Dosimetric evaluation of a MOSFET detector for clinical application in photon therapy.
ABSTRACT Dosimetric characteristics of a metal oxide-silicon semiconductor field effect transistor (MOSFET) detector are studied with megavoltage photon beams for patient dose verification. The major advantages of this detector are its size, which makes it a point dosimeter, and its ease of use. In order to use the MOSFET detector for dose verification of intensity-modulated radiation therapy (IMRT) and in-vivo dosimetry for radiation therapy, we need to evaluate the dosimetric properties of the MOSFET detector. Therefore, we investigated the reproducibility, dose-rate effect, accumulated-dose effect, angular dependence, and accuracy in tissue-maximum ratio measurements. Then, as it takes about 20 min in actual IMRT for the patient, we evaluated fading effect of MOSFET response. When the MOSFETs were read-out 20 min after irradiation, we observed a fading effect of 0.9% with 0.9% standard error of the mean. Further, we applied the MOSFET to the measurement of small field total scatter factor. The MOSFET for dose measurements of small field sizes was better than the reference pinpoint chamber with vertical direction. In conclusion, we assessed the accuracy, reliability, and usefulness of the MOSFET detector in clinical applications such as pinpoint absolute dosimetry for small fields.
- [Show abstract] [Hide abstract]
ABSTRACT: Radiation therapy in patients is planned by using computed tomography (CT) images acquired before start of the treatment course. Here, tumor shrinkage or weight loss or both, which are common during the treatment course for patients with head-and-neck (H&N) cancer, causes unexpected differences from the plan, as well as dose uncertainty with the daily positional error of patients. For accurate clinical evaluation, it is essential to identify these anatomical changes and daily positional errors, as well as consequent dosimetric changes. To evaluate the actual delivered dose, the authors proposed direct dose measurement and dose calculation with mega-voltage cone-beam CT (MVCBCT). The purpose of the present study was to experimentally evaluate dose calculation by MVCBCT. Furthermore, actual delivered dose was evaluated directly with accurate phantom setup. Because MVCBCT has CT-number variation, even when the analyzed object has a uniform density, a specific and simple CT-number correction method was developed and applied for the H&N site of a RANDO phantom. Dose distributions were calculated with the corrected MVCBCT images of a cylindrical polymethyl methacrylate phantom. Treatment processes from planning to beam delivery were performed for the H&N site of the RANDO phantom. The image-guided radiation therapy procedure was utilized for the phantom setup to improve measurement reliability. The calculated dose in the RANDO phantom was compared to the measured dose obtained by metal-oxide-semiconductor field-effect transistor detectors. In the polymethyl methacrylate phantom, the calculated and measured doses agreed within about +3%. In the RANDO phantom, the dose difference was less than +5%. The calculated dose based on simulation-CT agreed with the measured dose within±3%, even in the region with a high dose gradient. The actual delivered dose was successfully determined by dose calculation with MVCBCT, and the point dose measurement with the image-guided radiation therapy procedure.Medical dosimetry: official journal of the American Association of Medical Dosimetrists 12/2012; · 1.26 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We have developed a practical dose verification method for radiotherapy treatment planning systems by using only a Farmer ionization chamber in inhomogeneous phantoms. In particular, we compared experimental dose verifications of multi-layer phantom geometries and laterally inhomogeneous phantom geometries for homogeneous and inhomogenous dose calculations by using the fast-Fourier-transform convolution, fast-superposition, and superposition in the XiO radiotherapy treatment-planning system. We applied the dose verification method to three kernel-based algorithms in various phantom geometries with water-, lung- and bone-equivalent media of different field sizes. These calculations were then compared with experimental measurements by use of the Farmer ionization chamber. The fast-Fourier-transform convolution algorithm overestimated the dose by about 8% in the lung phantom geometry. The superposition algorithm and the fast-superposition algorithm were both accurate to better than 2% when compared to the measurements even for complex geometries. Our dose verification method was able to clarify the differences and equivalences of the three kernel-based algorithms and measurements with use only of commonly available apparatus. This will be generally useful in commissioning of inhomogeneity-correction algorithms in the clinical practice of treatment planning.Radiological Physics and Technology 01/2009; 2(1):87-96.
- [Show abstract] [Hide abstract]
ABSTRACT: For reducing the risk of skin injury during interventional radiology (IR) procedures, it has been suggested that physicians track patients' exposure doses. The metal-oxide semiconductor field effect transistor (MOSFET) dosimeter is designed to measure patient exposure dose during radiotherapy applications at megavoltage photon energies. Our purpose in this study was to evaluate the feasibility of using a MOSFET dosimeter (OneDose system) to measure patients' skin dose during exposure to diagnostic X-ray energies used in IR. The response of the OneDose system was almost constant at diagnostic X-ray energies, although the sensitivity was higher than that at megavoltage photon energies. We found that the angular dependence was minimal at diagnostic X-ray energies. The OneDose is almost invisible on X-ray images at diagnostic energies. Furthermore, the OneDose is easy to handle. The OneDose sensor performs well at diagnostic X-ray energies, although real-time measurements are not feasible. Thus, the OneDose system may prove useful in measuring patient exposure dose during IR.Radiological Physics and Technology 01/2009; 2(1):58-61.
Dosimetric evaluation of a MOSFET detector for clinical
application in photon therapy
Ryosuke Kohno Æ Æ Eriko Hirano Æ Æ Teiji Nishio Æ Æ Tomoko Miyagishi Æ Æ
Tomonori Goka Æ Æ Mitsuhiko Kawashima Æ Æ Takashi Ogino
Received: 5 July 2007/Revised: 19 October 2007/Accepted: 23 October 2007/Published online: 17 November 2007
? Japanese Society of Radiological Technology and Japan Society of Medical Physics 2007
silicon semiconductor field effect transistor (MOSFET)
detector are studied with megavoltage photon beams for
patient dose verification. The major advantages of this
detector are its size, which makes it a point dosimeter,
and its ease of use. In order to use the MOSFET detector
for dose verification of intensity-modulated radiation
therapy (IMRT) and in-vivo dosimetry for radiation
therapy, we need to evaluate the dosimetric properties of
the MOSFET detector. Therefore, we investigated the
reproducibility, dose-rate effect, accumulated-dose effect,
angular dependence, and accuracy in tissue-maximum
ratio measurements. Then, as it takes about 20 min in
actual IMRT for the patient, we evaluated fading effect of
MOSFET response. When the MOSFETs were read-out
20 min after irradiation, we observed a fading effect of
0.9% with 0.9% standard error of the mean. Further, we
Dosimetric characteristics of a metal oxide–
applied the MOSFET to the measurement of small field
total scatter factor. The MOSFET for dose measurements
of small field sizes was better than the reference pinpoint
chamber with vertical direction. In conclusion, we asses-
sed the accuracy, reliability, and usefulness of the
MOSFET detector in clinical applications such as pinpoint
absolute dosimetry for small fields.
Megavoltage radiation therapy ? Pinpoint dosimetry ?
Dose verification ? Small field
MOSFET detector ?
The accurate delivery of radiation in radiation therapy is
very important. Particularly, since an intensity-modulated
radiation therapy (IMRT) is maximizing tumor control and
minimizing normal-tissue complications, an accurate esti-
mation of the radiation dose in the patient through phantom
measurement is important for verifying that the expected
dose of radiation has been delivered. Therefore, we have to
measure and verify the photon dose under various clinical
conditions such as small fields, heterogeneities, irradiations
of irregular field and IMRT, which require pinpoint
A metal oxide–silicon semiconductor field effect tran-
sistor (MOSFET) detector has been widely used for dose
measurements in radiation therapy [1–4]. Furthermore,
Gurp et al.  reported the clinical implementation of the
MOSFET in electron beams, and Kohno et al.  evaluated
the feasibility of the MOSFET for proton dose measure-
ments. The major advantage of the MOSFET detector is
that it provides a direct reading with a very small active
area (0.04 mm2). The physical size of the packaged
R. Kohno (&) ? T. Nishio ? T. Miyagishi ? T. Goka ?
M. Kawashima ? T. Ogino
Particle Therapy Division, Research Center for Innovative
Oncology, National Cancer Center Hospital East,
6-5-1, Kashiwanoha, Kashiwa-shi, Chiba 277-8577, Japan
Department of Radiology, National Cancer Center Hospital East,
Chiba 277-8577, Japan
Department of Nuclear Engineering and Management,
Graduate School of Engineering, University of Tokyo,
2-11-16 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan
Department of Radiology, Graduate School of Medicine,
University of Tokyo, 7-3-1 Hongo, Bunkyo-ku,
Tokyo 113-8655, Japan
Radiol Phys Technol (2008) 1:55–61
MOSFET is 2.0 9 1.3 9 8.0 mm. The small size of the
detector could be advantageous for point dose measure-
ments. Other features are a direct and simple dose read-out
compared to a thermoluminescent dosimeter, which
requires pre-preparation and post-processing, the portabil-
ity of the system, and nonvolatility of the accumulated dose
of each MOSFET.
The direct dose reading from the gate-threshold voltage
involves uncertainties from uncorrected systematics of
accumulated-dose effects and angular dependence in
dosimeter placement [2–4, 7, 8]. Chuang et al.  reported
that the linearity and the angular dependence of the
MOSFET detector are about ±2 to 3% and ±2.5%. Thus,
the uncertainties of the MOSFET in photon dosimetry are
rather large. Such uncertainties need to be quantitatively
assessed in the commissioning process before clinical
The purpose of this study is to evaluate the basic char-
acteristicsof the MOSFET
reproducibility, dose-rate effect, accumulated-dose effect,
fading effect, angular dependence, accuracy in tissue-
maximum ratio, and total scatter factor measurements in
order to assess the accuracy, reliability, and usefulness in
clinical applications such as pinpoint absolute dosimetry
for small fields.
2 Materials and methods
2.1 MOSFET detector
Basically, a MOSFET consists of a P-type silicon semi-
conductor substrate, a layer of insulating oxide, and a metal
gate. Ionizing radiation generates electron-hole pairs in the
insulating layer and some of the holes drift toward the
substrate under appropriate bias voltage and get trapped
there semipermanently, which shifts the gate threshold
voltage for the source–drain conductivity. After exposure,
the gate threshold voltage can be repeatedly measured by
applying constant source–drain current, with which the
accumulated dose is read.
All measurements were performed with TN502RD
MOSFET detectors, produced by Thomson & Nielsen,
Canada. The TN502RD MOSFET detector is a dual bias
dual MOSFET detector that consists of two identical
MOSFETs fabricated on the same silicon chip and oper-
ating at two different gate biases . Here, the sensitivity
(mV/cGy) of a single MOSFET detector obviously
decreases as a function of accumulated radiation dose. The
sensitivity for individual MOSFET biased at 5, 10, and
15 V are parallel, indicating that the magnitude of the
sensitivity difference of two different, biased MOSFET, is
constant. Finally, the dualism of MOSFET achieves better
sensitivity and reproducibility within ±3% of that of the
single MOSFET detector. The sensitive volume is
0.2 9 0.2 mm cross-sectional and 0.5 lm thick, gate-
insulating layer of silicon dioxide.
The system consists of MOSFET detectors and a mo-
bileMOSFET. The mobileMOSFET has a software to
perform all dose data measurements and a small reader
module which includes reader, dual bias supply settings,
and ports for five MOSFETs. The PC is on-line with the
reader module and MOSFETs. The dose of the MOSFET is
obtained easily. The MOSFET detector is connected to the
mobileMOSFET, which provides a choice of high or
standard sensitivity bias voltage. According to the manu-
facture’s specification, dose reproducibilities for 100 cGy
on standard and high sensitivity bias are 3 and 1.2% at a
standard deviation of 1.
2.2 Experimental apparatus
Experiments were performed with a Siemens Primus linear
accelerator (Siemens, Concord, CA, USA) with a dual-
focus, multileaf collimator. Each inner leaf projects to a
width of 1 cm at the isocenter and can travel 10 cm over
the central beam axis. The uncertainty of the leaf positions
specified by the manufacturer is within ±1 mm. For all our
experiments, we used the MOSFET TN502RD and the
standard sensitivity bias voltage supply. MOSFETs were
read-out 10 s after each irradiation in order to obtain a
stable read-out without fading effect.
In most measurements, the MOSFET and a calibrated
0.6 cc Farmer ionization chamber type 30013 (PTW,
Freiburg, Germany) were placed in a dose calibration
phantom (DC) as shown in Fig. 1a. The DC was made of
PMMA, and the thickness of the DC was 3 cm. Some solid
water (SW) RMI-457 phantoms (GAMMEX RMI, Mid-
dleton, WI, USA) were stacked on the DC (Fig. 1b). Here,
Allahverdi et al.  reported that the SW phantom is
suitable for relative dosimetry in megavoltage photon
beams, and difference of up to 1% is observed for absolute
measurements. For angular dependence measurements, we
used a cylindrical PMMA phantom with a radius of 8.0 cm
as shown in Fig. 2.
2.2.1 Reproducibility, dose-rate effect, and accumulated-
Reproducibility, dose-rate effect, and accumulated-dose
effect measurement, for MOSFET response were measured
in a 10 9 10 cm2field geometry at depths of 10 and
100 cm source to axis distance (SAD). Figure 3 shows a
setup of dose measurement using the MOSFET and the
56R. Kohno et al.
Farmer ionization chamber. Dose measurement phantom
(DC + SW) consists of the dose calibration phantom (DC)
and solid water phantom (SW). For reproducibility mea-
surements, we used 6 MV photon beams with a dose-rate
of 300 MU/min. The MOSFET was repeatedly irradiated
with 109 100 MU. The response of the MOSFET was
averaged, and the standard deviation was calculated.
Additionally, the effect of the dose rates on the MOS-
FET response was also investigated. We used 50 and
300 MU/min dose-rate for 6 MV, and 50 and 500 MU/min
for 10 MV. The MOSFET was repeatedly irradiated with
100 MU, five times for each condition. The response of the
MOSFET, and sensitivity (mV/cGy), was calculated.
Then, in accumulated-dose effect measurements, we
considered a study of dose-escalation per fraction in the
future. A 6 MV photon beam was used in a dose range
from 0 to 940 cGy. The effect of accumulated-dose on the
response of the MOSFET was investigated.
Fig. 1 a Dose calibration acrylic phantom. b Phantom setup for dose
measurements using the MOSFET detector and the Farmer ionization
Fig. 2 Cylindrical acrylic phantom with a radius of 8.0 cm
Fig. 3 Experimental setup for dose measurements
Dosimetric evaluation of MOSFET detector in photon therapy57
2.2.2 Fading effect
We also investigated the MOSFET response following the
irradiation as a function of the time difference between the
irradiation and the read-out, to evaluate the fading effect.
The fading effect was also measured in the same experi-
mental setup as that of Sect. 2.2.1. The MOSFET was
irradiated by a 6 MV photon beam with a dose-rate of
300 MU/min, and was read-out after intervals from 10 to
1,200 s to obtain the MOSFET output. This procedure was
repeated ten times for each interval.
2.2.3 Tissue-maximum ratio
The tissue-maximum ratio was measured to examine the
response of the MOSFET to changes in the radiation
spectrum. The MOSFETs were placed in a stack of SW
slabs at depths ranging from 1 to 20 cm, with 100 cm
SAD, 10 9 10 cm2field size, and were irradiated at
100 MU with 6 and 10 MV beams.
2.2.4 Angular dependence
We evaluated the angular dependence to use in vivo
dosimetry at our hospital, since we use various direc-
tional beams for each patient. The cylindrical PMMA
phantom with a radius of 8.0 cm was used for this
through 360? in the phantom. The angular MOSFET
responses along the cable axis were measured at intervals
of 45? from 0 to 360?.
2.2.5 Total scatter factors
Total scatter factor (TSCF) at 10 cm depth for a 6 MV
photon beam by the MOSFET was measured in the dose
measurement phantom (DC + SW) for field sizes from 0.4
up to 20 cm, as shown in Fig. 3. Here, since a PTW
0.015 cc pinpoint ionization chamber type 31014 (2 mm/
9 4.5 mm) with vertical direction (PPV)  is a chamber
with the highest resolution at our hospital, we register the
TSCFs with the PPV in treatment planning system. In order
to compare with the TSCF of the MOSFET, the TSCF of
the PPV was measured in water.
3 Results and discussion
3.1 Reproducibility, dose-rate effect, and accumulated-
The estimate of reproducibility for a given MOSFET was
given by the standard deviation of repeated measurements
and found to be 1.9%. However, statistically, the true
reproducibility was somewhere between 1.5 and 3.3% with
a 90% confidence level. The current study indicated that
the reproducibility was within 3%, similar to the results
obtained by Soubra et al.  and Chuang et al. . Table 1
presents the results of the dose-rate effect for MOSFET
dosimeters. No measurable effect was observed by change
of the photon dose-rate between 50 and 300 MU for 6 MV
photon beam, and 50 and 500 MU for 10 MV.
The linearity plot in Fig. 4 showed that the dose was
linear up to 940 cGy. Figure 4 shows that MOSFET indi-
cates good linearity of the response. The effect of the
accumulated-dose on the sensitivity (mV/cGy) of the
MOSFET is shown in Fig. 5. The line in Fig. 5 indicates a
linear fitting. The sensitivity increased slightly with the
accumulated-dose. The result indicates that frequent cali-
bration of the sensitivity may improve the accuracy to use
the MOSFET detectors for absolute dosimetry.
3.2 Fading effect
It is known that the radiation stored in a MOSFET fades
over time [2, 6]. The fading effect is shown in Fig. 6. When
the MOSFETs were read 1,200 s after irradiation, we
observed a fading effect of 0.9% with 0.9% standard error
of the mean.
The line in Fig. 6 indicates a linear fitting. Here, it takes
about 1,200 s in IMRT. Namely, if the fading curve is
linear like this, i.e., fading(t) = 1 - 2% t/1,200 s, then the
effective fading over 1,200 s will be just the averaged
response, which is about 1%. Thus, we found that the
fading effect does not affect the MOSFET dose in the
IMRT dose verification.
3.3 Tissue-maximum ratio
The radiation energy spectrum changes with the thickness
of material through which a photon beam passes. The ideal
Table 1 Dose-rate effects of
the MOSFET detector for 6 and
10 MV photon beams
Photon energy (MV)610
Sensitivity (mV/cGy) 1.083 ± 0.0061.093 ± 0.010 1.085 ± 0.0121.102 ± 0.007
58R. Kohno et al.
dosimeter can measure absorbed dose regardless of energy.
Regarding the energy dependence for a photon beam,
Ramaseshan et al.  reported a uniform energy response
in the therapy range between 4 and 18 MV. Edwards et al.
 and Kron et al.  measured the energy dependence
for low-energy photons, and found that MOSFET dosi-
meters overestimate the dose at low energies.
We took the TMR ratios of TMR measured by the
MOSFET (TMR_MOSFET) to that measured by the
Farmer (TMR_Farmer). The TMR ratios in the SW phan-
tom for 6 and 10 MV photon beams, is shown in Fig. 7.
The TMR for 6 and 10 MV measured by the MOSFETs
agreed with the Farmer ionization chamber measurements
within ±0.9 and ±1.6%.
3.4 Angular dependence
The shape of the MOSFET is not designed with perfect
cylindrical symmetry. Previous studies investigated the
angular dependence of the MOSFET response [2, 3, 7].
They reported the angular dependence for a photon beam
was within 2–3%. Wang et al. [13, 14] simulated the
angular dependence of the MOSFET response by the
Monte Carlo method, and they evaluated the angular
dependence as a function of photon energy.
Figure 8 shows the angular dependence of the MOSFET
response for 6 and 10 MV photon beams. The MOSFET
responses were measured at 45? intervals from 0 to 360?.
These were normalized to the response at 0?. Error bars
represent the standard errors. From Fig. 8, it is clear that
the angular dependence of the MOSFETs was within
±3.0% for each energy level.
3.5 Total scatter factors
Figure 9 shows the comparison of the TSCF obtained in
the DC + SW by the MOSFET, and that measured in water
by the PPV. We took the TSCF ratios of TSCF measured
Fig. 4 Accumulated-dose effect of the MOSFET for photon doses up
to 940 cGy
Fig. 5 Effect of accumulated dose on the sensitivity (mV/cGy) of the
MOSFET. The line indicates a linear fitting
Fig. 6 Fading effect of MOSFET response. The line indicates a
Dosimetric evaluation of MOSFET detector in photon therapy 59