconstructed (b) without the TOF information or with a detector CRT of (c) 700 ps FWHM; (d) 500 ps FWHM; (e) 300 ps FWHM; and (f) 100 ps FWHM. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 9

Images reconstructed without TOF information using different numbers of groups of data.

Figure 10

Images reconstructed using different numbers of groups of data when the detector timing resolution was 300 ps FWHM.

Figure 11

Images reconstructed using list‐mode data from different numbers of sampling angles. (a) Sampling angle 1; (b) Sampling angles 1 + 2; (c) Sampling angles 1 + 2 +  3; (d) Sampling angles 1 + 2 +  3 + 4.

Figure 12

The average CRC as a function of tumor sizes estimated from the four images in Fig. 11. [Color figure can be viewed at wileyonlinelibrary.com]