Optimization of skin dose using in-vivo MOSFET dose measurements in bolus/non-bolus fraction ratio: A VMAT and a 3DCRT study.

In-phantom and in-vivo three dimensional conformal radiation therapy (3DCRT) and volumetric modulated arc therapy (VMAT) skin doses, measured with and without bolus in a female anthropomorphic phantom RANDO and in patients, were compared against treatment planning system calculated values. A thorough characterization of the metal oxide semiconductor field effect transistor measurement system was performed prior to the measurements in phantoms and patients. Patients with clinical indication for postoperative external radiotherapy were selected. Skin dose showed higher values with 3DCRT technique compared with VMAT. The increase in skin dose due to the use of bolus was quantified. It was observed that, in the case of VMAT, the bolus effect on the skin dose was considerable when compared with 3DCRT. From the point of view of treatment time, bolus cost, and positioning reproducibility, the use of bolus in these situations can be optimized.

and, on the other hand, required doses should be high enough to avoid tumor recurrence. 3,4,6,[10][11][12] Dose prescription and dosimetry assessment in the skin is not a straightforward task due to the limitation of treatment planning system (TPS) when calculating dose at the surface and at different structures of the skin (basal and dermal layers), as depth and location vary between patients and even in the same patient. The majority of available dose calculation algorithms are not sufficiently accurate to compute dose distributions in superficial and build-up regions where electron equilibrium is not established. Generally, TPS can calculate skin dose (build-up region) with accuracy up to 20%. [11][12][13][14][15] Different methodologies for skin dose measurement and detectors characterization have been reported in the literature: radiochromic films; thermoluminescent dosimeters (TLD), and other passive solid state dosimeters; diodes; metal oxide semiconductor field effect transistors (MOSFETs)-based dosimeter, etc. 7,13, The purpose of this work is to assess the skin doses in breast cancer treatments obtained with volumetric modulated arc therapy (VMAT) and three dimensional conformal radiation therapy (3DCRT) treatment techniques, using in-vivo dosimetry, and compare the measured values with the doses calculated by the TPS.

2.A | MOSFET calibration
In this study, dual-bias TN502RD MOSFET detectors fabricated by Thomson & Nielsen, Canada, were used. Before performing the skin dose measurements, the MOSFET detectors were calibrated using 4 and 6 MV photon beams using a Varian 2100C/D and a 2300IX linac

2.B | MOSFET measurement system characterization methodology
A thorough characterization of the MOSFET-based dosimetry system was performed regarding linearity, reproducibility, and angular dependence for two photon energies commonly used in radiotherapy treatments of the breast (4 and 6 MV). All measurements, with the exception of the angular dependence, were performed using the standard setup above described for the calibration. The angular dependence was performed using a spherical phantom of 14 cm in diameter (Lucy 3D QA phantom -Standard Imaging, Middleton, USA) acquiring measurements with several beam incidences (45 degree increments; Fig. 2). F I G . 1. System calibration setup. The ionization chamber positioned at a depth of 5 cm and the MOSFET detectors at Z max depth for the specific energy.
F I G . 2. Angular dependence measurement setup, with Lucy 3D quality assurance phantom.
To assess the variation of the dosimeters response with dose, three MOSFET located at the build-up position of each energy were irradiated in the range of 10 to 350 cGy with SSD = 100 cm. As a control measure (cross calibration), in the same procedure dose measurements were performed using the ionization chamber positioned at a 5 cm depth.
To assess the angular dependence, a MOSFET was placed in the center of the spherical phantom at the isocenter and a series of measurements was carried out by varying the rotation angle of the gantry. The MOSFET was positioned so that its flat side was facing the beam, with the gantry in the 0°position. The gantry angle varied between 0 and 315°in 45°increments.

2.C | In-phantom measurements methodology
The in-phantom measurements were performed using an anthropomorphic female RANDO phantom without breasts to simulate a postmastectomy patient. A CT scan of the phantom was acquired and four measurement points were selected on the phantom surface as seen in The clinical importance of measuring dose in this area is due to the relative high probability of future relapse. [6][7][8][9] The fourth point (P 4 ) was referenced on the axillar area, corresponding to the beam axis projection in the surface. The maximum sensitivity points of the MOSFET detectors were accurately placed on points P 1 -P 4 and measurements were performed with and without bolus.
The output from the TPS calculation on these points was later compared with the MOSFET measured doses. For this purpose, we used the TPS version 13.5 Eclipse ® Varian (Varian Medical Systems, Inc., Palo Alto, CA, USA) using the Anisotropic Analytical Algorithm (AAA) with a calculation grid of 2.5 mm. 3DCRT treatment plans were performed in the RANDO phantom for the breast area, using two oblique opposing tangential fields in field-in field technique with 4 MV photon beams. The VMAT plans were performed with 6 MV using two partial arcs, in the right side from 60°to 181°and in the left side from 340°to 179°.
A 3DCRT and a VMAT plan similar to a clinical case using 4 and 6 MV, respectively, was performed, with and without a bolus (Super-Flab, Eckert & Ziegler, 1 cm thick) for surface dose enhancement.
The prescription was 2 Gy (average dose in the PTV) per fraction for a total dose of 50 Gy in 25 fractions. In the TPS, the bolus was added, based on previous acquires CT unit calibration and no extra CT was performed to the patient.

3.A.2 | Angular dependence
The results are presented in the Fig. 6, for both studied energies.
The standard deviations (%σ) obtained from all the considered directions were 1.86% and 1.67%, for 4 and 6 MV respectively, which are comparable within the ±2% variation over 360°as stated by the manufacturer.

3.B | Phantom dose measurement
The results of the 3DCRT and VMAT plan irradiation of the female anthropomorphic phantom (D M ) were compared with TPS calculated values (D TPS ) with and without bolus.

3.C | Patient dose measurement
The obtained difference using the Eq. (1)

4.B | Patient doses
In both VMAT and 3DCRT techniques, it is also evident from  the bolus thickness in the cases where an increase or a decrease of surface dose is desirable.

| CONCLUSIONS
From the measurements made in the female anthropomorphic phantom RANDO, considering the treatment plan performed in the TPS, and without any correction factors, there was some discrepancy between the measured and the calculated values by the TPS, with more evidence for plan without bolus. However, this difference was within the ±20% error range, the value referred in the AAPM-TG 53 for the TPS calculation imprecision in the buildup region. These measurements also demonstrate that the surface dose increased in the presence of the bolus when considering VMAT and 3DCRT. This increased surface dose is clearly higher in the VMAT technique. Since the treatment plan is very similar to that performed for treatment of patients with breast carcinoma, measurements using the RANDO phantom seem to indicate an easy and straightforward method of verifying surface dose (in-vivo) applicable in actual clinical situations.
It should be noted that this method may be used in patients with other pathologies, always taking into account the associated error.

ACKNOWLEDGMENTS
The authors would like to acknowledge the radiotherapy technologists for their availability during in-vivo measurements throughout the treatment time.

CONF LICT OF I NTEREST
No conflicts of interest.