Scintillating fiber based in-vivo dose monitoring system to the rectum in proton therapy of prostate cancer: A Geant4 Monte Carlo simulation

Purpose: To construct a dose monitoring system based on an endorectal balloon coupled to thin scintillating fibers to study the dose to the rectum in proton therapy of prostate cancer. Method: A Geant4 Monte Carlo toolkit was used to simulate the proton therapy of prostate cancer, with an endorectal balloon and a set of scintillating fibers for immobilization and dosimetry measurements, respectively. Results: A linear response of the fibers to the dose delivered was observed to within less than 2%. Results obtained show that fibers close to the prostate recorded higher dose, with the closest fiber recording about one-third of the dose to the target. A 1/r2 (r is defined as center-to-center distance between the prostate and the fibers) decrease was observed as one goes toward the frontal and distal regions. A very low dose was recorded by the fibers beneath the balloon which is a clear indication that the overall volume of the rectal wall that is exposed to a higher dose is relatively minimized. Further analysis showed a relatively linear relationship between the dose to the target and the dose to the top fibers (total 17), with a slope of (-0.07 ± 0.07) at large number of events per degree of rotation of the modulator wheel (i.e., dose). Conclusion: Thin (1 mm × 1 mm), long (1 m) scintillating fibers were found to be ideal for real time in-vivo dose measurement to the rectum during proton therapy of prostate cancer. The linear response of the fibers to the dose delivered makes them good candidates as dosimeters. With thorough calibration and the ability to define a good correlation between the dose to the target and the dose to the fibers, such dosimeters can be used for real time dose verification to the target.


Introduction
Prostate cancer is the second leading cause of cancer related death in men, with estimated deaths of about 29,720 for 2013. [1][2][3][4] A number of treatment options are available for prostate cancer patients. The most commonly used radiation therapy modalities include: Brachytherapy implants [5][6][7] , in-tensity modulated radiation therapy (IMRT) [8][9][10][11][12][13] , and proton therapy. [14][15][16][17] Due to the characteristic Bragg peak and physical properties of the proton beam, it possesses the ability to penetrate deep before stopping in the medium. 18 Sharp lateral penumbra of the beam also allows sparing of the critical structures adjacent to the tumor volume. 19 The formation of the Bragg peak in the depth dose distribution and the ability to spread the dose out using a modulator wheel to generate a spread out Bragg peak (SOBP) allows one to highly localize the dose delivered to the target. This gives the advantage of sparing much of the healthy tissue around the target, especially in the distal region. Hence, proton beam therapy is an attractive option for the treatment of cancer. The ability to geometrically conform the dose to the target, however, should be accompanied with stringent requirements in the target localization to avoid error in the dose delivered. The two main sources of uncertainties for this are patient setup and organ motion. During prostate cancer therapy, bladder and rectum emptying and/or filling are the main contributors to the organ motion. 20,21 Research studies showed that such uncertainty can result in the introduction of prostate displacement in the inferior-superior (IS), anterior-posterior (AP), and right-left (RL) directions. Huang et al, 13 for instance, measured a maximum displacement in the AP (6.8 mm anteriorly to 4.6 mm posteriorly) and in the SI (3.5 mm superiorly to 6.8 mm inferiorly) directions. Such displacements are large enough to introduce a remarkable error in the dose delivered to the prostate, especially when proton beams are used to treat the tumor.
A number of techniques were devised to minimize the uncertainty introduced by organ motion in the treatment of prostate cancer. The use of an endorectal balloon, currently practiced by a number of cancer centers, was found to greatly reduce the prostate motion by internally immobilizing it. 20,22 This technique reduces the prostate displacement to less than 1 mm according to the study results from McGary et al. 22 In addition to internally immobilizing the prostate, researchers observed some additional benefits from the use of an endorectal balloon. First, it reduces greatly the overall volume of the rectal wall exposed to a higher dose. 20,23 Second, the balloon can be used to give a good prediction of the prostate location. 24 Most of the currently utilized balloons have an embedded fiducial marker that can be easily identified even in poor quality images. In the presence of a good reproducibility in the placement of the balloon, since it is also used during the simulation computed tomography (CT) scan, fiducial markers (instead of bonny landmarks) can be utilized for patient setup and prostate organ location. 24 Third, some researchers took advantage of the use of an endorectal balloon to develop an in-vivo dose measurement tool for the rectum. 25,26 This can be performed by the use of an appropriate dosimeters attached to the balloon to record the dose at the surface of the balloon, which is in contact with the inner rectal wall.
In this work, the Geant4 Monte Carlo toolkit 27, 28 was used to simulate the proton therapy of prostate cancer using an endorectal balloon as an internal immobilization device. Thin scintillating fibers were attached to the balloon for real time in-vivo dose measurement to the surface of the balloon. The characteristic properties that scintillating fibers possess such as water equivalency, temperature and pressure independence, fast response time etc. makes them ideal dosimetric tools for this task. 29 -33 The fact that they can be fabricated in any size and shape also makes them good candidates as they can be prepared in small sizes for a better resolution. This can also make the fibers less intrusive as far as the patient is concerned in clinical applications.

Methods and Materials
The Hadrontherapy Geant4 example application was used as a reference to design the simulation described in this work. 34 The modulator wheel that spreads out the Bragg peak to generate an SOBP was modified to accommodate high energy proton beams. A spherical geometry of diameter 2.9 cm, which is made up of a soft tissue equivalent material, G4_TISSUE_SOFT_ICRP, obtained from the National Institute of Standards and Technology (NIST) database 35 , was introduced to simulate the prostate. Geometries for the endorectal balloon as well as the thin long scintillating fibers (1 mm × 1 mm x 1 m) laying on the top surface and beneath the balloon were also introduced and their placements are detailed in Figure 2.
The beam line was designed with a simplified geometry to mimic the beam line of the Hampton University Proton Therapy Institute (HUPTI). 36 Hence, a basic difference exists between the beam line in the simulation design developed and the one at HUPTI.

Results
The Geant4 based simulation developed was validated against dedicated data acquired at HUPTI for a range of 257 mm and a modulation of 13 mm. The SOBP generated for a 200 MeV proton beam is compared to the measured data from HUPTI in Figure 1 (arbitrary units were used to normalize the data in the SOBP region).
The simulation data are in good agreement with the measured data. From Figure 1, the analysis of the data shows that at the beginning of the SOBP and at the dose fall off regions, the data points match within 4% and 3%, respectively. At the end of the SOBP, however, there is a remarkable difference between the two data sets. This is due to the combined effect of the beam line (differences in the geometry) and the modulator wheel residual. At the distal 10% dose, a depth difference of 2.5 mm was observed. The noticeable difference between the measured and the simulation data at the entrance dose is due to the effect of the different amount of secondary particles generated, a consequence of the beam line discrepancy between the simplistic simulation and the actual beam line at HUPTI.
In the simulation, a total of 34 fibers were placed around the endorectal balloon for dosimetry (17 atop and 17 beneath). The placement of the fibers around the balloon is shown in Figure 2. Figure 2, the fibers were laid on the surface of the balloon, with the fiber ID running from 1 to 17 from left to right. The beam direction is from left to right, irradiating the whole prostate. At the end of the simulation run, the energy deposited in all 34 fibers were recorded by requiring a cut in the volume of interest. Therefore, the knowledge of the geometrical location of each fiber is an important step.

As shown in
To acquire such information, the beam propagation axis (x-axis) was aligned with the center of the balloon. The plot of the interaction points for the fibers in the Y-X (transverse-longitudinal, cross sectional view) and Y-Z (transverse, beam eye view) planes is shown in As shown in Figure 2, the fibers were laid on the surface of the balloon, with the fiber ID running from 1 to 17 from left to right. The beam direction is from left to right, irradiating the whole prostate. At the end of the simulation run, the energy deposited in all 34 fibers were recorded by requiring a cut in the volume of interest. Therefore, the knowledge of the geometrical location of each fiber is an important step.
To acquire such information, the beam propagation axis (x-axis) was aligned with the center of the balloon. The plot of the interaction points for the fibers in the Y-X (transverse-longitudinal, cross sectional view) and Y-Z (transverse, beam eye view) planes is shown in Figure 3.  As shown in Figure 2, the fibers were laid on the surface of the balloon, with the fiber ID running from 1 to 17 from left to right. The beam direction is from left to right, irradiating the whole prostate. At the end of the simulation run, the energy deposited in all 34 fibers were recorded by requiring a cut in the volume of interest. Therefore, the knowledge of the geometrical location of each fiber is an important step.

International Journal of Cancer Therapy and Oncology
To acquire such information, the beam propagation axis (x-axis) was aligned with the center of the balloon. The plot of the interaction points for the fibers in the Y-X (transverse-longitudinal, cross sectional view) and Y-Z (transverse, beam eye view) planes is shown in Figure 3.  The energy deposited in the bottom fibers, which corresponds to the inner posterior wall of the rectum, was negligibly small compared to the energy deposited in the top fibers as well as inside the prostate: the maximum value recorded was 3.99 MeV for the central and for the 3000 events per degree of modulator rotation run. The energy deposited in the prostate as well as the total energy deposited in the top 17 fibers are given in Table 1. The table also shows the number of protons delivered in each setup. One can observe that the number of protons used is very small compared to the number required in the clinic to deliver a dose of 1.8 Cobalt Gray Equivalent (CGE) to a Kg of water, which is of the order of 10 10 . This is the dose per fraction most of the time prescribed for prostate cancer patients at HUPTI. In the simulation, the dose delivered was of the order of a few μGy due to computational limitations.   The energy deposited in the bottom fibers, which corresponds to the inner posterior wall of the rectum, was negligibly small compared to the energy deposited in the top fibers as well as inside the prostate: the maximum value recorded was 3.99 MeV for the central and for the 3000 events per degree of modulator rotation run. The energy deposited in the prostate as well as the total energy deposited in the top 17 fibers are given in Table 1. The table also shows the number of protons delivered in each setup. One can observe that the number of protons used is very small compared to the number required in the clinic to deliver a dose of 1.8 Cobalt Gray Equivalent (CGE) to a Kg of water, which is of the order of 10 10 . This is the dose per fraction most of the time prescribed for prostate cancer patients at HUPTI. In the simulation, the dose delivered was of the order of a few μGy due to computational limitations.   The energy deposited in the bottom fibers, which corresponds to the inner posterior wall of the rectum, was negligibly small compared to the energy deposited in the top fibers as well as inside the prostate: the maximum value recorded was 3.99 MeV for the central and for the 3000 events per degree of modulator rotation run. The energy deposited in the prostate as well as the total energy deposited in the top 17 fibers are given in Table 1. The table also shows the number of protons delivered in each setup. One can observe that the number of protons used is very small compared to the number required in the clinic to deliver a dose of 1.8 Cobalt Gray Equivalent (CGE) to a Kg of water, which is of the order of 10 10 . This is the dose per fraction most of the time prescribed for prostate cancer patients at HUPTI. In the simulation, the dose delivered was of the order of a few μGy due to computational limitations.  The calculation of the energy deposited per unit volume (MeV/cc) gave 4062 and 1214 for the prostate and the top middle fiber (Fiber ID 9), respectively. These correspond to an absorbed dose of 0.65 μGy and 0.19 μGy for the prostate and the fiber, respectively. This result showed that the section of the rectal wall represented by Fiber ID 9 is pushed by the inflated balloon to the high dose region, receiving about a third of the dose deposited in the prostate. As one moves away from the middle fiber towards the frontal and distal regions, the value of the energy deposited per unit volume decreases dramatically. For Fiber ID 1 (frontal region) and Fiber ID 17 (distal region) the values were 517 (factor of 1/8) and 428 (factor of 1/9), respectively, corresponding to an absorbed dose of approximately 0.08 μGy and 0.07 μGy to the frontal and distal fibers, respectively. In the actual treatment of a prostate cancer patient, the anterior section of the rectal wall is usually included in the PTV. Consequently, the top Fiber ID 9 will be inside the PTV, receiving the prescribed dose. Hence, this fiber can be used for real time dose verification to the PTV during irradiation. In the simulation, however, all the top fibers were kept outside of the field, a reason for the absorbed dose to be lower for the fibers than that of the prostate. Similar results were obtained for the other simulation runs. The corresponding plot of the energy deposited versus the Fiber IDs is shown in Figure 4. The energy deposited as a function of the number of events per degree of modulator rotation (i.e., dose), for the prostate and the total energy deposited to the 17 top fibers, is shown in Figure 5.  the rectal wall is usually included in the PTV. Consequently, the top Fiber ID 9 will be inside the PTV, receiving the prescribed dose. Hence, this fiber can be used for real time dose verification to the PTV during irradiation. In the simulation, however, all the top fibers were kept outside of the field, a reason for the absorbed dose to be lower for the fibers than that of the prostate. Similar results were obtained for the other simulation runs. The corresponding plot of the energy deposited versus the Fiber IDs is shown in Figure 4. The energy deposited as a function of the number of events per degree of modulator rotation (i.e., dose), for the prostate and the total energy deposited to the 17 top fibers, is shown in Figure 5.  the rectal wall is usually included in the PTV. Consequently, the top Fiber ID 9 will be inside the PTV, receiving the prescribed dose. Hence, this fiber can be used for real time dose verification to the PTV during irradiation. In the simulation, however, all the top fibers were kept outside of the field, a reason for the absorbed dose to be lower for the fibers than that of the prostate. Similar results were obtained for the other simulation runs. The corresponding plot of the energy deposited versus the Fiber IDs is shown in Figure 4. The energy deposited as a function of the number of events per degree of modulator rotation (i.e., dose), for the prostate and the total energy deposited to the 17 top fibers, is shown in Figure 5.

Discussion
From the results obtained, one observes that increasing the number of events per degree of rotation (i.e., dose) increases the energy deposited inside the fibers. In Figure 4, it can be observed that the energy deposited is largest around the middle fibers, especially for the central (Fiber ID 9), which is the closest to the prostate. In moving to the frontal (Fiber ID 1) and distal (Fiber ID 17) regions, the energy deposited decreases as 1/r 2 (r is defined as center-to-center distance between prostate and fiber), which implies that the overall rectal wall volume exposed to higher radiation is smaller. This result is in good agreement to the results obtained by other researchers. 20,23 In the actual planning of the treatment, the top section of the rectum is usually enclosed within the planning target volume (PTV). Hence, the Fiber ID 9, located at the top middle, is usually enclosed within the PTV receiving the prescribed dose. With appropriate calibration, this fiber can be used to monitor the dose delivered to the PTV.
Further analysis of the data showed that there is also a linear response of the scintillating fibers to the radiation dose delivered to the prostate, with a precision of less than 2% as show in Figure 5: slope of 0.853±0.077 MeV/events per degree for the prostate and 0.275±0.023 MeV/events per degree for the fibers. This result is one of the characteristic properties studied by a number of other researchers that make the fibers ideal as a dosimetric tool. [29][30][31][32][33] Moreover, the ratio of the energy deposited in the prostate to the total energy deposited in the top 17 fibers gets relatively constant, especially for larger number of events (slope of -0.07±0.07 MeV per Gy to the prostate). Once a clear relationship is defined between the dose to the prostate and the dose to the fibers, those fibers can be used as dose verification tools to the target.

Conclusion
In this work, the use of an endorectal balloon coupled to thin scintillating fibers was studied as a potential dose monitoring system in proton therapy of prostate cancer. Based on the results obtained, the energy deposited in the fibers placed atop the balloon was observed to decrease as 1/r 2 as one goes to the distal and proximal ends from the central fiber. In addition, the total energy deposited inside the 17 fibers located beneath the balloon near the posterior section of the rectal wall was 0.27% of the energy deposited in the top middle fiber (Fiber ID 9) alone. These results indicate that the top section of the rectal wall is pushed to the high dose region while the overall volume of the rectal wall exposed to a higher dose is very small. Further analysis of the data shows that the ratio of the dose deposited in the prostate to the total dose deposited in the top 17 fibers is relatively flat with a slope of -0.07±0.07, especially at large number of events. Hence, with appropriate calibration, these fibers can be used to give a precise prediction of the dose delivered to the prostate, opening the capability of an in-vivo dose monitoring system during irradiation. An experimental setup will be designed based on the simulation presented herein. The results will be presented in a forthcoming paper.