The Phase I Proton CT Scanner and Test Beam Results at LLUMC, invited

INTRODUCTION

Loma Linda University, in collaboration with Northern Illinois University and UC Santa Cruz, has built and tested a pre-clinical prototype proton CT (pCT) scanner based on a design presented previously [1]. The system is designed for scanning of head-sized phantoms and small animals. This report presents some of our preliminary test results for the scanner installed on the research beam line at Loma Linda University Medical Center (LLUMC).

METHODS

The prototype pCT scanner consists of two major components: the silicon tracker for tracking the paths of individual protons and the cesium iodide crystal calorimeter for measuring the residual energy of individual protons after passing through the image object. A schematic representation of the system is shown in Fig. 1.

Fig. 1.

Schematic representation of the prototype pCT scanner.

The tracker is comprised of 16 silicon strip detectors arranged into 4 planes each with a sensitive area of 8.95 × 17.9 cm². The silicon strip detectors have a strip pitch of 228 um. To achieve a continuous sensitive area without any gaps, the detectors in a plane are overlapped slightly. This is visible in Fig. 1. The planes are separated into two “telescopes” positioned before and after the object to measure the position and direction of individual protons as they enter and exit the object.

The calorimeter consists of 18 thallium doped cesium iodide (CsI(T1)) crystals arranged to form a 3×6 rectangular matrix encompassing the sensitive area of the tracker [2]. Individual protons are stopped in the calorimeter and the energy is converted to light by scintillation. The light is collected by a photodiode paired to each crystal and a conversion to a digital value is performed with an ADC.

An FPGA-based data acquisition system processes and records both the tracking and energy information at a maximum rate of approximately 10⁵ protons per second.

Calorimeter Calibration

The resolution of the relative stopping power in pCT image reconstruction is largely dictated by the accuracy of the measurement of the exit energy. A calibration method has been developed to convert the calorimeter signal into a water-equivalent path length (WEPL) to be used in the pCT reconstruction. This method is designed to be accurate and simple enough to be performed daily. The method involves varying the thickness of polystyrene plates of known relative stopping power in front of the calorimeter and measuring the response to a mono-energetic proton beam, after passing through the plates. Fig. 2 shows a photograph of the scanner during this calibration. The mean calorimeter response is then recorded for several different thicknesses of plate to establish a relationship between WEPL and response.

Fig. 2.

pCT scanner during calorimeter calibration. Two polystyrene plates are visible in the center of the photograph.

Relative Stopping Power Verification

To verify proper calibration of the calorimeter and check the values of relative stopping power (RSP) obtained with the pCT scanner independent of the image reconstruction algorithms, the RSP of a set of tissue equivalent plates was measured using the system. This was done using plates of uniform thickness and density, placed in front of the calorimeter and between the tracking telescopes. Each plate was exposed to approximately 100,000 protons with an energy of 100 MeV and a roughly parallel beam, and the histories and calorimeter responses were recorded for each. Using the WEPL calibration curve obtained previously, the calorimeter responses were converted to WEPL. Histories which crossed the plate with an angle greater than 1 degree were discarded. Finally, the RSP of the material was calculated as the mean WEPL divided by the physical thickness of the plate. The results were compared to RSP measurements obtained for the same plates using a standard water phantom depth-dose range shift measurement.

Imaging Studies

In the prototype pCT scanner, the beam line and detectors are fixed while the imaging subject is rotated. This is analogous to a gantry-based design, in which the subject remains fixed and the gantry and detectors rotate around it.

A number of different phantoms have been scanned with the pCT system. The first is the Lucy phantom (shown in Fig. 4), a 14 cm diameter polystyrene sphere with tissue equivalent inserts. The phantom was scanned using a 200 MeV cone beam. A 3D reconstruction of the phantom was produced based on 10⁸ proton histories. Another phantom scanned was a water phantom, constructed out of a 15 cm acrylic cylinder. This phantom was scanned with a 200 MeV cone beam and 8×10⁸ proton histories. Both phantoms were scanned with 90 projections in 4 degree steps. Results will be presented from each of these imaging studies.

Fig. 4.

The Lucy phantom used in imaging studies assembled with the Leksell frame adapter.

RESULTS

Relative Stopping Power Verification

The RSP for each of the tissue equivalent materials was measured with the method described above and compared to measurements obtained with a standard water phantom depth-dose range shift measurement. The results show very good agreement (better than 1%). It should also be noted that the measurements with the pCT scanner took significantly less time and delivered less dose to the phantom, by a factor of approximately 10⁵. The results are shown in Table I.

Table I. Comparison of RSP measurements obtained using the pCT scanner and using a water phantom depth-dose range shift method

Material	RSP	σ RSP
	Range shift method
Lung	0.267	0.005
Adipose	0.947	0.007
Muscle	1.032	0.008
Brain	1.062	0.007
Liver	1.076	0.005
Cortical Bone	1.599	0.007
	pCT scanner method
Lung	0.268	0.001
Adipose	0.943	0.002
Muscle	1.037	0.002
Brain	1.064	0.002
Liver	1.078	0.002
Cortical Bone	1.595	0.002

Reconstructed Images

The first phantom successfully scanned and reconstructed with the Phase I pCT scanner was the Lucy phantom. The phantom was reconstructed in 3D with a 16 × 16 × 8 cm³ reconstruction volume. The volume was divided into voxels of 0.625 × 0.625 × 2.5 mm³ size, forming a 256 × 256 × 32 element reconstruction volume. Fig. 5 shows an axial slice from this reconstruction. The gray values shown in the image represent RSP values from 0 to 2.

Fig. 5.

Axial slice from the reconstructed pCT scan of the Lucy phantom.

To check the absolute value of the reconstructed RSP values against a precisely known material, a water phantom was constructed specifically for the pCT scanner and filled with distilled water and degassed in a vacuum chamber. The phantom was scanned and reconstructed with the same parameters as the Lucy phantom. Fig. 6 shows an axial slice from this reconstruction.

Fig. 6.

Axial slice from the reconstructed pCT scan of the water phantom.

There are some visible ring artifacts in the water phantom reconstruction, which are systematic and can be corrected for. The artifacts are most likely due to variations in the signal from the individual calorimeter crystals, which need to be carefully calibrated relative to each other. Fig. 7 shows a central band profile across an axial slice.

Fig. 7.

Profile across the central band of an axial slice of the water phantom. The central dip artifact can be seen at x=8 cm and a larger ring artifact between 5 and 11 cm is noticeable. The 0.5 cm thick acrylic walls are seen as a slight rise at the edges of the profile.

The reconstructed RSP values for the water phantom are very close to the expected value of 1, differing by about 1%. A histogram of RSP values selected from a region of interest excluding the central artifact and acrylic walls is shown in Fig. 8. The mean value is 0.9952 with an RMS variation of 0.006. Improved calibration and corrections should be decrease the variation and systematic error. This work is ongoing.

Fig. 8.

Histogram of RSP values for the water phantom excluding the central artifact and acrylic walls.

Fig. 3.

Fitted second order polynomial from the WEPL calibration method.