Water-equivalent path length calibration of a prototype proton CT scanner.

PURPOSE: The authors present a calibration method for a prototype proton computed tomography (pCT) scanner. The accuracy of these measurements depends upon careful calibration of the energy detector used to measure the residual energy of the protons that passed through the object. METHODS: A prototype pCT scanner with a cesium iodide (CsI(Tl)) crystal calorimeter was calibrated by measuring the calorimeter response for protons of 200 and 100 MeV initial energies undergoing degradation in polystyrene plates of known thickness and relative stopping power (RSP) with respect to water. Calibration curves for the two proton energies were obtained by fitting a second-degree polynomial to the water-equivalent path length versus calorimeter response data. Using the 100 MeV calibration curve, the RSP values for a variety of tissue-equivalent materials were measured and compared to values obtained from a standard depth-dose range shift measurement using a water-tank. A cylindrical water phantom was scanned with 200 MeV protons and its RSP distribution was reconstructed using the 200 MeV calibration. RESULTS: It is shown that this calibration method produces measured RSP values of various tissue-equivalent materials that agree to within 0.5% of values obtained using an established water-tank method. The mean RSP value of the water phantom reconstruction was found to be 0.995 ± 0.006. CONCLUSIONS: The method presented provides a simple and reliable procedure for calibration of a pCT scanner.

Experimental realization of dynamic fluence field optimization for proton computed tomography.

Proton computed tomography (pCT) has high accuracy and dose efficiency in producing spatial maps of the relative stopping power (RSP) required for treatment planning in proton therapy. With fluence-modulated pCT (FMpCT), prescribed noise distributions can be achieved, which allows to decrease imaging dose by employing object-specific dynamically modulated fluence during the acquisition. For FMpCT acquisitions we divide the image into region-of-interest (ROI) and non-ROI volumes. In proton therapy, the ROI volume would encompass all treatment beams. An optimization algorithm then calculates dynamically modulated fluence that achieves low prescribed noise inside the ROI and high prescribed noise elsewhere. It also produces a planned noise distribution, which is the expected noise map for that fluence, as calculated with a Monte Carlo simulation. The optimized fluence can be achieved by acquiring pCT images with grids of intensity modulated pencil beams. In this work, we interfaced the control system of a clinical proton beam line to deliver the optimized fluence. Using three phantoms we acquired images with uniform fluence, with a constant noise prescription, and with an FMpCT task. Image noise distributions as well as fluence maps were compared to the corresponding planned distributions as well as to the prescription. Furthermore, we propose a correction method that removes image artifacts stemming from the acquisition with pencil beams having a spatially varying energy distribution that is not seen in clinical operation. RSP accuracy of FMpCT scans was compared to uniform scans and was found to be comparable to standard pCT scans. While we identified technical improvements for future experimental acquisitions, in particular related to an unexpected pencil beam size reduction and a misalignment of the fluence pattern, agreement with the planned noise was satisfactory and we conclude that FMpCT optimized for specific image noise prescriptions is experimentally feasible.

Leakage and scatter radiation from a double scattering based proton beamline.

Proton beams offer several advantages over conventional radiation techniques for treating cancer and other diseases. These advantages might be negated if the leakage and scatter radiation from the beamline and patient are too large. Although the leakage and scatter radiation for the double scattering proton beamlines at the Loma Linda University Proton Treatment Facility were measured during the acceptance testing that occurred in the early 1990s, recent discussions in the radiotherapy community have prompted a reinvestigation of this contribution to the dose equivalent a patient receives. The dose and dose equivalent delivered to a large phantom patient outside a primary proton field were determined using five methods: simulations using Monte Carlo calculations, measurements with silver halide film, measurements with ionization chambers, measurements with rem meters, and measurements with CR-39 plastic nuclear track detectors. The Monte Carlo dose distribution was calculated in a coronal plane through the simulated patient that coincided with the central axis of the beam. Measurements with the ionization chambers, rem meters, and plastic nuclear track detectors were made at multiple locations within the same coronal plane. Measurements with the film were done in a plane perpendicular to the central axis of the beam and coincident with the surface of the phantom patient. In general, agreement between the five methods was good, but there were some differences. Measurements and simulations also tended to be in agreement with the original acceptance testing measurements and results from similar facilities published in the literature. Simulations illustrated that most of the neutrons entering the patient are produced in the final patient-specific aperture and precollimator just upstream of the aperture, not in the scattering system. These new results confirm that the dose equivalents received by patients outside the primary proton field from primary particles that leak through the nozzle are below the accepted standards for x-ray and electron beams. The total dose equivalent outside of the field is similar to that received by patients undergoing treatments with intensity modulated x-ray therapy. At the center of a patient for a whole course of treatment, the dose equivalent is comparable to that delivered by a single whole-body XCT scan.

Calibration of a proton beam energy monitor.

Delivery of therapeutic proton beams requires an absolute energy accuracy of +/-0.64 to 0.27 MeV for patch fields and a relative energy accuracy of +/-0.10 to 0.25 MeV for tailoring the depth dose distribution using the energy stacking technique. Achromatic switchyard tunes, which lead to better stability of the beam incident onto the patient, unfortunately limit the ability of switchyard magnet tesla meters to verify the correct beam energy within the tolerances listed above. A new monitor to measure the proton energy before each pulse is transported through the switchyard has been installed into a proton synchrotron. The purpose of this monitor is to correct and/or inhibit beam delivery when the measured beam energy is outside of the tolerances for treatment. The monitor calculates the beam energy using data from two frequency and eight beam position monitors that measure the revolution frequency of the proton bunches and the effective offset of the orbit from the nominal radius of the synchrotron. The new energy monitor has been calibrated by measuring the range of the beam through water and comparing with published range-energy tables for various energies. A relationship between depth dose curves and range-energy tables was first determined using Monte Carlo simulations of particle transport and energy deposition. To reduce the uncertainties associated with typical scanning water phantoms, a new technique was devised in which the beam energy was scanned while fixed thickness water tanks were sandwiched between two fixed parallel plate ionization chambers. Using a multitude of tank sizes, several energies were tested to determine the nominal accelerator orbit radius. After calibration, the energy reported by the control system matched the energy derived by range measurements to better than 0.72 MeV for all nine energies tested between 40 and 255 MeV with an average difference of -0.33 MeV. A study of different combinations of revolution frequency and radial offsets to test the envelope of algorithm accuracy demonstrated a relative accuracy of +/-0.11 MeV for small energy changes between 126 and 250 MeV. These new measurements may serve as a data set for benchmarking range-energy relationships.

Application of fluence field modulation to proton computed tomography for proton therapy imaging.

This simulation study presents the application of fluence field modulated computed tomography, initially developed for x-ray CT, to proton computed tomography (pCT). By using pencil beam (PB) scanning, fluence modulated pCT (FMpCT) may achieve variable image quality in a pCT image and imaging dose reduction. Three virtual phantoms, a uniform cylinder and two patients, were studied using Monte Carlo simulations of an ideal list-mode pCT scanner. Regions of interest (ROI) were selected for high image quality and only PBs intercepting them preserved full fluence (FF). Image quality was investigated in terms of accuracy (mean) and noise (standard deviation) of the reconstructed proton relative stopping power compared to reference values. Dose calculation accuracy on FMpCT images was evaluated in terms of dose volume histograms (DVH), range difference (RD) for beam-eye-view (BEV) dose profiles and gamma evaluation. Pseudo FMpCT scans were created from broad beam experimental data acquired with a list-mode pCT prototype. FMpCT noise in ROIs was equivalent to FF images and accuracy better than -1.3%(-0.7%) by using 1% of FF for the cylinder (patients). Integral imaging dose reduction of 37% and 56% was achieved for the two patients for that level of modulation. Corresponding DVHs from proton dose calculation on FMpCT images agreed to those from reference images and 96% of BEV profiles had RD below 2 mm, compared to only 1% for uniform 1% of FF. Gamma pass rates (2%, 2 mm) were 98% for FMpCT while for uniform 1% of FF they were as low as 59%. Applying FMpCT to preliminary experimental data showed that low noise levels and accuracy could be preserved in a ROI, down to 30% modulation. We have shown, using both virtual and experimental pCT scans, that FMpCT is potentially feasible and may allow a means of imaging dose reduction for a pCT scanner operating in PB scanning mode. This may be of particular importance to proton therapy given the low integral dose found outside the target.

Particle beam radiation therapy in prostate cancer: is there an advantage?

Hadron therapy uses heavy particles to deliver therapeutic ionizing energy. Each particle's inherent attributes determine the pattern of energy deposited by its beam, expressed in macro (conformability to a three-dimensional target volume) and micro (radiobiologic properties) distributions. Mass and charge regulate the inherent properties; beam energy provides a controllable, variable characteristic. Generally, heavy charged particles provide superior macrodosimetric properties; heavy particles (charged or not) have microdosimetric characteristics that produce high linear energy transfer (LET). Neutron macrodosimetry is similar to that of photons. Protons and helium ions possess superior macrodosimetric properties, plus microdosimetric characteristics resulting in low LET, yielding beam characteristics that approach the ideal for clinical radiotherapy. Hadron therapy for prostate cancer has been limited by the availability of appropriate treatment facilities. Nonetheless, encouraging results have been obtained. Neutron therapy demonstrated improved overall survival in a multi-institutional randomized trial, and improved local disease control in a subsequent trial. Proton radiation forms the boost component of several conformal dose-escalation studies. A Loma Linda University study demonstrated low treatment-related morbidity despite a prostate dose of 75 CGE; late-morbidity data were superior to published reports from multi-field, conformal photon therapy. A Phase III dose-escalation study of protons for early prostate cancer is proceeding.

A beam intensity monitor for the Loma Linda cancer therapy proton accelerator.

A beam intensity monitor was tested in a 230-MeV proton beam at the Loma Linda Proton Therapy Accelerator during its commissioning at Fermi National Accelerator Laboratory. The intensity monitor was designed to regulate the beam intensity extracted from the proton synchrotron. The proton beam is tunable between 70 and 250 MeV with an adjustable intensity between 10(10) and 10(11) protons per spill. A beam spill is typically 1 s long with a 2-s repetition period. The intensity monitor must be radiation hard, expose minimum mass to the beam, and measure intensity to 1% in 1-ms time intervals. To this end, a 5-cm-thick xenon gas scintillator optically coupled to a photomultiplier tube (PMT) was tested to measure its response to the proton beam. The gas cell was operated at 1.2 atm of pressure and has 12.7-microns-thick titanium entrance and exit foils. The total mass exposed to the beam is 0.14 g/cm2 and is dominated by the titanium windows. This mass corresponds to a range attenuation equal to 1.4 mm of water. The energy lost to the xenon gas is about 70 keV per proton. Each passing proton will produce approximately 2000 photons. With a detection efficiency on the order of 0.05% for this UV light, one would anticipate over 10(10) photoelectrons per second. In a 1-ms time bin there will be approximately 10(7) photoelectrons. This yields a resolution limited by systematics. For unregulated 0.4-s proton spills, we observe a response bandwidth in excess of 10(4) Hz. While signal-to-noise and linearity were not easily measured, we estimate as few as 10(3) protons can be observed suggesting a dynamic range in excess of 10(5) is available.

A performance study of the Loma Linda proton medical accelerator.

More than three years have passed since Loma Linda treated the first cancer patient with the world's first proton accelerator dedicated to radiation therapy. Since that time, over 1000 patients have completed treatments and the facility currently treats more than 45 patients per day. With a typical intensity of 3 x 10(10) protons per pulse and 27 pulses per minute, dose rates of 90-100 cGy/min are easily achieved on a 20-cm diameter field. In most cases, patient treatment times are 2 min, much less than the patient alignment time required before each treatment. Nevertheless, there is considerable medical interest in increasing field sizes up to 40-cm diameter while keeping dose rates high and treatment times low. In this article, beam measurements relevant to intensity studies are presented and possible accelerator modifications for upgrades are proposed. It is shown that nearly all intensity losses can be ascribed to the large momentum spread of the injected beam and occur at or near the injection energy of 2 MeV. The agreement between calculations and measurements appears quite good. In addition, optimum beam characteristics for a new injector are discussed based upon the momentum acceptance and space charge limits of the Loma Linda synchrotron.

A prototype beam delivery system for the proton medical accelerator at Loma Linda.

A variable energy proton accelerator was commissioned at Fermi National Accelerator Laboratory for use in cancer treatment at the Loma Linda University Medical Center. The advantages of precise dose localization by proton therapy, while sparing nearby healthy tissue, are well documented [R. R. Wilson, Radiology 47, 487 (1946); M. Wagner, Med. Phys. 9, 749 (1982); M. Goitein and F. Chen, Med. Phys. 10, 831 (1983)]. One of the components of the proton therapy facility is a beam delivery system capable of delivering precise dose distributions to the target volume in the patient. To this end, a prototype beam delivery system was tested during the accelerator's commissioning period. The beam delivery system consisted of a beam spreading device to produce a large, uniform field, a range modulator to generate a spread out Bragg peak (SOBP), and various beam detectors to measure intensity, beam centering, and dose distributions. The beam delivery system provided a uniform proton dose distribution in a cylindrical volume of 20-cm-diam area and 9-cm depth. The dose variations throughout the target volume were found to be less than +/- 5%. Modifications in the range modulator should reduce this considerably. The central axis dose rate in the region of the SOBP was found to be 0.4 cGy/spill with an incident beam intensity of 6.7 x 10(9) protons/spill. With an accelerator repetition rate of 30 spills/min and expected intensity of 2.5 x 10(10) protons/spill for patient treatment, this system can provide 50 cGy/min for a 20-cm-diam field and 9-cm range modulation.(ABSTRACT TRUNCATED AT 250 WORDS)

Microdosimetry spectra of the Loma Linda proton beam and relative biological effectiveness comparisons.

Protons have long been recognized as low LET radiation in radiotherapy. However, a detailed account of LET (linear energy transfer) and RBE (relative biological effectiveness) changes with incident beam energy and depth in tissue is still unresolved. This issue is particularly important for treatment planning, where the physical dose prescription is calculated from a RBE using cobalt as the reference radiation. Any significant RBE changes with energy or depth will be important to incorporate in treatment planning. In this paper we present microdosimetry spectra for the proton beam at various energies and depths and compare the results to cell survival studies performed at Loma Linda. An empirically determined biological weighting function that depends on lineal energy is used to correlate the microdosimetry spectra with cell survival data. We conclude that the variations in measured RBE with beam energy and depth are small until the distal edge of the beam is reached. On the distal edge, protons achieve stopping powers as high as 100 keV/micron, which is reflected in the lineal energy spectra taken there. Lineal energy spectra 5 cm beyond the distal edge of the Bragg peak also show a high LET component but at a dose rate 600 times smaller than observed inside the proton field.

Two-dimensional and quasi-three-dimensional dosimetry of hadron and photon beams with the Magic Cube and the Pixel Ionization Chamber.

Two detectors for fast two-dimensional (2D) and quasi-three-dimensional (quasi-3D) verification of the dose delivered by radiotherapy beams have been developed at University and Istituto Nazionale di Fisica Nucleare (INFN) of Torino. The Magic Cube is a stack of strip-segmented ionization chambers interleaved with water-equivalent slabs. The parallel plate ionization chambers have a sensitive area of 24 x 24 cm2, and consist of 0.375 cm wide and 24 cm long strips. There are a total of 64 strips per chamber. The Magic Cube has been tested with the clinical proton beam at Loma Linda University Medical Centre (LLUMC), and was shown to be capable of fast and precise quasi-3D dose verification. The Pixel Ionization Chamber (PXC) is a detector with pixel anode segmentation. It is a 32 x 32 matrix of 1024 cylindrical ionization cells arranged in a square 24 x 24 cm2 area. Each cell has 0.4 cm diameter and 0.55 cm height, at a pitch of 0.75 cm separates the centre of adjacent cells. The sensitive volume of each single ionization cell is 0.07 cm3. The detectors are read out using custom designed front-end microelectronics and a personal computer-based data acquisition system. The PXC has been used to verify dynamic intensity-modulated radiotherapy for head-and-neck and breast cancers.

Relative biological effectiveness and microdosimetry of a mixed energy field of protons up to 200 MeV.

We have studied radiation effects utilizing the new 250 MeV Synchrotron at Loma Linda University Medical Center. In this paper we present the data collected for the survival of Chinese hamster lung (V79) cells, that were irradiated with a beam of mixed energy protons up to 200 MeV. The RBE for protons, when compared to 60Co gamma rays, ranged from a low of 1.2 at the high energy portion of the field to 1.3+ at the low energy portion of the field. These results are consistent with the measured lineal energy (microdosimetric) spectra.

Dose error analysis for a scanned proton beam delivery system.

All particle beam scanning systems are subject to dose delivery errors due to errors in position, energy and intensity of the delivered beam. In addition, finite scan speeds, beam spill non-uniformities, and delays in detector, detector electronics and magnet responses will all contribute errors in delivery. In this paper, we present dose errors for an 8 × 10 × 8 cm(3) target of uniform water equivalent density with 8 cm spread out Bragg peak and a prescribed dose of 2 Gy. Lower doses are also analyzed and presented later in the paper. Beam energy errors and errors due to limitations of scanning system hardware have been included in the analysis. By using Gaussian shaped pencil beams derived from measurements in the research room of the James M Slater Proton Treatment and Research Center at Loma Linda, CA and executing treatment simulations multiple times, statistical dose errors have been calculated in each 2.5 mm cubic voxel in the target. These errors were calculated by delivering multiple treatments to the same volume and calculating the rms variation in delivered dose at each voxel in the target. The variations in dose were the result of random beam delivery errors such as proton energy, spot position and intensity fluctuations. The results show that with reasonable assumptions of random beam delivery errors, the spot scanning technique yielded an rms dose error in each voxel less than 2% or 3% of the 2 Gy prescribed dose. These calculated errors are within acceptable clinical limits for radiation therapy.