Abstract
Aim: To examine the compatibility of the measured and calculated dose for the treatment of lung lesions by helical tomotherapy. Materials and Methods: The administered dose was measured a total of 55 times at 22 points with a radiophotoluminescence glass dosimeter (RPLGD) inserted in the position of an anthropomorphic Rando Phantom. Two Gy were prescribed and calculated with a tomotherapy planning machine for a 3-cm diameter spherical planning target volume (PTV) created in the lung area. Compatibility (measured dose/calculated dose and σ value=(Dmeas-Dcalc)/Dprescribed) ×100 (%)) was analyzed according to dosimeter location. Results: Deviations between measured and calculated doses for the lung lesion were within 4% for planning target volume, indicating that adequate dose delivery to the PTV was achievable. On the other hand, we found dose deviations up to 15% for the lower prescribed dose range (64% or less) for the measured dose/calculated comparison and a 6% deviation according to the σ value in or near inhomogeneous tissue. Conclusion: Although the measured dose satisfied the clinical requirement in almost all areas including PTV, we should note that there may be discrepancies between expected calculated dose and irradiated dose in or near inhomogeneous area.
- Tomotherapy
- lung tumor
- stereotactic body radiotherapy
- radiophotoluminescence glass dosimeter
- Rando phantom
Stereotactic body radiotherapy (SBRT) is a method for increasing the radiotherapy dose delivered to the tumor while minimizing that offered to surrounding normal tissue. This technique has been well-described in the literature, and clinical trials suggest promising results (1-3). These findings prompted us to initiate SBRT using helical tomotherapy (HT), which is a technology that delivers fan-beam intensity modulated radiation therapy (IMRT) under megavoltage computed tomography (MVCT) guidance through continuous and synchronous gantry rotation and couch movement during radiation delivery, which facilitates daily image-guided radiation therapy (IGRT) administration. There is limited information on the clinical use of HT in lung cancer. In spite of the advantages of high setup accuracy, adaptive RT-planning capability, and sparing of normal tissues, low-dose shower is of concern in HT, where radiation is given over 360 degrees (4-6). Side-effects by low-dose shower may be enhanced by additional chemotherapeutic agents. Shueng et al. reported severe radiation pneumonia after 30 Gy/10 fractions of HT radiotherapy and subsequent chemotherapy for T8-T10 metastasis for symptom relief (4). Song et al. reported that HT has produced a somewhat high rate of fatal pulmonary complications. They suggest that the percentage volume of lung irradiated by more than 5 Gy; V5 should be considered and kept as low as possible in addition to the conventional dosimetric factors (5). SBRT using a linear accelerator was found to produce a rather homogeneous dose distribution in planning target volume, but IMRT showed dose inhomogeneity with a steep dose gradient which may lead to inadequate target coverage or a higher dose to wider surrounding normal tissue, and caution should be taken especially when a new chemotherapeutic agent is used (3,6). Hsieh et al. advised caution for combination SBRT using HT and erlotinib regarding the potential risks of enhanced adverse effects (6).
SBRT using HT is performed in several institutes including ours, using custom-made phantom as an accompanying quality assurance item for HT. As there is much ambiguity still within the quality assurance system, we independently attempted to confirm the reliability of dose delivery for consistency of lung tumor SBRT using HT. We selected a 3-D anthropomorphic phantom (Rando phantom) equipped with a radiophotoluminescence glass dosimeter (RPLGD) because of its similarity to the real clinical conditions. The Rando phantom is a widely-available anthropomorphic phantom that has been used for many years for dose measurements in a variety of applications (7-9). The RPLGD was developed in the 1950s and has subsequently been used for radiotherapeutic dosimetry because of its several superior features (10-13). RPLGDs possess good properties for in vivo dosimetry, including small size, ruggedness, nontoxicity, photon-energy and dose rate independence over the energy range 0.2-0.3 MeV, high sensitivity, good reproducibility, and repeat readability until the detectors become annealed. Linearity and reproducibility were better than for previous thermoluminescent dosimeters (TLD)(14-16). These advantages have made RPLGDs more robust than TLDs for in vitro dosimetry. We, therefore, used RPLGDs to establish quality assurance systems for brachytherapy and found deviation between the calculated and measured dose of 10% or more (10,11). Although RPLDGs have less accuracy than an ionization chamber, it enables us to examine the dose to a smaller volume, which is also suitable for IMRT because of its steep dose gradient. For these reasons, we used Rando phantom and RPLGD for quality assurance analysis of HT SBRT.
Materials and Methods
The element of the RPLGD (Dose Ace; Chiyoda Technol Corporation, Tokyo, Japan) is 1.5 mm in diameter and 8.5 mm in length, and the holder is 2.8 mm diameter and 9.5 mm length. A reader (FGD-1000, Chiyoda Technol, Tokyo, Japan) stimulates the RPLGD using a pulsed ultraviolet laser, and readout range is 10 μGy to 10 Gy. The linear dose response for 1 to 136 Gy for the RPLGD has been confirmed (17). RPLGDs are made of uniform glass with an effective atomic number of 12.039, containing 11.00% Na, 31.55% P, 51.16% O, 6.12% Al, and 0.17% Ag by weight. RPLGDs with their holders were inserted into an adult female Rando phantom (The Phantom Laboratory, Salem, NY, USA) (Figure 1).
Each cross section of Rando phantom has a 2.5 cm thickness, and each slice of phantom material contained a grid of plugs that could be removed to allow the insertion of RPLDGs with holders. We obtained measurements at 18 points in the central plane including four points on the body surface (Figure 2).
We also examined two cross sections in both upper and lower directions, at a point corresponding to the central position of the PTV in the central plane. We measured the point dose three times in general, except for the distant area (13, 14, 15, 16, 17, 18, 17) from PTV, for the limitation of entire number of RPLGDs. CT images with a 3-mm slice thickness were acquired with an Aquilion 64 (Toshiba Medical Co., Tokyo, Japan). The images were transferred to a treatment planning system (TPS) radiotherapy plan (Tomotherapy Planning Station Version 3.1.5.3 Hi-Art system; TomoTherapy Incorporated, Madison, WI, USA). The gross tumor volume (GTV) was rendered as a sphere 20 mm in diameter and the PTV (27.79 cm3) was automatically generated by the addition of a 5 mm margin to GTV. For SBRT, we used the prescribed dose of 50 Gy in five fractions for D95. Organs at risk were delineated in lung, spinal cord, and trachea. Dose constraints for the prescription dose of 50 Gy delivered in five fractions were a maximum tolerated dose of 28 Gy for the spinal cord, 44 Gy for the esophagus (we could not use this constraint because the Rando phantom does not include an esophageal structure), and lung V20 <20% mean lung dose <10 Gy. The total lung volume equaled that of both lungs minus that of the GTV, while V20 is the percentage of the lung volume receiving 20 Gy or more. For this mock RT plan, we divided 10 Gy into a five-fraction plan by dividing all values by 5 (Figure 3), and for this investigation, the prescribed dose was set at a single fraction of 2 Gy.
All these parameters were found to be adequate (maximum dose for the spinal cord was 0.452 Gy, V20=15%, mean lung dose=0.95 Gy). The maximum dose for PTV was 10.37 Gy and the minimum dose was 9.63 Gy. The superposition, convolution algorithm was adopted for dose calculations. Irradiation was performed with a Hi-Art Helical Tomotherapy system (Accuray Co. Sunnyvale, CA USA). Following irradiation, RPLGDs were removed from the phantom for estimation. Doses were measured for the one or three administrations, depending on the location, and the average value was regarded as the calculated dose for the given RPLGD. Calibration of RPLGDs and linearity was confirmed by methods described elsewhere (10, 11). Individual correction factors were then generated for each RPLGD. The procedures were repeated five times.
Absorbed dose from MVCT. The absorbed dose from MVCT was assessed by using RPLDGs in similar positions (1, 3, 5, 11, 14, 15, 16, 17, 18) which were then exposed to 3 MV X-rays.
Results
Results for RPLGDs. The absorbed doses measured by RPLGDs after irradiation of the female Rando phantom are presented in Table I. Deviations between measured and calculated doses for the lung lesion were determined as a ratio. The calculation grid size is 3.8 mm. Table II and Figure 2 show differences of more than 5% between the calculated and estimated dose. We also calculated σ=(Dmeas-Dcalc)/Dprescribed) ×100 (%) for reference (Tables I and II).
We found that deviations were within 5% of the PTV, which demonstrates that an adequate dose delivery to the PTV is feasible. Although we found dose deviations of up to 15% for the lower prescribed dose range (64% or less) and/or the inhomogeneous area adjacent to bone (Figure 4), only one point (7; 3 cm to the right of the PTV center, ipsilateral lung) had 6% deviation according to the σ value. Interestingly, good concordance was found in locations in vertical positions and deviations of more than 5% were detected only in the horizontal direction of PTV (Figure 2).
Absorbed dose from MVCT. The PTV area (3 cm in length) was subjected to MVCT. The slice thickness is 4 mm in MVCT. The following doses were absorbed from MVCT: 1, 1.6 cGy; 3, 1.6 cGy; 5, 1.6 cGy; 11, 1.7 cGy; 14, 0.1 cGy; 15, 0.3 cGy; 16, 1.6cGy; 17, 0.8 cGy; 18, 0.1 cGy. The central plane received 1.6-1.7 cGy, the adjacent plane 0.3-0.8 cGy and the 5 cm caudal or cranial plane 0.1 cGy.
Discussion
Helical tomotherapy is a relatively new tool capable of SBRT with both advantages and disadvantages compared to other treatment systems. Firstly, as a CT-based technology with greatly expanded degrees of freedom, it allows for excellent dose sculpting, and provides outstanding conformity of the prescribed dose to the target. The megavoltage CT scan performed immediately before treatment potentially allows for smaller PTV expansions due to reduced setup uncertainty, thereby permitting acceptance of a higher treatment dose (18). On the other hand, tomotherapy currently allows only for coplanar treatment, which results in less capability than that of non-coplanar treatment methods for spreading the exit and entrance dose outside the radial plane. An additional concern is tumor motion, which may theoretically cause tumor underdosing. However, several previous investigations have shown that such concerns are somewhat unrealistic and not likely to have any significant dosimetric consequences (19, 20). Consequently, SBRT using tomotherapy is considered acceptable for clinical usage (21, 22) and this prompted us to initiate lung SBRT using tomotherapy.
The Rando phantom is a widely available anthropomorphic phantom that has been used for many years for dose measurements in a variety of applications (7-9). Measuring the dose at any given point in the Rando phantom is straightforward, and it has been used in diagnostic radiology, radiotherapy, and radiation protection (23, 24). In principle, the Rando phantom should enable investigators to perform an objective comparison of the radiation doses received by normal-sized adults in a wide range of settings. In the present study, we used a newly-developed RPLGD, which is characterized by easier handling and high reliability (14-16). Dispersion of response among dosimeters is small (coefficient of variation 0.82%), and reproducibility of repeat measurements by a single element is excellent (coefficient of variation 0.29%), and better than commercially available TLDs (8). The largest source of error in calculations exist in tissue inhomogeneities, such as the case found in the lung or bony anatomy and the resultant loss of electronic equilibrium. It is reported that the presence of large heterogeneities cannot be entirely accounted for by the superposition-convolution algorithm when compared to dosimetric measurements, especially in lower-dose areas (25). We found a borderline 6% deviation according to the σ value at 3 cm to the right of the PTV center in the ipsilateral lung, the clinical significance of this deviation is not clear and requires further investigation.
In conclusion, although the measured doses satisfied the clinical requirement in almost all areas including PTV, it should be borne in mind that there may be discrepancies between the expected calculated dose and the irradiated dose in inhomogeneous areas.
- Received January 21, 2013.
- Revision received March 5, 2013.
- Accepted March 5, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved