Design and commissioning of an image-guided small animal radiation platform and quality assurance protocol for integrated proton and x-ray radiobiology research

The Roberts Proton Therapy Center at the University of Pennsylvania consists of the IBA Proteus Plus with a C230 cyclotron (IBA Proton Therapy, Louvain-La-Neuve, Belgium), five clinical treatment rooms (four gantry and 1 fixed beam) and a sixth dedicated research room with two fixed beam lines, herein designated RA6 and RA7. The C230 cyclotron and energy selection system (ESS) can modulate the beam energy in terms of water equivalent thickness (WET) from a maximum energy of 230 MeV (approx. 33 cm WET) to a minimum energy of approximately 70 MeV (approx. 4.1 cm WET). To reduce the range of the proton beam to be useful in small animal models, an additional 4 cm WET range shifter was added to the end of the RA7 beamline.

The SARRP is installed on rails (figure 1(a)) in the shielded research room allowing movement into and out of the RA7 beam path. The integrated SARRP shielding was removed to make loading of animals easier. A remotely operated anesthesia system and multiple video camera views were installed for safely monitoring animal status. The SARRP was mounted on rails so that the imaging capabilities and robotic positioning stage can be shared with the proton beam system providing the capability to use the on-board CBCT for imaging small animals with proton studies. The ability for the SARRP to move in an out of the proton beam, by sliding it along the rails, allows for versatile use of the proton beam with larger equipment, such as water tanks and scintillation detectors, that would not fit on the SARRP stage. By removing the SARRP from the beam path, measurement can easily be made for commissioning and QA procedures. At installation, the SARRP stage was levelled with height adjustable pedestals connecting the SARRP frame to the sliding rail platform. Furthermore, positioning of the SARRP stage, gantry, and rails were designed to be square with the proton beam line to reduce any rotational uncertainty and requiring correction of translational shifts only.

Zoom In Zoom Out Reset image size

A collimation system was designed to shape the proton field size for small animals. The design allows for use of interchangeable brass collimators with the proton beam. Using these collimators, a range of field sizes from 0.5 to 10 mm widths can be achieved. The collimation system is mounted on a movable platform at the end of the RA7 line, which allows for multiple degrees of freedom including translational movement, pitch, and tilt (figure 1(b)) so that the collimator can be optimally aligned with the proton beam at any position.

The IBA Zebra device (IBA Dosimetry, Bartlett TN) was used to measure the percent depth dose (PDD) for a range of energies from 77 MeV to 96 MeV with 0.55 mm resolution (figure 2(a)). The intrinsic resolution of the Zebra detector is 2.0 mm but 1.0 mm resolution is attainable by repeating measurements with a 1.0 mm window inserted. Higher resolution with the Zebra device was achieved using paper index cards. The WET of a stack of 50 cards was first determined. Then a group of five cards (corresponding to 0.55 mm WET) was placed in front of the detection chamber. Data was merged with the corresponding 0.55 mm WET shift from the index cards, 1 mm WET window, and measurement with no WET shift.

Zoom In Zoom Out Reset image size

PDD measurements were also taken using a PTW Bragg Peak Chamber (PTW, Freiburg, Germany) in an IBA 1D water tank (figure 2(b)). The scanning motor and track were mounted in the horizontal plane of the water tank parallel with the proton beam path, as opposed to the standard vertical orientation. The Bragg peak chamber was aligned with the proton beam using EBT3 GafChromic film (Ashland Advanced Materials, Bridgewater, NJ, USA) to verify the position of the proton beam. Measurements were taken in steps of 0.5 mm along the dose distribution. The measurement position was determined for the nominal beam configuration used for small animal irradiations, which includes a 4 cm WET degrader. For measurement of PDD in water, the Bragg peak chamber wall (4.11 mm) and the water tank wall (11.0 mm) with the addition of 2.5 cm of WET solid water were placed in the beam path for a resulting WET of 4.022 cm. The difference in WET between the measurements with the water tank set-up and the nominal beam (0.022 cm) was the shallowest measurement depth shown in figure 2(b).

In addition, lateral beam profiles were measured using EBT3 film for multiple energies and depths. The beam profiles were scanned using an Epson 1000XL flat-bed scanner and analyzed using ImageJ (NIH, Bethesda, MD, USA) and Matlab (version R2017b, MathWorks Inc., Natick, MA, USA) software. Flatness and symmetry were analyzed using the Varian definition with the data smoothed with a 10% span for the moving average.

Absolute dose measurements were performed with an Advanced Markus Chamber (PTW, Freiburg, Germany) according to International Atomic Energy Agency TRS-398 code of practice recommendations. The sensitive volume of the detector (2.5 mm radius) was placed in the middle of a spread out Bragg peak (SOBP), with distal range of 2.3 cm and modulation (width) of 1 cm, of the large uncollimated open field to mitigate volume averaging. The uncollimated open field has a Gaussian shape with a full width half max of 16.8 mm. The chamber was centered on the uncollimated beam and placed in the middle of the SOBP with 1.5 cm of solid water build up. EBT3 film response was calibrated using procedure described in Zhao and Das (2010) and Reinhardt et al (2012). Proton Bragg peak under response of around 20% has been observed for both EBT (Zhao and Das 2010) and EBT2/EBT3 film (Reinhardt et al 2012). To account for the intrinsic proton energy response characteristics of gafchromic film, the EBT3 films were cross-calibrated with the Markus chamber measurement of the uncollimated field at the same measurement depth to ensure a consistent proton energy spectrum for both film and ionization chamber measurements. It was assumed that the effect of scatter on output due to the collimation would be negligible so that the open field film calibration curve could be applied to film irradiated with collimated fields. Dose calculation for films used for dosimetry was performed using parameters from calibrated film of the same batch.

To ensure the reproducibility of studies between different systems, a QA program was developed to maintain the reliability and characteristics specific to each proton and x-ray system. A QA phantom was designed to hold two pieces of Gafchromic film: one at the front face to verify proton beam alignment and another in line with the beam direction to show the depth dose of a Bragg peak. Figure 3(a) shows the resulting QA phantom that was 3D printed at the University of Pennsylvania using a Form2 printer (Formlabs, Boston, MA, USA) and clear engineering resin. A small high-z material (aluminum of 1.2 mm in length) was affixed on the front face of the phantom to target upon CBCT acquisition.

Zoom In Zoom Out Reset image size

Table 1 shows the daily QA requirements to be checked prior to acquiring experimental data. This ensures accurate alignment of the proton beam with the SARRP isocenter and provides a check of the consistency of the beam characteristics as well as the safety interlocks. The front of the phantom, mentioned above, holds a 30 mm  ×  40 mm film (figure 3(b)). The two pieces fit together with another 30 mm  ×  40 mm film between them. The film between the two pieces is tilted at a 5° angle from the horizontal. This ensures that the depth dose profile is not unduly altered by the effects of an air gap between film and phantom. Zhao and Das (2010) found that EBT film placed parallel to a proton beam in a comparable phantom can be used to measure proton range with an uncertainty of 0.5 mm. This accuracy is sufficient for QA and well within the experimental uncertainties for most small animal research. Since the film sits at an angle relative to the horizontal, the position on the film is slightly longer than the depth the proton beam traveled. Each position on the film corresponds to a specific depth in the phantom, however, through the relation R  =  f  · cos(5°) where R represents the depth the proton beam traveled and f  represents the distance between the front end of the film and the point of interest on the film. An example of the depth dose profile from the film is shown in figure 3(c).

The alignment protocol process is shown in figure 3(d). After the SARRP is placed in the proton beam path, the QA phantom can be placed on the SARRP animal positioning stage. A CBCT of the phantom is then acquired. The high-z material in the phantom is easily visible in the image. The center of the high-z material is identified and the robotic stage is moved so that the SARRP isocenter is placed at the high-z material. A proton field is then delivered with collimation. Figure 2(b) shows alignment with a 5  ×  5 mm² collimator, but a smaller collimator can be used for more accurate alignment. After the proton field is delivered, the necessary shift between the high-z material and the center of the proton field (determined by ImageJ) is recorded. The shift and collinearity of the SARRP mechanical and imaging isocenters is then verified with delivery of a second proton beam onto a second piece of film and after the robotic animal stage shift is performed.

A 5  ×  5 mm² collimator was used to deliver an 89 MeV (6.3 g cm⁻²) beam to flank tumor on a C57BL/6 mouse to a dose of 4 Gy. 2 cm WET of bolus was placed in front of the target so that the Bragg peak would be located within the tumor at 3 mm depth. Optimal cutting temperature OCT compound embedded B16F10 tumor tissues were processed and stained for DNA double strand breaks as described previously (Verginadis et al 2017).

PDDs are shown for each measured energy in figure 2. Energies from 77 MeV to 96 MeV, with 2–4 MeV steps, were measured with the IBA Zebra device and a 4 cm range shifter at the end of the beam pipe. A 0.11 mm resolution with a minimum measurement depth of 1.9 mm was achieved with the Zebra device. Data shown in figure 2 does not include under- or over-responding ionization chambers in the Zebra device. Table 2 gives more specific information for the 90% peak width, 90% dose range, and the ratio of entrance (1.9 mm depth) to peak readings. The minimum measurement depth can be further decreased by removing the Bragg Peak chamber (3 mm WET) that was installed between the collimator and the beam pipe. This Bragg Peak chamber will be used for online verification of dose delivery. For each energy, the distal dose falls to  <10% of its peak within 2 mm.

With the range of energies available for this system, a spread-out Bragg peak (SOBP) can be created and delivered. This is done by pre-determining the number of monitor units required for each energy and stopping the beam when that limit is reached. Unique SOBPs can be created for each experiment. The range of the SOBP is limited by the range of the highest energy beam involved. SOBPs with a range of 29.5 mm and width of 24 mm to a range of 8.4 mm and width of 2.9 mm have been commissioned. Pristine Bragg peak weighting to generate an SOBP can be calculated with in-house software developed in Matlab. Figure 4 shows the calculated versus measured SOBP. Generally, the measured SOBP exhibits lower dose relative to the maximum value than the calculated SOBP. The beam delivery system limits the accuracy of beam weighting to 10 monitor units (MU). The calculated SOBP has weighted individual monoenergetic Bragg peaks with higher precision, and during delivery, these weights were rounded to the nearest 10 MU.

Zoom In Zoom Out Reset image size

The lateral beam profile in the x and y  direction (horizontal and vertical directions, respectively) for a 3 mm  ×  3 mm and 5 mm  ×  5 mm collimated beam at multiple depths in a solid water phantom placed on the SARRP stage is shown in figure 5. As expected, the FWHM of the beam profile began to increase slightly with depth from 0 mm depth to 9.8 mm depth and is summarized in table 3. Lateral profiles were determined from film scanned and analyzed for relative density using ImageJ. Film was placed between thin pieces of solid water to determine the lateral profile for multiple depths using the same beam.

Zoom In Zoom Out Reset image size

Flatness and symmetry data are summarized in table 4. The y  profile shows that a generally flat and symmetric distribution is achievable with the system. The x profile, however, shows a slightly higher degree of asymmetry. Collimator size also affects the flatness and symmetry measured. The 5  ×  5 mm² field sizes improved flatness and symmetry values (<6%) compared to the 3  ×  3 mm² field sizes (<12%). Increased field asymmetry can indicate a slight misalignment between the proton beam and the collimator. If the collimator is tilted with respect to the beam axis, the beam can be partially scattered, resulting in a tilted profile. This effect is used in the QA protocol to check the alignment of the collimator. The current collimator mounting system benefits from having finely tunable stages; however, it is easily susceptible to movement due to external forces or movement of the collimator in and out of position along the beam axis for clearance when moving the SARRP on its rails. Furthermore, analysis of flatness and symmetry can be limited for small field sizes and film profiles.

Using the QA phantom and protocol developed, the alignment shift between the SARRP isocenter and the proton beam center was determined both before and after shifting the robotic stage. By applying the robotic stage shifts, the discrepancy between the SARRP and the proton beam isocenters was improved from 2.67  ±  0.38 mm to 0.12  ±  0.04 mm (figure 6). The alignment and reproducibility were improved upon anchoring the x-ray SARRP in position against the rail stops. Without anchoring, the SARRP would drift to a slightly different position, resulting in larger discrepancies between the final x-ray-proton alignment (4.98  ±  2.15 mm pre-alignment to 1.46  ±  0.59 mm post-alignment). Example films from the protocol are shown in figures 3(b) and (c).

Zoom In Zoom Out Reset image size

Prior to any small animal experiments, this QA protocol should be performed to ensure that the beam quality is consistent and to determine the necessary shift between the CBCT isocenter and the proton field. After acquiring a CBCT of the small animal, the target is determined, and a shift of the robotic animal stage by the pre-determined amount from the QA protocol will ensure that the proton field is delivered accurately. The protocol takes 10–15 min depending on the beam availability and will be easy to implement prior to the day's experiments.

One of the main objectives of this study was to determine whether damage to the tumor tissue could be restricted to a specified depth using the data commissioned for the proton beam line and the QA protocol that was developed. A very well established method to analyze the direct damage post-irradiation is the detection of the phosphorylation of histone H2AX at serine 139 (γH2AX). This method is considered to be very sensitive and quantitative for detecting DNA-double strand breaks. Immunofluorescence staining for γH2AX on the B16F10 tumors harvested at 1 h after 4 Gy of proton irradiation showed that indeed the depth of the damage is restricted to 3 mm, corresponding to the depth of the Bragg peak set using energy selection and appropriate build-up (figure 7). Moreover, the width of the damaged area (around 4–5 mm) corresponds to the size of the 5  ×  5 mm² collimator used (figure 7), while the rest of the tissue has no visible DNA damage.

Zoom In Zoom Out Reset image size