Microdosimetric characterisation of radiation fields for modelling tissue response in radiotherapy

Uncertainty in dose per monitor unit estimates for passively scattered proton therapy: The role of compensator and patient scatter
in prostate cases
Wayne Newhauser1,2,3,4,*, Annelise Giebeler1,2,5, Ronald Zhu1,2, Uwe Titt1,2, Andrew Lee1,6, Rui Zhang1,2,3,4

1Department of Radiation Physics and Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
2The University of Texas Graduate School of Biomedical Sciences at Houston, Houston, Texas, USA
3Department of Physics and Astronomy, Louisiana State University, Medical Physics Program, Baton Rouge, Louisiana, USA
4Mary Bird Perkins Cancer Center, Baton Rouge, Louisiana, USA
5Scripps Proton Therapy Center, Summers Ridge Road, San Diego, California, USA

6Texas Center for Proton Therapy, Irving, Texas, USA

Cite this article as: Newhauser WD, Giebeler A, Zhu R, Titt U, Lee A, Zhang R. Uncertainty in dose per monitor unit estimates for passively scattered proton therapy: The role of compensator and patient scatter in prostate cases. Jour Proton Ther 2015; 1:116.
DOI: http://dx.doi.org/10.14319/jpt.11.6

Original Article


Standard methods for determining dose per monitor unit values in a patient do not yet exist for proton therapy. Indeed, some aspects of D/MU estimationremain poorly understood, such as the conversion of absorbed dose in a water phantom to absorbed dose in a patient. This study focused on the water-to-patient absorbed dose conversion factor, FCSPS, which accounts for differences in scatter (from the range compensator and internal patient anatomy) between patient treatments and their corresponding calibration irradiation in a homogeneous water-box-phantom. We estimated FCSPS for 32 prostate fields using a pencil beam dose algorithm in the treatment planning system (TPS). The mean FCSPS value was 1.006; its standard deviation of the mean was ±0.001. The lower bound for uncertainty in FCSPS, μFCSPS, was estimated for a sub-set of fields through comparisons of TPS dose predictions with measurements and Monte Carlo (MC) simulations. Comparison of TPS predictions and measurements yielded μFCSPS of 0.4% - 0.8%. Comparison of TPS predictions and MC simulations yielded μFCSPS < 0.3%. For a prostate treatment, a comparison of FCSPS values from TPS predictions with the historical value of 1.0 yielded μFCSPS < 3% and a mean μFCSPS of 0.6%. Regardless of estimation method, μFCSPS was approximately 1%, suggesting that uncertainty in FCSPS for proton treatments of prostate cancer is clinically acceptable.

Keywords: Dose Per Monitor Unit; Prostate Cancer; Proton Therapy; Uncertainty

1. Introduction

Proton therapy is gradually becoming more available to the general patient population.1, 2 Specifically, the number of proton treatment centers increased from 15 to 46 from 2000 to 2014.3 An analysis of cancers treated with proton therapy revealed that the most commonly treated diseases were uveal melanoma (35%) and prostate cancer (26%).4 Of these, treatments for prostate cancer are of particular interest because of its high incidence.5 In addition, recent studies report that relative to conventional photon radiotherapy, proton therapy for localized prostate cancer permits reduced dose to critical structures6-13 and successful dose escalation7, 14, 15. However, because proton therapy was previously limited to a few centers worldwide, there have been no large-scale randomized clinical trials of proton vs. photon radiotherapy for prostate cancer.16, 17 Thus, there is a strong impetus to standardize the dosimetry of proton radiotherapy so that patient treatment and outcome data can be directly compared. In this respect, advisory bodies have published proton dosimetry protocols for determining proton beam output, or absorbed dose per monitor unit (D/MU) values, in water under reference conditions (such as the American Association of Physicists in Medicine18, European Clinical Heavy Particle Dosimetry Group (ECHED)19,20, International Commission on Radiation Units and Measurements1, and the International Atomic Energy Agency 21.  Still, in contrast to photon and electron therapy, to date, a protocol to harmonize methods for estimating absorbed dose from proton therapy in a patient is lacking.

Nonetheless, progress is being made toward a more complete understanding of the determination of absorbed dose in a patient receiving proton therapy.22-31 Specifically, progress includes the creation of a geometrical framework and an initial estimate of 1.0 for a water-to-patient absorbed dose conversion factor22 as well as validation of Monte Carlo (MC)-simulated D/MU data within 1 to 1.5% of measured data25, 27, 29, 31. In addition, studies by Fontenot et al.26, Akagi et al.24, and Titt et al.28 indicate that absorption and scatter of the treatment beam in field-specific collimation devices, range compensators (RC), and patient anatomy have the potential to increase uncertainty in estimates of absorbed dose per monitor unit in the patient, (D/MU)p, by 1% or more. In this respect, Akagi et al.24 demonstrated that the combined effects of scatter in the RC, scatter in the patient, and scatter from the patient-specific collimator could yield (D/MU)p values 2% to 3% higher or lower than those measured in a phantom. These results were confounded by field-size effects, which were subsequently addressed by Titt et al.28 Finally, Fontenot et al.26 addressed uncertainty in D/MU measurements under patient-specific fields associated with the presence of the range compensator and they recommended measurement without the range compensator. Together, these works underscore the need for standard methods of determining (D/MU)p values, the potential complexities of doing so, and the need to better understand the total uncertainty in  (D/MU)p

The objective of this study was to estimate total uncertainty in (D/MU)p, values for patients who receive proton therapy for prostate cancer. In particular, we used treatment planning system (TPS) calculations, measurements, MC simulations, and a comparison to the historical value of 1.0 to quantify the least understood factor in the water-to-patient absorbed dose conversion factor, FCSPS, the multiplicative factor that takes into account scatter from the RC and internal patient scatter, and its uncertainty.

2. Methods and Materials

2.1 D/MU formalism

The formalism for calculating the beam output, or D/MU value, for proton therapy was based in part on previously reported methods.30,32 Our formalism includes D/MU estimates under reference conditions, field-specific treatment conditions, and D/MU estimates in the patient (Figure 1).

Figure 1: Schematic diagram for the measurement formalism in the reference (a), treatment-field-specific (b), and patient (c) conditions. The parameters describing these conditions are the beam range, R; modulation, M; depth to calibration point, d; field size, FS; and source-to-calibration point distance, SCD, which is isocenter for the fields in this project. In (c) 'RC' specifies the field-specific range compensator.

We first defined D/MU in a water phantom under reference conditions, shown in Figure 1a and denoted by (D/MU)refw ≡ 1 cGy MU-1. The reference condition comprises a collimated 10 cm × 10 cm field with a range of 28.5 cm (250 MeV proton beam), a 10-cm spread-out Bragg peak (SOBP), and a center of modulation at 23.5-cm depth in water that was located at isocenter. This range corresponds to the most penetrating beam available from the treatment unit at our institution with a medium field size. Note that a water equivalent phantom, e.g., a solid plastic phantom, may also be used. For simplicity and brevity, we shall consider these as being interchangeable.

We next defined D/MU value in a water phantom as before, except the proton beam parameters were taken from a treatment field (Figure 1b), as


where Fw,ref is a conversion factor that takes into account differences between the D/MU values at the reference condition and the treatment-field-specific condition in water.  Fw,ref is defined as

where Fi are as follows. FOF corrects for changes in beam output which result from differences in the proton beam energy spectrum relative to the reference condition; these changes are due to differences in the properties of the beam that is injected into the treatment head and due to scatter and absorption in the range modulator wheel. FRS corrects for changes in beam output due to the range shifter. FSOBP corrects for changes in the beam output due to differences in the SOBP relative to the reference field. F­InvSq corrects for changes in beam output which result from differences in beam divergence relative to the reference condition; changes in beam divergence occur when there are changes in the distance from the effective source to the field specific point of measurement, or calibration point. (The calibration point for the reference condition was located at isocenter; the distance from the effective source to isocenter for our passively scattered beamlines was 270 cm.) FFS corrects beam output for differences in proton fluence due to changes in the uncollimated field size (i.e., the amount of lateral beam spreading), and FColS corrects beam output for differences in scatter from the reference aperture (10 cm x 10 cm) to the patient-specific aperture.28

We defined the (D/MU) value in the patient (Figure 1c) according to

where, Fp,w is a conversion factor that accounts for differences in the D/MU values between the field-specific calibration condition in water and the patient-specific treatment condition in tissue, or

FMS corrects for effect of differences in the proton mass stopping power in tissue relative to that in water. It can be determined using MC simulations or analytical calculations.33, 34 The compensator scatter and patient scatter factor, FCSPS, takes into account differences in the beam output due to differences in the scattering and attenuation within the patient and RC together relative to that of a water box phantom and no RC.

FCSPS was the focus of our study and we considered two methods for estimating it. In the first method, which we shall refer to as the treatment planning system method (TPS method), we define

where DpRC/MU is the absorbed dose per MU in the patient at the calibration point (described below) with the patient-specific RC present in the field, and DWno_RC/MU is the absorbed dose per MU in water (or water equivalent material) at the calibration point without the RC in the field.

In the second method, named the treatment planning system and measurement method (TPS+M method), we approximated FCSPS as  

where, FCS corrects beam output for compensator scatter and is given by

and FPS corrects beam output for internal patient scatter and is given by

DwRC/MU in equation (7) is the absorbed dose per MU in water (or water equivalent material) at the calibration point with the RC in the field,  and the other terms in equations (7) and (8) are as defined for equation (5). 

The calibration pointis the location where the relative absorbed dose (in MU) is calibrated to the prescribed absolute absorbed dose (in Gy). It typically corresponds to a region in the patient or phantom that will receive uniform dose and is close to the center of the SOBP. Because the prostate treatment plans used in this work were isocentric, all treatment fields within a plan shared the same calibration point, i.e., isocenter (IEC 1989). As such, the calibration point for each individual treatment field occupied the same point in space, whether in water or in the patient.

2.2 Uncertainty budget

We used an uncertainty budget for (D/MU)p as an a priori guide in studying the numerical impact of uncertainty in various factors on uncertainty in (D/MU)p. Specifically, we prepared a lookup table (see Table 1) to determine which intervals of uncertainty in (D/MU)w, FColS, and FCSPS, respectively, would cause ≤ 5% uncertainty in(D/MU)p. In it, we considered three methods for determining (D/MU)p. The main difference between methods was the way in which (D/MU)w was determined, i.e., through use of a TPS, measurements, or MC simulations.

When using a TPS to predict (D/MU)w, we have

where, (D/MU)wTPS was estimated by the pencil beam algorithm (PBA) algorithm35 in TPS, FColS is defined in equation (2), and FMS, FCSPS are defined according to equation (4).  (D/MU)wTPS and FMS were obtained from interpolation of measured values, and FCSPS was obtained using a TPS (Eclipse, Varian Medical Systems, Palo Alto, CA).  However, because the TPS did not calculate the contribution of lateral scatter from the edges of the field-specific collimator28, we included FColS (as a modifier of (D/MU)w) in the estimation of (D/MU)p

When using measurements of (D/MU)w (with the patient-specific collimator in place), we have

where, FColS does not appear because it is implicitly taken into account in (D/MU)wmeas. Finally, when (D/MU)w was estimated using MC simulations, we have

where,  (D/MU)wMC inherently includes consideration of each material in the proton beam path (and its mass stopping power) as well as lateral scatter from the field-specific collimator. Thus, FCSPS is the only correction factor in the equation, provided (D/MU)wMC is determined in a water box phantom.

We used standard methods for error propagation to estimate the relative uncertainty in (D/MU)p as,

where, μ was the symbol used to represent uncertainty. Each quantity in equations (9) through (11) has an associated uncertainty, and several of those uncertainties were variable or unknown. Thus, the estimation of relative uncertainty in (D/MU)p (Table 1) was based on a combination of three methods, including standard propagation of errors36,  use of uncertainty values from the literature, and sensitivity testing to quantify the impact of contributing uncertainties that were not available from the literature or that were not determined in this work. The uncertainty factors μ(D/MU)w, μFCS, μFCSPS, and μFMS were found to be uncorrelated, and intervals of each factor were found that satisfied the 5% uncertainty criterion of (D/MU)p. Values of μ(D/MU)w were restricted to the interval from 2.5% to 4.5%, μFColS was set to 1.0%, as was μFMS, and μFCSPS were restricted from 1.5% to 4.5%. In other words, the results of the a priori estimates of relative uncertainty in (D/MU)p (Table 1) were used to estimate plausible values of the uncertainty in (D/MU)p when various values of μ(D/MU)w, μFColS, μFCSPS, and μFMS were taken into account.

2.3 Parameter values used to populate the uncertainty budget

Table 2 lists the parameter values used in equation (12) to populate Table 1, their associated uncertainties, and the corresponding literature sources. We used a value of 1.02 for (D/MU)w because it is typical for a prostate patient treated at our institution. However, values of (D/MU)w potentially depend on how (D/MU)wref is defined by individual facilities and the protocols they follow.1,21,37 Therefore, an interval of (D/MU)w values was provided. This interval takes into account reported differences between the ICRU and IAEA protocols.38,39,40 Consequently, the interval of values used to estimate the relative uncertainty in (D/MU)w was ±2.5% to ±4.5%.

The value used for FColS, 1.02, reflects the importance of the collimator in proton beam dosimetry.28 The uncertainty for FColS, μFColS, was estimated using an interval of ±1%, which accounted for deviations in the treatment field from a 10 cm x 10 cm collimated field size, variation in beam energies from 160 MeV to 250 MeV, and variation in the location of the calibration point relative to the center of the SOBP. 

The value used for FMS, 1.0, is an estimate of the dosimetric effect that a medium other than water causes due to differences in the proton mass stopping powers. The uncertainty in FMS, μFMS, was estimated at 1% using mass stopping power ratios41 of water to muscle.

The values for FCS and FCSPS were determined using data from the only publications which directly addressed patient and compensator scatter.24, 30 Akagi et al.24 used a water phantom to determine values for FCS and FCSPS, and Sahoo et al.30 used a commercial TPS and its verification plan feature to report a mean FCSPS value of 1.00±0.04 for a sample of unspecified anatomical treatment locations. However, because neither study provided an estimate of FCSPS for prostate treatment fields, we used the arithmetic average of values from Akagi et al.24 and Sahoo et al.30, yielding the FCSPS value of 1.03 listed in Table 2. The uncertainty in our estimate of FCSPS in Table 2 was unknown, thus determining a value of μ(FCSPS) was a central focus of this work (section 2.5) as it is needed for the estimation of uncertainty in (D/MU)p.

2.4 Estimation of FCSPS for prostate treatment fields

2.4.1 Estimation of FCSPS using the TPS method

The TPS method (eq. 5) was applied to each of 32 prostate treatment fields taken from a representative sample of patients (n = 16, 2 treatment fields each) from our practice.  Patients were selected using the consecutive sampling method42 to minimize selection bias and indexed as 1 through 16. Patients in this study (1) received passively scattered proton treatments for stage I or II prostatic adenocarcinoma and (2) a D/MU calibration date within a year of this study's inception. The first date in the calibration interval was selected arbitrarily, and the end date was based on the date on which the desired number of consecutive patients had been treated.

To calculate absorbed dose to water (or water equivalent material), i.e., DWno_RC in equation 5, we used the verification plan feature of the TPS, utilizing the same beam energy, lateral scatterer, range shifter, range modulation, and collimation as in the patient's treatment plan but with the patient's CT anatomy replaced by a water-box-phantom. The procedure for creating verification plans was taken from Newhauser.43 Briefly, in all fields, the calibration point location remained fixed at isocenter (Figure 2) to minimize the dosimetric impact of differences in beam divergence. Also, as described by Newhauser22, the water-equivalent depth of the calibration point was made equal for a treatment plan and the corresponding verification plan by shifting the water phantom in the verification plan. By maintaining the same location and water-equivalent depth of the calibration point in the patient and water phantom, dosimetric differences due to differences in scatter in the patient and phantom were isolated, so their respective effects on dose delivered to the calibration point could be revealed.

2.4.2 Estimation of FCSPS using the TPS+M method

The TPS-plus-measurement (TPS+M) method used a combination of measurements of absorbed dose in a water phantom and TPS calculations to estimate FCSPS according to equation (6). This method was applied to four treatment fields from two patients. These patients met the same inclusion criteria as patients 1 to 16 except they were selected prospectively so that we were able to make the additional measurements required for this method. The corresponding patient indices were 17 and 18 (Table 3). Estimation of FCS using measurements

Our measurements of FCS, using equation (7), utilized a 0.015-cm3 air-filled ionization chamber (PTW pinpoint chamber, model TN31041, serial number 0079; Freiburg, Germany), an electrometer (Scandronix Wellhofer Dose 1, serial number 0293; Schwarzenbruck, Germany) and a plastic phantom (polymerized methyl methacrylate; C5H3O2, ρ = 1.19 g cm-3; GE Plastics Inc., Pittsfield, MA).  Measurements were taken twice for each field: once at the calibration point with the RC in place, DWRC , and once at the calibration point without the RC in place, DWno_RC.  In the measurement for DWno_RC , the water-equivalent thickness of the phantom was increased to preserve a fixed location of the calibration point.  Finally, FCS was estimated according to equation (7). Estimation of FCS using the TPS

The second step in generating FCSPS values with the TPS+Mmethod was to determine FPS using equation (8). Because in vivo measurements were not feasible, the PBA in the TPS was used to determine the ratio of absorbed dose at the calibration point in the patient, Dp,calRC, to that in a water phantom, Dw,calRC .

2.5 Uncertainties in FCSPS

2.5.1 Estimation of uncertainty in FCSPS   from the TPS method

When FCSPS was determined using the TPS method, i.e., TPS dose estimates applied to equation (5), the corresponding uncertainty in FCSPS, (μFCSPS)TPS, was estimated by comparing values of absorbed dose in the patient generated by the TPS calculations and MC simulations.  As a result, we estimated that

where, DTPS represents absorbed dose from the TPS method at the calibration point in the patient, and DMC represents absorbed dose from MC simulations at the calibration point in the patient. In practice, we used differences in dose profiles to estimate DTPS and DMC for patients 17 and 18. These profiles were generated using the PBA35 in the TPS and the Monte Carlo Proton Radiotherapy Treatment Planning (MCPRTP) code44, 45. In general, Contemporary proton PBAs in the TPS system provide excellent accuracy44, especially in homogeneous media, superior to that of broad beam algorithms in heterogeneous media. The improvement in accuracy comes mostly at the cost of greater computation times. Because of inherent approximations in the PBA, it may not provide sufficient accuracy in extremely heterogeneous media, at material and/or density interfaces, or in other complex situations.

The MCPRTP used the Monte Carlo N-particle eXtended radiation transport code46 with parallel processing as a radiation dose calculation engine. Each component of the proton treatment unit was modeled in detail and patient's CT images were converted to voxelized phantom in the MCNPX code. The accuracy of the Monte Carlo simulation model has been previously evaluated by Titt et al.47 More details of Monte Carlo simulations can be found in previous reports from our group.44, 48, 49, 50

DMC was considered to provide the best estimate of the true absorbed dose at the calibration point. This distinction was made for two reasons. (1) The MC simulation model used more complete and realistic physics models to describe multiple coulomb scattering and nuclear interactions.27 (2) The MC simulations took into account the variations in elemental composition and mass density of various tissues, whereas the TPS approximated all tissues as water of varying density.45 Input data to both algorithms (i.e., the CT data set, aperture, RC, and beam line parameters) were identical, so differences in predicted dose distributions were attributed solely to differences in the pencil beam and Monte Carlo dose algorithms.

2.5.2 Estimation of uncertainty in FCSPS from the TPS+M method

When FCSPS was determined using the TPS+M method, i.e., TPS dose estimates and measurements (section 2.4.2), the corresponding uncertainty in FCSPS, (μFCSPS)TPS+M, was estimated using a statistical approach: differences between FCSPS values generated with the TPS+M and TPS methods were used to estimate upper and lower bounds of (μFCSPS)M. The lower bound of the absolute uncertainty in FCSPS,, was estimated according to

where,are the mean FCSPS values from the TPS and TPS+M methods, respectively. The value was averaged over 32 fields (patients 1-16), and the   value was averaged over 4 fields (patients 17 and 18). Mean values were used because it was assumed that most random variations in the data would be averaged out of the respective data sets, so the resulting difference would represent a clinically representative estimate of the differences between calculation methods, i.e., a lower bound for the true uncertainty in FCSPS.

The upper bound of uncertaintyin FCSPS, denoted by , was estimated from the maximum absolute values of difference in paired FCSPS values, or

where, the subscripts xi indicate that the TPS and TPS+M calculations were performed for each individual field in the sample, i.e., patients 17 and 18 (see section 2.4.1).  Because was calculated for each field individually, it represented the differences in the TPS and TPS+M calculation methods solely. With this approach, we avoided confounding factors such as inter-patient differences in anatomy or treatment design that would have occurred otherwise.

2.5.3 Estimation of uncertainty in FCSPS from historical methods

When the historical value of 1.0 was used for FCSPS, the corresponding uncertainty in FCSPS, (μFCSPS)HIST, was estimated according to,

where, FCSPS(HIST) is 1.0 and FCSPS(TPS) was determined with the TPS method. This estimation was done for patients 1 through 16.

3. Results

3.1 Estimation of FCSPS using the TPS method

Table 4 lists descriptive statistics that compare FCSPS values from the two data sets studied here (patients 1-16 and 17-18). The values for the two data sets were not significantly different from one another.

3.2 Estimation of FCSPS using the TPS+M method

Table 4 reveals good agreement between values of FCSPS calculated with the TPS method and those calculated with the TPS+M method. The standard deviation in FCSPS for the TPS method was smaller than the corresponding value from the TPS+M method. Also, there was a larger interval in values for the TPS+M method than for the TPS method. There are several possible explanations for this. One is that additional statistical uncertainty was introduced by the measurements. Another is that the measurements better indicate the true standard deviation, while the TPS method artificially smooths out some of the true variation.

3.3 Estimation of uncertainty in FCSPS

3.3.1 Uncertainty in FCSPS using the TPS method

Absorbed dose predictions from MC and pencil beam algorithms were compared for a prostate treatment in one of the patients (patient 17) following the methods of 2.5.1. Differences in estimates of absorbed dose between MC simulations and the TPS method resulted in a 10.5 cGy difference between profiles at isocenter for the right lateral field and a 4.7 cGy difference for the left lateral field. The absorbed doses from the right and left lateral fields for the treatment plan at isocenter, which were used to represent Dw as an approximation, were 3550 cGy and 3570 cGy, respectively. Therefore, the (μFCSPS)TPS was less than 0.3% for the individual fields. These estimates suggest that the contribution of μFCSPS to uncertainty in (D/MU)p for this particular patient was negligible.

3.3.2 Uncertainty in FCSPS using the TPS+M method

Following the methods in section 2.5.2, we estimated the upper and lower bounds for uncertainty in FCSPS. The lower bound,, was 0.004 and the upper bound,, was 0.008. 

3.3.3 Uncertainty in FCSPS using historical estimate

Estimates of FCSPS done using the TPS method were compared to the historical value of 1.0 for FCSPS. Descriptive statistics for this analysis are listed in Table 5. The (μFCSPS)HIST was less than 0.029, and the mean (μFCSPS)HIST was 0.006.

4. Discussion

We estimated uncertainty in (D/MU)p for patients receiving proton therapy for cancer of the prostate. In particular, we compared estimates of uncertainty in FCSPS by means of measurements, MC simulations, and pencil beam dose calculations. Our results confirm that when FCSPS is included in the estimation of (D/MU)p,, the uncertainty in (D/MU)p is less than 5%, regardless of the method used to calculate FCSPS.

Our findings on FCSPS are similar to those of Newhauser22,51 and Sahoo et al.30 At the outset of this work, the standard of care at our institution followed the approach described by Newhauser22 in which FCSPS was taken as unity.  Subsequently Newhauser et al.51 for prostate treatment fields was near the historical value of 1.0 based on results of a phantom study. Likewise, Sahoo et al.30 reported that FCSPS spanned the interval 0.957 to 1.089 with a mean value of 1.00 ± 0.04. However, we note that Sahoo et al.30 did not specify which anatomical treatment sites these values were from. Therefore, it is difficult to make a direct comparison of those results with the findings from this work. 

As noted in section 2.3, we found no directly comparable reports on μFCSPS in the literature for prostate treatment. Our findings on μFCSPS differ from those in the works most similar to ours, that is, reports from Akagi et al.24 and Sahoo et al.30 Akagi et al.24 reported a relative uncertainty of 3.6% in the measured value of FCSPS; this value exceeds ours by approximately a factor of 10. However, the uncertainty reported by Akagi et al.24 took into account the errors in reproducing collimator scatter (FColS) and inaccuracy in analytical modeling of patient anatomy which were not applicable to our study.

One of the clinical implications of this work is that accuracy of (D/MU)p values, in the special case of prostate treatment fields, does not depend strongly on accurate knowledge of the FCSPS factor. Thus, this work also provides an evidential basis and rationale for standardizing absolute proton dosimetry, which is a key requisite step to conduct multi-institution clinical trials.

This study had several limitations. First, it considered normal patient anatomy, e.g., the effects of implanted fiducial markers, hip prostheses, and organ motion on (D/MU)p and its uncertainty were not included. However, these are not serious limitations because solutions for fiducial markers are known52,53,54,55, and although hip prostheses are relatively rare, there are MV/kV imaging solutions that are available to correct for their effects56. Second, we considered only the lateral opposed-pair treatment technique, while involvement of pelvic lymph nodes or treatment of other more complex treatment strategies57,58 would require a much different application of our present findings. Third, our findings are specific to the treatment planning and delivery systems in use at our institution for a passively scattered treatment; there may be additional differences between this study and other treatment techniques, such as intensity-modulated proton therapy. Nonetheless, the methods and results of this study may serve as a qualitative guide for similar studies of other proton therapy systems.

Given the complexities and uncertainties associated with estimation of absorbed dose in the patient, additional studies are needed to test whether the findings of this work will hold for other anatomical sites. In our laboratory, additional studies are now under way to address estimation of uncertainties in (D/MU)p in the thorax.

5. Conclusion

In conclusion, our study investigated the water-to-patient absorbed dose conversion factor, FCSPS, one of the least well-understood factors in proton output calculation, and found that mean FCSPS value was 1.006 and uncertainty in FCSPS was approximately 1%, suggesting that uncertainty in FCSPS for proton therapy of prostate cancer is clinically acceptable.

Conflict of Interest

The authors declare that they have no conflicts of interest. The authors alone are responsible for the content and writing of the paper.


  1. International Commission on Radiation Units and Measurements (ICRU) 2007 Prescribing, Recording, and Reporting Proton-Beam Therapy. ICRU Report 78 (Bethesda, MD: ICRU).
  2. Durante M, Loeffler JS. Charged particles in radiation oncology. Nat Rev Clin Oncol. 2010; 7:37-43. [CrossRef]
  3. PTCOG, 2014 PTCOG patient statistics of particle therapy centers  in 2013.
  4. Sisterson J. Ion beam therapy in 2004. Nucl Instr Meth B. 2005;241: 713-6. [CrossRef]
  5. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA: Cancer J Clin. 2012; 62:10-29. [CrossRef]
  6. Slater JD, Rossi CJ Jr, Yonemoto LT, et al. Proton therapy for prostate cancer: the initial Loma Linda University experience. Int J Radiat Oncol Biol Phys. 2004;59: 348-52. [CrossRef]
  7. Rossi CJ Jr. Conformal proton beam radiation therapy for prostate cancer: concepts and clinical results. Commun Oncol. 2007; 4: 235-50. [CrossRef]
  8. Mayahara H, Murakami M, Kagawa K, et al. Acute morbidity of proton therapy for prostate cancer: the Hyogo Ion Beam Medical Center experience. Int J Radiat Oncol Biol Phys 2007;69:434-43. [CrossRef]
  9. Vargas C, Mahajan C, Fryer A, et al. Rectal dose-volume differences using proton radiotherapy and a rectal balloon or water alone for the treatment of prostate cancer. Int J Radiat Oncol Biol Phys. 2007;69:1110-6. [CrossRef]
  10. Coen JJ, Zietman AL. Proton radiation for localized prostate cancer. Nat Rev Urol. 2009;6: 324-30. [CrossRef]
  11. Chera BS, Vargas C, Morris CG, et al. Dosimetric study of pelvic proton radiotherapy for high-risk prostate cancer. Int J Radiat Oncol Biol Phys. 2009;75: 994-1002. [CrossRef]
  12. Mendenhall NP, Li Z, Hoppe BS, et al. Early outcomes from three prospective trials of image-guided proton therapy for prostate cancer. Int J Radiat Oncol Biol Phys. 2012;82:213-21. [CrossRef]
  13. Sheets NC, Goldin GH, Meyer AM, et al. Intensity-modulated radiation therapy, proton therapy, or conformal radiation therapy and morbidity and disease control in localized prostate cancer. JAMA. 2012;307:1611-20. [CrossRef]
  14. Zietman AL, DeSilvio ML, Slater JD, et al. Comparison of conventional-dose vs high-dose conformal radiation therapy in clinically localized adenocarcinoma of the prostate: a randomized controlled trial. JAMA. 2005; 294:1233-9. [CrossRef]
  15. Zietman AL, Bae K, Slater JD, et al. Randomized trial comparing conventional-dose with high-dose conformal radiation therapy in early-stage adenocarcinoma of the prostate: long-term results from proton radiation oncology group/american college of radiology 95-09. J Clin Oncol 2010;28:1106-11. [CrossRef]
  16. Olsen DR, Bruland OS, Frykholm G, Norderhaug IN. Proton therapy - a systematic review of clinical effectiveness. Radiother Oncol. 2007; 83:123-32. [CrossRef]
  17. Terasawa T, Dvorak T, Ip S, et al. Systematic review: charged-particle radiation therapy for cancer. Ann Intern Med. 2009;151:556-65. [CrossRef]
  18. AAPM. Report 16 – Protocol for heavy charged-particle therapy beam dosimetry. New York, NY: AIP - American Association of Physicists in Medicine, Task Group 20, 1986.
  19. Vynckier S, Bonnett DE, Jones DTL. Code of practice for clinical proton dosimetry. Radiother Oncol. 1991;20:53-63. [CrossRef]
  20. Vynckier S, Bonnett D E, Jones DTL. Supplement to the code of practice for clinical proton dosimetry. Radiother Oncol. 1994;32:174-79. [CrossRef]
  21. International Atomic Energy Agency (IAEA). Absorbed dose determination in external beam radiotherapy. IAEA Technical Report Series 398, 2000 (Vienna: IAEA).
  22. Newhauser W. Dosimetry for the gantry beams at the Northeast Proton Therapy Center Part I: Dimensions and geometric relationships. Massachusetts General Hospital, Report HD-112, 2001.
  23. Kooy HM, Rosenthal SJ, Engelsman M, et al. The prediction of output factors for spread-out proton Bragg peak fields in clinical practice. Phys Med Biol. 2005;50:5847-56. [CrossRef]
  24. Akagi T, Kanematsu N, Takatani Y, et al. Scatter factors in proton therapy with a broad beam. Phys Med Biol. 2006;51: 1919-28. [CrossRef]
  25. Paganetti H. Monte Carlo calculations for absolute dosimetry to determine machine outputs for proton therapy fields. Phys Med Biol. 2006;51:2801-12. [CrossRef]
  26. Fontenot JD, Newhauser WD, Bloch C, et al. Determination of output factors for small proton therapy fields. Med Phys. 2007;34:489-98. [CrossRef]
  27. Herault J, Iborra N, Serrano B, Chauvel P. Spread-out Bragg peak and monitor units calculation with the Monte Carlo code MCNPX. Med Phys. 2007; 34:680-8. [CrossRef]
  28. Titt U, Zheng Y, Vassiliev ON, Newhauser WD. Monte Carlo investigation of collimator scatter of proton-therapy beams produced using the passive scattering method. Phys Med Biol. 2008; 53: 487-504. [CrossRef]
  29. Koch N, Newhauser WD, Titt U, et al. Monte Carlo calculations and measurements of absorbed dose per monitor unit for the treatment of uveal melanoma with proton therapy. Phys Med Biol. 2008;53:1581-94. [CrossRef]
  30. Sahoo N, Zhu X, Arjomandy B, et al. A procedure for calculation of monitor units for passively scattered proton radiotherapy beams. Med Phys. 2008;35:5088-97. [CrossRef]
  31. Koch N , Newhauser WD. Development and verification of an analytical algorithm to predict absorbed dose distributions in ocular proton therapy using Monte Carlo simulations. Phys Med Biol. 2010;55:833-53. [CrossRef]
  32. Moyers MF, Vatnitsky SM. Practical implementation of light ion beam treatments.  Medical Physics Publishing, Madison, WI, 2012.  ISBN 978-1-930524-55-2.
  33. Siebers JV, Keall PJ, Nahum AE, Mohan R. Converting absorbed dose to medium to absorbed dose to water for Monte Carlo based photon beam dose calculations. Phys Med Biol. 2000;45:983-95. [CrossRef]
  34. Paganetti H. Dose to water versus dose to medium in proton beam therapy. Phys Med Biol. 2009;54:4399-421. [CrossRef]
  35. Schaffner B, Pedroni E, Lomax A. Dose calculation models for proton treatment planning using a dynamic beam delivery system: an attempt to include density heterogeneity effects in the analytical dose calculation. Phys Med Biol. 1999; 44:27-41. [CrossRef]
  36. Mohr PJ, Taylor BN. CODATA recommended values of the fundamental physical constants: 1998. Rev Mod Phys 2000;72:351-495. [CrossRef]
  37. International Commission on Radiation Units and Measurements (ICRU). Clinical Proton Dosimetry Part I: Beam production, beam delivery and measurement of absorbed dose. ICRU Report 59, 1998 (Bethesda, MD: ICRU).
  38. Medin J, Andreo P, Vynckier S. Comparison of dosimetry recommendations for clinical proton beams. Phys Med Biol. 2000;45:3195-211. [CrossRef]
  39. Newhauser W, Myers K, Rosenthal S, Smith AR. Proton beam dosimetry for radiosurgery: implementation of the ICRU Report 59 at the Harvard Cyclotron Laboratory. Phys Med Biol. 2002;47:1369-89. [CrossRef]
  40. Newhauser WD, Burns J, Smith AR. Dosimetry for ocular proton beam therapy at the Harvard Cyclotron Laboratory based on the ICRU Report 59.  Med Phys. 2002;29:1953-61. [CrossRef]
  41. International Commission on Radiation Units and Measurements (ICRU). Stopping powers and ranges for proton and alpha particles. ICRU Report 49, 1993 (Bethesda, MD: ICRU).
  42. Lunsford TK, Lunsford BR. The research sample, Part I: Sampling. J Prosthet Orthot 1995;7:105-12. [CrossRef]
  43. Newhauser W. Procedure for creating a verification plan in the Eclipse treatment planning system for patients receiving proton therapy for cancer of the prostate, Report MDACC-10-0043, University of Texas MD Anderson Cancer Center, Houston, TX, 2007.
  44. Newhauser W, Fontenot J, Zheng Y, et al. Monte Carlo simulations for configuring and testing an analytical proton dose-calculation algorithm. Phys Med Biol. 2007;52:4569-84. [CrossRef]
  45. Newhauser W, Zheng Y, Taddei P, et al. Monte Carlo proton radiation therapy planning calculations Trans. Am Nucl  Soc. 2008;99:63.
  46. Hendricks J, McKinney G, Durkee J, et al. MCNPX, Version 26c (Los Alamos, NM: Los Alamos National Laboratory) 2006.
  47. Titt U, Sahoo N, Ding X, et al. Assessment of the accuracy of an MCNPX-based Monte Carlo simulation model for predicting three-dimensional absorbed dose distributions. Phys Med Biol. 2008;53:4455-70. [CrossRef]
  48. Taddei PJ, Mirkovic D, Fontenot JD, et al. Stray radiation dose and second cancer risk for a pediatric patient receiving craniospinal irradiation with proton beams. Phys Med Biol 2009;54:2259-75. [CrossRef]
  49. Zhang R, Fontenot J, Mirkovic D, et al. Advantages of MCNPX-based lattice tally over mesh tally in high-speed Monte Carlo dose reconstruction for proton radiotherapy. Nucl Technol. 2013; 183:101-6. [CrossRef]
  50. Newhauser W, Zhang R. The physics of proton therapy. Phys. Med. Biol. 2015;60:R155-209. [CrossRef]
  51. Newhauser W, Zhu R, Sahoo N, et al. Small corrections to measured D/MU values for patients receiving proton therapy for cancer of the prostate, Report MDACC-10-0065, University of Texas MD Anderson Cancer Center, Houston, TX, 2007.
  52. Newhauser W, Fontenot J, Koch N, et al. Monte Carlo simulations of the dosimetric impact of radiopaque fiducial markers for proton radiotherapy of the prostate. Phys Med Biol. 2007;52:2937-52. [CrossRef]
  53. Giebeler A, Fontenot J, Balter P, et al. Dose perturbations from implanted helical gold markers in proton therapy of prostate cancer. J Appl Clin Med Phys. 2009;10:2875. [CrossRef]
  54. Cheung J, Kudchadker RJ, Zhu XR, et al. Dose perturbation and image artifacts caused by carbon-coated ceramic and stainless steel fiducials used in proton therapy for prostate cancer. Phys Med Biol. 2010; 55:7135-47. [CrossRef]
  55. Huang JY, Newhauser WD, Zhu XR, et al. Investigation of dose perturbations and the radiographic visibility of potential fiducials for proton radiation therapy of the prostate. Phys Med Biol. 2011;56:5287-302. [CrossRef]
  56. Newhauser WD, Giebeler A, Langen KM, et al. Can megavoltage computed tomography reduce proton range uncertainties in treatment plans for patients with large metal implants? Phys Med Biol. 2008;53:2327-44. [CrossRef]
  57. Rechner LA, Howell RM, Zhang R, et al. Risk of radiogenic second cancers following volumetric modulated arc therapy and proton arc therapy for prostate cancer. Phys Med Biol. 2012;57:7117-32. [CrossRef]
  58. Rechner LA, Howell RM, Zhang R, Newhauser WD. Impact of margin size on the predicted risk of radiogenic second cancers following proton arc therapy and volumetric modulated arc therapy for prostate cancer. Phys Med Biol. 2012;57:N469-79. [CrossRef]

Submission: May 11, 2015; First Revision: June 17, 2015; Acceptance: June 24, 2015; Publication: August 22, 2015

*Corresponding author: : Wayne Newhauser, PhD; Mary Bird Perkins Cancer Center and Louisiana State University, Baton Rouge, LA, 70803, USA.

© Newhauser. Published by EJourPub. Journal of Proton Therapy. All rights reserved.

Copyright (c) 2015 Wayne D. Newhauser

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.


© Journal of Proton Therapy (ISSN 2469-5491)

To make sure that you can receive messages from us, please add the 'protonjournal.org' and 'protonjournal.com' domains to your e-mail 'safe list'. If you do not receive e-mail in your 'inbox', please check your 'bulk mail' or 'junk mail' folders.