Journal of Cancer Research and Therapeutics Medknow Publications on behalf of the Association of Radiation Oncologists of India (AROI) ISSN: 0973-1482 EISSN: 1998-4138 Vol. 5, Num. 4, 2009, pp. 284-289

Journal of Cancer Research and Therapeutics, Vol. 5, No. 4, October-December, 2009, pp. 284-289

Brief Communication

Characterization of metal oxide field-effect transistors for first helical tomotherapy Hi-Art II unit in India

Kinhikar RajeshA, Pai Rajeshree, Master Zubin, Deshpande DeepakD

Department of Medical Physics, Tata Memorial Centre, Parel, Mumbai

Correspondence Address:Department of Medical Physics, Tata Memorial Hospital, Parel, Mumbai - 400 012
rkinhikar@rediffmail.com

Code Number: cr09067

PMID: 20160363

DOI: 10.4103/0973-1482.59911

Abstract

Keywords: Ion chamber, metal oxide semiconductor field-effect transistor, thermoluminescence dosimeter, tomotherapy

Introduction

Hi-Art II tomotherapy (Tomotherapy Inc., Corporation, Madison, WI, USA) has emerged as a novel approach for intensity-modulated radiotherapy (IMRT). ^{[1],[2],[3],[4]} The 6-MV linear accelerator is mounted on a slip ring gantry with mega-voltage computed tomography (MVCT) xenon-based detectors mounted opposite to it. The details about the tomotherapy machine design and the features have been described elsewhere. ^[5],[6]

Helical tomotherapy unit is a dedicated IMRT device with on-board MVCT imaging capability. Its radiation beam characteristics are in several aspects different from characteristics of other treatment units. As a consequence, it is recommended to first calibrate the metal oxide field-effect transistor (MOSFET) dosimeters before clinical use with tomotherapy. In addition, the detector sensitivity needs to be derived from a reference class dosimeter.

The objective of this study was to validate MOSFET detectors at phantom surface for a 6-MV beam generated by a tomotherapy machine and to characterize the response for reproducibility, field size dependence, angular dependence, and fade effect. This MOSFET is not supposed to show any temperature dependence due to its design. ^[7] Moreover, the fading effect does not significantly vary the characteristics of the detector. ^[7] Hence, among these characteristics, only the determinations of field size dependence and angular dependence are strictly needed for the tomotherapy purpose. However, to confirm the expected independence of this MOSFET detector response of temperature and readout time of the detector after irradiation, the tests were performed.

Materials and Methods

Tomotherapy machine

The helical tomotherapy accelerator is mounted on a slip ring gantry with computed tomography (CT) xenon-filled linear detector array mounted opposite to the source. This model utilizes the same accelerator as the Siemens PrimArt 6 MV linear accelerator (Siemens Inc., Concord, CA, USA). It employs a slit beam of radiation that continuously rotates around the patient while the patient continuously translates through the beam. The beam is 40 cm wide in the direction transverse to patient movement. A moveable set of tungsten jaws collimates the beam from closed to 5 cm wide in the inferior-superior direction of the patient. The first helical tomotherapy machine in India was installed at Advanced Centre for Treatment Research and Education in Cancer (ACTREC) in July 2007.

MOSFET detector

A standard MOSFET (TN-502RD sensors plus low-sensitivity-bias supply, Best Medical, Springfield, VA, USA) was used for calibration. The diagrammatic details of the unit are available at the manufacturer′s website. It is a software-controlled system (v 2.2) with semiconductor transistors (sensitive volume of 0.2 mm × 0.2 mm × 0.0005 mm). For all measurements, the bias supply was set at a standard sensitivity (1 mV/cGy). For these measurements, which required a lot of repetitive measurements, the MOSFET AutoSense system developed for on-line radiotherapy, was applied to shorten the measurement times, because the control box of the MOSFET system can be read out without entering the treatment room.

Calibration

Calibration was performed to convert the radiation-induced dosimeter voltage shift to dose, and the detector sensitivity was defined as the ratio of the measured voltage shift of the dosimeter and the actual dose delivered with the ionization chamber at the depth of the maximum dose. In this case, both the MOSFET and ion chambers were positioned at the 85-cm source-to-axis distance (SAD). Once calibrated, the dosimeter can be used with or without the build-up material when applying appropriate conversion factors.

Thus,

[INLINE:1]

The calibration was performed for five sensors. The epoxy side (bubble up) was positioned [Figure - 1] upstream facing the tomotherapy 6-MV beam representing the normal incidence of irradiation. An Exradin A1SL ion chamber (Standard Imaging, Middleton, WI, USA) was placed at a 1.5-cm depth with an SAD of 85 cm in virtual water slabs (density 1.04 gm/cc, Standard Imaging, Middleton, WI, USA) as shown in [Figure - 2]. The A1SL has a small volume (0.056 cm ³ ), which makes it a good candidate for measurements in high-dose-gradient regions during IMRT. The ion chamber wall of the A1SL is 1.1 mm thick and both the ion chamber wall and the central electrode are made of C552 air-equivalent plastic. The ion chamber was irradiated for 30 s for three jaw settings (1 cm × 40 cm, 2.5 cm × 40 cm, and 5 cm × 40 cm) at the isocenter with both static and standard rotational modes of treatment. The charge collected by the TomoElectrometer was recorded and converted into the absorbed dose as per the American Association of Physicists in Medicine (AAPM) Task Group (TG51) dosimetry protocol. ^[11] This calculated dose from the ion chamber was considered as a reference reading for estimating the detector sensitivity of the MOSFET.

The MOSFET system was initialized for at least 1 h prior to calibration. The MOSFETs were placed at the surface of virtual water slabs with a source-to-surface distance (SSD) of 85 cm. The 10-cm-thick virtual water slab was used as a backscatter medium. The movable jaws were set to a calibration field size of 2.5 cm × 40 cm at the isocenter. The dose (3 Gy) was delivered with the machine in a static mode (gantry at 0°) with all leaves open for 30 s. The reading from MOSFET was noted for 10 consecutive measurements for estimating the detector reproducibility. This was repeated for five MOSFET sensors. A known calculated dose (3 Gy) was delivered to each sensor and the sensitivity was estimated. Similarly, the readings were noted for the field size of 5 cm × 40 cm and 1 cm × 40 cm as well.

In addition to the calibration in a static mode, the MOSFET was also calibrated for the tomotherapy unit operating in a standard rotational mode for jaw settings of 1, 2.5, and 5 cm. The cylindrical phantom supplied by the manufacturer was used for this purpose. The measurements for various field sizes were performed to estimate the field size dependence of the MOSFET response. A known (3 Gy) dose was delivered to each sensor and the sensitivity was estimated. In rotational mode, the gantry keeps rotating continuously allowing 52 projections (each projections of 7.2° approximately) in a complete rotation of 360°. In each projection, MLCs may be modulated. This methodology was performed to verify the effect of intrinsic angular dependence (irradiation rotational mode) upon the dose reading under charge particle equilibrium (CPE).

The angular dependence of the MOSFET sensor response was also estimated. For this, the MOSFET was placed on the surface of the virtual water slabs at an SSD 85 cm. The bubble facing the beam (static and zero degree). A procedure of 30 s with a dose of 3 Gy was performed with the above-mentioned field size and the detector reading was recorded. The MOSFET was then rotated by every 45° and the same procedure was repeated.

The MOSFETs were also tested for the fade effect by reading it out instantly postirradiation and 15 min postirradiation. This was performed for the same irradiations with the dose of 3 Gy. The effect of temperature on the MOSFET was also estimated by using the detectors at standard (room) temperature (18°C) and at 30°C.

Subsequently, to test the accuracy of the measurement of the detector sensitivity when used with the tomotherapy unit, the detectors were also calibrated for a 6-MV photon beam generated by Primus (Siemens Inc.) linear accelerator (linac) for comparison. A total of five sensors, one at a time, were positioned at a depth of 5 cm in virtual water slabs. A 10-cm slab of virtual water was used for backscatter. The calibrations were done with an SSD of 100 cm at a field size of 10 cm × 10 cm. For "bubble-up" calibrations, the MOSFET was placed with the epoxy bubble facing the beam and for "bubble-down" calibrations, the epoxy bubble was facing opposite of the beam. A known dose (3 Gy) was delivered to each sensor and the sensitivity was calculated.

Thermoluminescence dosimeter

The measurements were also carried out with a thermoluminescence dosimeter (TLD). Prior to each irradiation, the TLD-100 (LiF:Mg, Ti) powder (The Harshaw Chemical Co., Solon, OH, USA) was annealed using a thermal cycle: 400°C (±5°) for 1-h cooling for 5 min, −100°C for 2 h in a programmable muffle furnace (model-126, Fisher Scientific Co., Pittsburgh, PA, USA), and then cooling to normal room temperature. For annealing, the TL powder was placed inside a glass Petri dish with cover. A Rexon UL-320 TLD Reader (TLD Systems Inc., USA) was used to record TL output at a maximum acquisition temperature of 280°C using a constant heating rate of 14°C/s. A constant time gap of 24 h was maintained between the irradiation and readout. A dose response curve for the TLD-100 powder was generated in a Co-60 gamma ray beam (Equinox 80, Best Theratronics Ltd., Ottawa, Canada) and was found linear in the range of 0.5-4.0 Gy and also for the energy up to 6 MV. ^[13] For measurements using the TLD, about 40 mg of the freshly annealed TLD-100 powder was packed in square polyethylene pouches (approximately 1 cm × 1 cm). This pouch was kept in the virtual water slabs at a 1.5-cm depth with a 10-cm slab as backscatter. The TLD was irradiated for the static and rotational procedures as mentioned above for both the jaw settings. The TL output of about 10-mg powder pouches was recorded using the REXON TLD reader. Thus four readings were obtained from each TL pouch. The mean value of net TL output per unit weight (nC/mg) of these four readings was used for calculation. The uncertainty on TLD-100 powder measurements was ±3%. One person performed both the annealing and heating and followed strict procedures of careful handling of the TLDs.

Results

For both static and rotational measurements, the MOSFET was reproducible (+ 0.5%). [Table - 1] shows the reproducibility of the MOSFET response for tomotherapy and the Primus linear accelerator. The table shows the readings when the jaw setting was 2.5 cm × 40 cm and 10 cm × 10 cm for tomotherapy and linac, respectively. The MOSFET showed the reproducibility within + 0.5% (SD 1.4) for tomotherapy.

The subsequent MOSFET readings were within + 2% for available jaw settings (1, 2.5, and 5 cm). Thus the MOSFET response was found to be independent of the field size. [Table - 2] and [Table - 3] summarize the results for 1-cm and 5-cm field size respectively.

[Table - 4] shows the MOSFET detector sensitivity as a function of field size for tomotherapy. The sensitivity of the MOSFET was also measured for the Primus linear accelerator to estimate the accuracy. Five sensors were used. The jaw setting for the linac was 10 cm × 10 cm. The mean detector sensitivity was found to be 1.10 (SD 0.02) mV/cGy for the static and 1.12 (SD 0.01) mV/cGy for the rotational irradiation modes, respectively. Clearly, these are not statistically different numbers. Thus, the mean detector sensitivity found was the same within the experimental uncertainty for static and rotational dose deliveries. Based on that, we conclude that the same detector sensitivity can be used for each irradiation mode. From the experiment performed with the MOSFET detector on linac 6-MV beam, we found the mean sensitivity as 1.10 mV/cGy. Thus this confirmed the accuracy of the sensitivity for the MOSFET detectors for the 6-MV photon beam generated by tomotherapy and the linac.

The TLD agreed with ion chamber readings for both static and rotational procedures within th + 2%. The measurements in a rotational procedure with MOSFETs revealed the angular dependence of less than 1%. [Table - 6] shows the MOSFET detector response as a function of detector rotation (every 45°) for three field sizes (1 cm × 40 cm, 2.5 cm × 40 cm, and 5 cm × 40 cm). From the results, it is seen that the detector response of less than 1% was observed between the 0° and 180° orientation of the detector. Thus, the MOSFET did not show significant angular dependence with the static irradiation procedure for three field sizes. The maximum variation in the detector response between bubble-up (facing the beam) and bubble-down (facing opposite of the beam) was less than 1%. The MOSFET readings were consistent and reproducible (+ 0.5%) and they showed no significant dependence on the field size. There was a negligible fading effect of the MOSFET. The MOSFET response was found independent of temperature in the range 18°-30° ^. [Table - 5] shows the sensitivity at various temperatures.

Discussion

Helical tomotherapy is the first treatment unit dedicated to IMRT and a fully integrated image-guided radiotherapy (IGRT) system with the on-board mega-voltage CT (MVCT) capability. This is, in some aspects, a nonstandard radiotherapy system because of the field sizes involved and rotational treatment delivery. Hence the dosimetry protocols applied to tomotherapy cannot be similar as the conventional AAPM TG51. The overall uncertainty in dose determination using MOSFETs and the calibration method as applied in this study was about 2%.

As to the characterization of MOSFET detectors for "reproducibility, field size dependence, angular dependence, fade effects, and temperature dependence," only the determinations of field size dependence and angular dependence are strictly needed for tomotherapy purposes. The other dependences concern the general MOSFET behavior, and have been extensively reported in the literature. ^[8],[9] The MOSFET sensitivity of 1.11 mV/cGy was in good agreement with the earlier study. ^[8]

It should be noted that the terminology "detector sensitivity" has been used in this study. Usually, sensitivity refers to the ratio of the detector response and the known dose delivered to the detector. It has the unit, in this case, mV/cGy. The reciprocal of sensitivity is referred to as the calibration factor. However, in the MOSFET user manual, this has not been clearly mentioned. Hence we used the term detector sensitivity in this study with a unit mV/cGy. When compared with the reports, ^[8] we found that the calibration factor of 0.901 cGy/MV has been estimated. Thus the inverse of the mean sensitivity achieved (0.9 cGy/MV) in our study for the standard sensitivity MOSFET was in good agreement (0.11%) with the results obtained by the study mentioned above.

This study deals with the measurements performed with the MOSFET on the surface of the phantom thus simulating actual clinical measurements. The MOSFET response was found independent of temperature in the range 18°-30°. This is in agreement with the earlier reports. ^[9] The AAPM Task Group report ^[12] (TG-21) suggests the calibration of ion chamber or any detector under the reference condition (field size of 10 cm × 10 cm, SSD 100 cm with full build-up). In tomotherapy, the maximum possible field size is 5 cm × 40 cm at the isocenter (85 cm). This gives the equivalent field size close to 10 cm × 10 cm. Tomotherapy has only three jaw widths (1, 2.5, and 5 cm) available for treatment. Hence we used these field sizes. Also the helical tomotherapy SAD is 85 cm, and while the SSD of 100 cm could be achieved, it is not the geometry that patients would be treated with.

In most of the clinical situations, we use either 2.5- or 5-cm jaw settings; however, the MOSFET was calibrated for all the three available field sizes. The maximum field for tomotherapy is 5 cm × 40 cm. The equivalent square of this maximum jaw setting is nearly equal to the reference field size (10 cm × 10 cm) from a standard linac.

The MOSFET response was constant and reproducible. It was revealed from the measurements that the sensitivity of the MOSFET detector was found to be independent of the mode of the tomotherapy unit operation (static or rotational). Tomotherapy treatment delivery is always rotational. The angular dependence of the MOSFET response resulted to be less than 1% during rotational treatments. However, MOSFET readings were reproducible in both the cases. The mean MOSFET reproducibility was 0.5% for a reading of 200 mV, 10 successive measurements, and standard sensitivity of the bias supply. For both static and rotational measurements, the MOSFET was reproducible and the readings were constant for available jaw settings (1, 2.5, and 5 cm). Thus the MOSFET response was found to be independent of the field size. This is in agreement with the earlier literature report. ^[8]

The TLDs were found very useful due to their small size and reusability. Moreover, unlike the ion chamber or MOSFETs, they do not require any bias voltage and hence can directly be placed at any desired location on the patient. The TLDs used were from a single batch and an extreme care was taken during handling with forceps and gloves. From a single TLD packet of the powder, minimum four to five samples were used for readings, and the mean of all the readings was estimated. The uncertainty in TLD measurements was + 2%. The ion chamber is supposed to be a benchmark detector for absolute dosimetry. The same A1Sl chamber was used as a reference detector for the calibration of both MOSFETs and TLDs.

The MOSFET sensitivity was estimated for the 6-MV photon beam generated both by tomotherapy and the linac. The experimental results confirmed the accuracy of these sensitivity factors for MOSFET detectors.

Conclusion

This study deals only with the measurements and calibration performed at the surface of the phantom. The MOSFET was calibrated and validated for phantom measurements for the 6-MV photon beam generated by a tomotherapy machine. The MOSFET was successfully calibrated for the tomotherapy 6-MV static as well as rotational beam. The sensitivity of the detector was the same for both modes of treatment delivery. The performance of the MOSFET was validated and found satisfactory for the helical tomotherapy Hi-Art II unit. However, the MOSFET may be used for in vivo surface dosimetry only after it is calibrated under the conditions replicating as much as possible the manner in which the dosimeter will be used clinically.

References

Copyright 2009 - Journal of Cancer Research and Therapeutics

The following images related to this document are available:

Photo images