HDR Brachytherapy Dose Distribution is Influenced by the Metal Material of the Applicator

Applicators containing metal have been widely used in recent years when applying brachytherapy to patients with cervical cancer. However, the high dose rate (HDR) treatment-planning system (TPS) that is currently used in brachytherapy still assumes that the treatment environment constitutes a homogeneous water medium and does not include a dose correction for the metal material of the applicator. The primary purpose of this study was to evaluate the HDR 192Ir dose distribution in cervical cancer patients when performing brachytherapy using a metal-containing applicator. Thermoluminescent dosimeter (TLD) measurements and Monte Carlo N-Particle eXtended (MCNPX) code were used to explore the doses to the rectum and bladder when using a Henschke applicator containing metal during brachytherapy. When the applicator was assumed to be present, the absolute dose difference between the TLD measurement and MCNPX simulation values was within approximately 5%. A comparison of the MCNPX simulation and TPS calculation values revealed that the TPS overestimated the International Commission of Radiation Units and Measurement (ICRU) rectum and bladder reference doses by 57.78% and 49.59%, respectively. We therefore suggest that the TPS should be modified to account for the shielding effects of the applicator to ensure the accuracy of the delivered doses.

Brachytherapy for patients with cervical cancer involves positioning an applicator within the patient's body, and the characteristics of the applicator material differ markedly from those of the soft tissues of the body. Henschke applicators have been widely used in patients with cervical cancer in recent years 10,11 . This type of applicator has an ovoid metal structure that is designed to reduce the doses applied to the bladder and rectum. However, the dose calculation formula in AAPM TG-43 (American Association of Physicists in Medicine Task Group No. 43) is based on the assumption of a homogeneous water environment 12 ; the formula does not contain corrections for the inhomogeneous media inside the applicator or evaluate the dose distribution in water in the presence of the applicator. It is thus unclear whether the formula can be used to ensure the accuracy of the delivered doses.
The purpose of this study was to determine the effects of the shielding materials contained in the ovoid on the applied doses for multiple 192 Ir source indwelling positions inside the Henschke applicator. This study used Monte Carlo N-Particle eXtended (MCNPX) code to simulate the dose distribution in water in the presence and absence of the Henschke applicator and made measurements using a thermoluminescent dosimeter (TLD). The values measured with the TLD were compared with the values calculated with the treatment planning system (TPS) and the MCNPX simulation results to explore the differences between the dose distributions in the presence and absence of the Henschke applicator as well as the doses at the ICRU bladder and rectum reference points. Comparison of the TPS calculation, MCNPX simulation, and TLD measurement values when including the Henschke applicator. The dose in water in the presence of the Henschke applicator was simulated using MCNPX code, and the TLD measurement values were used to assess the dose distribution near the Henschke applicator. The dose differences among the TLD measurement, MCNPX simulation, and TPS calculation values are listed in Table 1. The greatest differences in doses between the MCNPX simulation and TLDs measured values were within 5%, with the ratio of the simulation and measured values ranging from 0.93 to 1.07. A comparison with the TPS calculation value indicated dose reductions from 89.33% to 1.89% for the TLD measurement value and from 76.49% to 7.16% for the MCNPX simulation value. Therefore, if the shielding effect of the Henschke applicator is neglected, the dose predicted by the TPS can be overestimated.  Figure 1(a) represents the z-y plane at x = 0 cm, which passes through the long axis of the tandem; Fig. 1(b) represents the x-y plane at z = 1.0 cm, which passes through the center of the ovoid; and Fig. 1(c) represents the z-x plane at y = 0 cm. The MCNPX results reveal that the 2D dose distribution was significantly reduced in the   Table 2. When the Henschke applicator was present and the patient accepted single-fraction irradiation, the shielding material could reduce the bladder and rectum average doses by 30.32% and 38.95%, respectively. The dose-volume histograms (DVHs) of the rectum and bladder are shown in Fig. 2 with a spatial resolution of 1 mm 2 . Clearly, the presence of the Henschke applicator can reduce the rectum and bladder doses. The doses covering 50% and 20% of the organ volume (D 50 and D 20 , respectively) decreased by 33.93% and 32.38% for the bladder between the presence and absence of the Henschke applicator, respectively, whereas they decreased by 41.90% and 37.98% for the rectum.

Discussion
The TPS is used in brachytherapy when treating patients with cervical cancer. However, the TPS does not account for the presence of the Henschke applicator. To determine how the applicator affects the TPS calculation values, simulations were performed for an 192 Ir source in the absence of the applicator for the 17 indwelling positions. The MCNPX and TLDs results were both consistent with the conclusion that the TPS would overestimate each point due to the presence of the Henschke applicator (Table 1). These results show that brachytherapy with a Henschke applicator may cause large dose differences when the TPS ignores the applicator.
The bladder (locations R8, R15, R16, and R17) and rectum (locations R9, R13, and R14) dose reference points were located above and beneath the Henschke applicator, respectively; these areas are in the shielding direction of the ovoid, and the doses in these areas were significantly reduced due to the shielding effect of the ovoid. The dose distribution and treatment quality will be affected if water is considered to be present at the location of the Henschke applicator. The metal material of the applicator used in brachytherapy should not be assumed to be the soft tissue. The results also imply that if the tumor is located behind the ovoid, the tumor dose will be overestimated if the shielding effect of the metal is not taken into consideration. A bladder accepting excessive radiation doses will produce incontinence, radiation cystitis, and hematuria, and the rectum will exhibit radiation proctitis, bloody stools, and chronic rectal ulcers [13][14][15] .
A typical clinical treatment plan was applied to the homemade water phantom to enable dose measurements and Monte Carlo simulations. The point dose measurements were consistent with the MCNPX simulation values. Our Monte Carlo simulation results accurately assessed the DVH and the critical organ doses, including at the ICRU rectum and bladder reference points (Tables 1 and 2 and Fig. 2). TG-43 recommends that the agreement between independent dose measurements and TPS dose calculations be within 15%. The MCNPX results indicated that the TPS overestimated the ICRU rectum and bladder reference doses by 57.78% and 49.59%, respectively. Thus, the TG-43 dose calculations produce unacceptable dose distributions when the Henschke applicator is present during brachytherapy. The Henschke applicator has been widely used to treat patients with cervical cancer, and the present study assessed the effects of such an applicator on the dose distribution. When the TPS ignored the effect of the applicator, the MCNPX simulation values were consistent with the TPS calculation values. These results indicate that considering the Henschke applicator has relatively large effects on the dose distribution when brachytherapy is applied to patients with cervical cancer. The TPS currently appears to be insufficient in this case. This study thus recommends that the TPS be modified to take into account an applicator containing metal material to ensure the accuracy of the delivered doses.

192
Ir source and Henschke applicator. The 192 Ir source used in this study had an active length of 3.6 mm, a diameter of 0.65 mm, and a density of 22.42 g/cm 3 ; it was encapsulated by a stainless steel outer cover with an outer diameter of 0.9 mm that was welded to a steel cable for attachment to a remote after-loading machine (microSelectron-HDR v2, Nucletron, The Netherlands). The air kerma strength of this source is 45,730 cGy.cm 2 /h (11.2 Ci) when first installed, and the GENIE TPS (microSelectron, Nucletron) was used. In the Henschke tandem and ovoid applicator, the ovoid contains tungsten alloy as a shielding material with a density of 17.0 g/cm 3 and the following atomic composition: 91% W, 4.5% Ni, and 4.5% Fe. The geometry of the shielding structure inside the ovoid was obtained with the aid of Kodak X-Omat V films.
TLD and the dose readout system. To minimize the influence of the dose gradient on the measurements, the TLD-100 H cube (Harshaw, USA) with dimensions of 1 × 1 × 1 mm 3 was used in this study. TLD-100 H (LiF: Mg, Cu, P) has an effective atomic number of 8.20 and a density of 2.64 g/cm 3 . The dose readings ranged between 1 μ Gy and 10 Gy in the TLD reading system (system UL-320, Rexon, USA).

Homemade water phantom.
A water phantom hosted in an acrylic container was designed for brachytherapy dose measurements. The water phantom had dimensions of 52 × 38 × 29 cm 3 , 2-cm-thick top and button   Table 3. Measurement and calculated dose reference points in a realistic brachytherapy situation.
covers, and 1-cm-thick walls. Holes were drilled in the covers at 1-cm spacings, with 361 holes distributed in a matrix of 19 × 19 cm 2 , giving a total of 722 holes, as shown in Fig. 3. Acrylic rods of various lengths were constructed with a cavity to accommodate the TLD. The rods enabled the TLD to be placed in different holes for dose measurements. The rectum and bladder cavity phantoms were designed based on human anatomy, and the cavities were filled with water when performing the measurement experiments.
Measurements with multiple source indwelling positions. Before performing the measurements, we referred to the Manchester system and ICRU38 report to determine the appropriate measurement points. The Manchester system of brachytherapy is the most widely used for cervical cancer and includes four points: A, B, rectum, and bladder. Point A is located 2 cm superior to the cervical opening and 2 cm lateral to the middle of the cervical canal, and point B is defined as being 3 cm lateral to point A. The ICRU report suggests that the reference points include those in the bladder, rectum, lymph trapezoid, and pelvic wall. In addition to the point A, point B, and ICRU bladder and rectum reference points, we added several measurement points in the region corresponding to the critical organ with the other points arranged along the tandem. This study used the orthogonal imaging method, which takes an anterior-posterior film and a lateral film to define the spatial coordinates of the source. This information enabled the TPS to be created, and a prescribed dose of 6 Gy was applied to point A for absolute dose comparisons. TLD readout values with variance coefficients less than 3% were selected to implement the dose measurement; the resulting 17 measurement points are indicated in Table 3 and Fig. 4. Monte Carlo simulation. The reliability of the TLD results was verified by comparisons with Monte Carlo simulations. The MCNPX code was developed by Los Alamos National Laboratory, and this study used MCNPX version 2.7.0 to calculate the 192 Ir dose distribution 16 . The 192 Ir photon spectrum was obtained from Brookhaven National Laboratory 17 . The simulation was divided into two parts: (A) simulations ignoring the Henschke applicator, with the simulation and TPS results being compared, and (B) simulations that included the Henschke applicator, which simulated the clinical treatment condition. The simulation results could be used to verify the TLD results and examine the accuracy of the TPS.
The mesh tally in MCNPX code was used to calculate the two-dimensional (2D) dose distribution around the 192 Ir source. The mesh tally employed DE and DF cards to convert the fluence into the doses at particular points. The values of DE and DF are the energy bin and dose conversion factor, respectively, for a water medium. The mesh size for dose calculation was 1 × 1 × 1 mm 3 , with at least 10 8 particles being simulated, which yielded 1σ statistical errors of less than 3% at all tested locations. Radiation particles were removed from the simulation when their energy was less than 10 keV.
The 17 source indwelling positions and indwelling times were simulated according to the TPS output file. To analyze the doses to the rectum and bladder, separate simulations were performed in the absence and presence of the Henschke applicator. A simple geometry was input to the MCNPX code to represent the rectum and bladder. The rectum comprised a hollow 7-cm-long and 0.2-mm-thick cylindrical shell with a diameter of 3 cm, and the bladder comprised a 0.2-cm-thick spherical shell with a diameter of 7.6 cm.