PT  - JOURNAL ARTICLE
AU  - Julien Rossignol
AU  - Réjean Fontaine
AU  - Rosana Martinez Turtos
AU  - Stefan Gundacker
AU  - Etiennette Auffray
AU  - plecoq LECOQ
AU  - Yves Bérubé-Lauzière
TI  - Time-of-Flight Computed Tomography a Proof of Principle Study
DP  - 2019 May 01
TA  - Journal of Nuclear Medicine
PG  - 47--47
VI  - 60
IP  - supplement 1
4099  - http://jnm.snmjournals.org/content/60/supplement_1/47.short
4100  - http://jnm.snmjournals.org/content/60/supplement_1/47.full
SO  - J Nucl Med2019 May 01; 60
AB  - 47Objectives: The introduction of time-of-flight (TOF) in medical imaging such as in positron emission tomography (PET) or diffuse optical tomography lead to improvements on image quality. In computed tomography (CT), the signal-to-noise ratio (SNR) can be improved when photon counting along with energy measurements is used. The aim of this study is a) to assess the feasibility of using TOF in CT to discriminate between ballistic and scattered photons based on their travel path in order to reduce the effect of scatter noise and b) identify detector requirements for a first prototype development and to assess if this new approach could be implemented along with TOF-PET in a bimodality scanner using only one detector system. Methods: Herein, a complete flat panel cone-beam scanner is simulated using Geant4 Application for Tomographic Emission (GATE). The simulated scanner comprises a monochromatic X-ray source, irradiating with a cone angle of 16°, plexiglass phantoms of 50, 100, 150 and 200 mm thickness used to generate increasing levels of scatter radiation and a flat panel made of a 256 x 256 array of 1 x 1 x 1 mm3 detectors. A digital scatter rejection algorithm (DSR) compares the TOF of each individual photon with the expected TOF for a direct trajectory and removes the photons who failed to arrive in time. The DSR performance is evaluated first with single projection acquisitions for different configurations, then a full 360° acquisition of a water cylinder with two cylindrical bone inserts is realised to measure the effect of DSR on image quality. In this case, a perfect time resolution is used and 100 keV photons are emitted in a 28° cone towards the detection system comprising a 256 x 256 array of 2 x 2 mm2 pixels. The image was reconstructed using the implementation of the algorithm of Feldkamp, David and Kress (FDK) from the Reconstruction Toolkit (RTK). The experimental setup comprises a pulsed X-ray source N5084 from Hamamatsu and a 3 x 3 mm2Silicon Photomultiplier with a 200 µm thick LYSO:Ce crystal. Two acquisitions, with and without a beam-blocker used to force scattering of photons on the X-ray enclosure, were made to highlight statistical differences between the TOF of ballistic and scattered photons. Results: After traveling through 200 mm of plexiglass, 85% and 40% of all scattered radiation arrives at least 100 and 500 ps later than expected with a median TOF difference of 400 ps. This value decreases to 82 ps with a 50 mm phantom. With photon energies lower than 60 keV and scanner diameter smaller than 500 mm, the time window shrinks and a total time resolution of 10 ps or less is needed to correctly identify almost all primary and scattered photons since these parameters decrease the mean scattering angle and travel path. In a TOF-CT model with perfect time resolution and 120 keV photons, the scatter-to-primary ratio (SPR) can be reduced to less than 3%, even with an initial SPR as high as 300%. With 10 and 100 ps FWHM total resolution, SPR is reduced from 225% to respectively 8% and 40%. In the reconstructed 360° acquisition (Figure 1), when applying the DSR algorithm the SPR decreases from 300% to 4%, the CNR doubles and the cup artifacts are greatly reduced. When applying the DSR a similar CNR to the uncorrected image could be obtained with a 4 fold dose reduction. The experimental setup shows a shift between the TOF of scattered photons and ballistic photons of 389 ps, fitting the expected time for photons to circle the beam-blocker by scattering on the X-ray enclosure. Conclusions: Our results confirms the possibility in both simulation and experiment to measure statistical differences between scattered and ballistic photons. Reducing the time resolution to 10 ps is required in order to almost completely remove the effect of scatter noise, which fits with the goals for TOF-PET. This improvement could lead to a reduction of dose such as in our simulated setup where a 75% dose is observed. As the timing requirements are similar, TOF-CT could be implemented using the same detection system as TOF-PET.