Dosimetric optimization of a conical breast brachytherapy applicator for improved skin dose sparing

Transcription

1 Dosimetric optimization of a conical breast brachytherapy applicator for improved dose sparing 5 Yun Yang Biomedical Engineering and Biotechnology, University of Massachusetts Lowell, Massachusetts Mark J. Rivard a) Department of Radiation Oncology, Tufts University School of Medicine, Boston, Massachusetts Purpose: Both the AccuBoost D-shaped and round applicators have been dosimetrically characterized and clinically used to treat patients with breast cancer. While the round applicators provide conformal dose coverage, under certain clinical circumstances the breast dose may be higher than preferred. The purpose of this study was to modify the round applicators to minimize dose while not substantially affecting dose uniformity within the target volume and reducing the treatment time. Methods: In order to irradiate the intended volume while sparing critical structures like the, the current round applicator design has been augmented through addition of an internal truncated cone (i.e., frustum) shield. Monte Carlo methods and clinical constraints were used to design the optimal cone applicator. With the cone applicator now defined as the entire assembly including the surrounding tungsten-alloy shell holding the HDR 192 Ir source catheter, the applicator height was reduced to diminish the treatment time while minimizing dose. Monte Carlo simulation results were validated using both radiochromic film and ionization chamber measurements based on established techniques. Results: The optimal cone applicators diminished the maximum dose by 15% to 32% (based on applicator diameter and breast separation) with tumor dose reduced by less than 3% for a constant exposure time. Furthermore, reduction in applicator height diminished the treatment time by up to 30%. Radiochromic film and ionization chamber dosimetric results in phantom agreed with Monte Carlo simulation results typically within 3%. Larger differences were outside the treatment volume in low dose regions or associated with differences between the measurement and Monte Carlo simulation environments. Conclusions: A new radiotherapy treatment device was developed and dosimetrically characterized. This set of applicators significantly reduces the dose and treatment time while retaining uniform target dose. Key words: Ir-192, brachytherapy, applicator, optimization, dosimetry I. INTRODUCTION The AccuBoost (Advanced Radiation Therapy LLC, Billerica, MA) D-shaped applicators and round applicators were designed to deliver HDR 192 Ir breast brachytherapy through collimation 40 of emissions by an external shield. 1,2 Thus, dose is delivered to tissues within the applicator 84782_3_art_file_474313_l8dzpz.doc Page 1 of 15 9/7/2010

2 aperture uniformly along the applicator central axis. Although both types of applicators can provide lower dose than single-field external electron beam and MammoSite system for applying breast brachytherapy boost to the tumor bed, dose may be too high under certain clinical circumstances such as for combinations of small applicators and large breast 45 separations. 3,4 Though no significant reactions have been observed clinically to date when using the AccuBoost applicators to deliver whole breast irradiation boost at 2 Gy per fraction for Gy, it is possible that reactions such as moist desquamation may occur over the treated region. For accelerated partial breast irradiation (APBI) using this fractionation scheme, the maximum acceptable daily dose is less than 4.25 Gy based on 50 nationwide observations of healthy tissue toxicities. Consequently, the design of the AccuBoost applicators was reexamined to consider improvements for the superficial doses as would be delivered to the breast. For simplicity and due to their cylindrical symmetry, improvements of only the round applicators were considered. For a single axis of treatment with the AccuBoost round applicators, the maximum dose is 55 located on the surface of the breast. By adding of an internal tungsten-alloy shielding device (i.e., truncated cone or frustum) inside the current round applicator, it is possible to irradiate the intended volume while diminishing the dose. By reducing the overall applicator height, it is also possible to reduce the treatment time. This study improves the clinical capabilities of the round applicators through Monte Carlo (MC) design optimization by 60 reducing the dose and treatment time while retaining uniform tumor dose. II. MATERIALS AND METHODS 84782_3_art_file_474313_l8dzpz.doc Page 2 of 15 9/7/2010

3 II.A. Monte Carlo simulations To guide the MC simulations for applicator design, an assessment of thickness was 65 needed. Normal human is composed of epidermis layer, dermis layer and subcutaneous layer. While breast thickness has been simulated as 4.0 mm thick for standardization of MC breast studies, 5,6 it is larger than the measured range. In a mammography study of 150 women, 7 the thickness of normal human breast was found to range from mm. In a second study of 49 women using CT, the breast thickness, including epidermis layer 70 and dermis layer, was measured to be mm. 8 A study 4 of 11 patients and 22 treatment plans with parallel-opposed AccuBoost beam arrangements showed that the distances between target volumes and the mammography paddle were > 5 mm. Consequently, the hypothesis of sparing breast while retaining uniform target dose seemed plausible. Based on the measured data and confidence intervals, a maximum breast 75 thickness of 3.0 mm was used in this study with defined as 0 < d < 3.0 mm. The methods used to perform the MC simulations using MCNP5 version 1.40 radiation transport code 9 were similar to those used to characterize the dose distributions for the AccuBoost D-shaped and round applicators. 1,2 To determine the optimal frustum design for the interior shield, frustum variables were studied individually towards developing a cone 80 applicator (Fig. 1). The frustum was simulated (Fig. 2a) with variable projected radius r (relative to the HDR 192 Ir source plane), frustum height h (inside the applicator), and apical depth f (into the breast tissue relative to the compression paddle). Based on a study of 11 patients, 4 as well as data from a patient registry including multiple institutions, 10 the average breast separation S is about 60 mm, and thus the average target (tumor bed) depth d is 30 mm _3_art_file_474313_l8dzpz.doc Page 3 of 15 9/7/2010

4 85 While this value (S = 60 mm) was used for the design optimization, evaluations of dose profiles and depth-dose data for other S values were considered. To determine the optimal frustum design, r was simulated to align the cone edge 3.0 mm to 0.0 mm with 0.5 mm increments from the applicator internal aperture relative to the source dwell plane, h was simulated from 5.5 mm to 20.5 mm with 5.0 mm increments, and f was 90 simulated to be 2.0 mm < f < 7.0 mm with 0.5 mm increments (an extreme value of f = 10.0 mm was also simulated) Seven frustums with different r, four frustums with different h, and eleven frustums with different f were simulated for cone applicators having internal diameters of 4 cm to 8 cm (4C, 5C, 6C, 7C, and 8C), for comparisons to the round applicators (4R, 5R, 6R, 7R, and 8R). 2 The overall applicator height H was reduced by 3 mm to 18 mm after determining 95 the optimal cone design, and was designed to balance the benefits of reduced dose and treatment time (in comparison to the round applicator). Three-dimensional (3D) dose distributions in the phantom were calculated using the MCNP5 track-length volumetric estimation of collisional kerma (F6 tally, 0.5 mm voxels) which stochastically-approximated absorbed dose. With the optimal design and 3D dose distributions, the ratio of maximum 100 dose to the dose at the center of the compressed breast along the central axis was obtained. center D max Towards estimating the variation of dosimetric results based on the practical nature of the applicator design, a sub-analysis of the applicator components (i.e., frustum, catheter, source) was performed to assess the effects of engineering tolerances. Frustum shift within the 105 applicator aperture by 1 mm and 2 mm was calculated using a 2D *FMESH4 tally. 9 Each side of the cubic mesh elements had 0.5 mm resolution. Dose was sampled on the X-Y plane at Z 84782_3_art_file_474313_l8dzpz.doc Page 4 of 15 9/7/2010

5 = 40 mm and 10 mm and also on the X-Z (Y = 0) and Y-Z (X = 0) planes. Variation of source position within the applicator (Fig. 2b) was assessed by varying source position within the catheter and catheter position within the catheter guide. The extreme positions for maximum 110 and minimum source shielding by the frustum were simulated with the aforementioned F6 tally, and corresponded to the source positioned farthest from and closest to the, respectively. II.B. Experimental measurements After the optimal frustum geometry was determined using MC methods, the optimal 6C applicator (6R applicator + optimal frustum) was measured using both radiochromic film and 115 ionization chamber. The measurement techniques were identical to those used previously for the AccuBoost D-shaped and round applicators. 1,2 Depth-dose data were measured using radiochromic type EBT film (International Specialty Products in Wayne, NJ) and measurements were converted to absorbed dose using film batch calibration. 1 Ionization chamber measurements were made with a calibrated Markus ionization chamber (model N23343, 120 PTW-Freiburg GmbH in Freiburg, Germany). Since the change in chamber response between 192 Ir and 60 Co was negligible for this study, 11 dose to water for the ion chamber was calculated from integrated charge using ADCL-provided N D,W factors for 60 Co. This approach was discussed in detail by Yang and Rivard. 1 Although there are some uncertainties in the wall correction values for the Markus 125 chamber that would affect its beam quality correction factor k(q) at these energies, the chamber response between 192 Ir and 60 Co is not expected to vary by more than 2%. 12 The results from the Monte Carlo calculations and the radiochromic film measurements support this assumption. To determine the depth of maximum dose, measurements were performed with 1 mm 84782_3_art_file_474313_l8dzpz.doc Page 5 of 15 9/7/2010

6 increments at 3.0 d 10.0 mm for the optimal 6C applicator. Both sets of measurements 130 were performed in an 80-mm thick mm 2 polystyrene (PS) phantom with one 0.7 mm PS sheet and three 1 mm PS sheets were used to mimic the 3.67 mm polycarbonate mammography paddle. 2 II.C. Uncertainty analysis Dosimetric uncertainty analyses were performed based on the analysis of Rivard 13 and by 135 Yang and Rivard. 1 Type A and Type B dosimetric uncertainties are presented in Table I. III. RESULTS III.A. Monte Carlo simulations III.A.1. Optimal frustum design 140 Fig. 3a illustrates comparisons of dose profiles at d = 0 mm (approximating the ) and d = 30 mm (approximating the tumor bed) for the 6R applicator and 6C applicators with different r. For 6C applicators with r = 29.5 mm or r = 30.0 mm, lower dose was observed than for the 6R applicator, but the average tumor dose t umord avg was reduced by > 13%. For 27.0 < r < 29.0 mm, maximum tumor dose t umor D max reduction was within 3% and reduced as r 145 increased. Similarly, dose decreased as r increased. The r = 28.0 mm and r = 28.5 mm designs were preferred as they provided equivalent tumor dose coverage with 2% higher tumord avg than the r = 29.0 mm design. Although r = 28.5 mm provided 2% lower maximum dose D max and average dose D avg than r = 28.0 mm, r = 28.0 mm was determined to be the optimal projected radius for the 6C applicator to avoid oblique source 150 shielding given the potential for source/catheter motion within the applicator. This approach 84782_3_art_file_474313_l8dzpz.doc Page 6 of 15 9/7/2010

7 was extended to the other applicator diameters, and resulted in a trend of optimal frustum r design associated with applicator design guideline of internal radius minus 2 mm. Comparisons of dose profiles (Fig. 3b) for the 6C applicators with different h values at d = 0 mm and d = 30 mm show D max and D avg were reduced by > 24% compared to the 155 6R applicator. This held true both inside and outside the applicator aperture. Inside the applicator aperture, h variations reduced t umor D max and t umord avg by less than 3% and 7%, respectively, which were less than the resultant dose reductions. Similar results were obtained for the other four applicator sizes. To avoid potential collision between the frustum and the compression paddle, the optimal frustum height (19.5 mm) was set to be 1 mm 160 less than the maximum height in all cases, and did not significantly diminish the benefit of a full-height h. Comparisons of dose profiles (Fig. 3c) for the 6C applicators with different f values at d = 0 mm and d = 30 mm show D max and D avg were reduced by > 24% compared to the 6R applicator. This also held true both inside and outside the applicator aperture. Because 165 of the prior discussion and citations on thickness, f < 4.0 mm was not considered to be optimal because the frustum shielding would not protect the. For 4.0 < f < 10.0 mm, tumord avg decreased 1%/mm of apical depth relative to 6R applicator. The dose profiles for d = 4.0 mm (Fig. 4) indicate that f = 5.0 mm protects tissues at d < 4.0 mm. The dose gradient is high, and changes by a factor of 2.8/mm on the central axis at d = 4.5 mm. This accuracy of 170 the model is emphasized by the fundamentally different depth dose behavior shown by the f = 4.5 mm (red arrow) and f = 5.0 mm (black arrow) trials as shown in Fig. 4 at these two depths. Considering a maximum clinical setup tolerance of 1.0 mm due to possible gaps between the 84782_3_art_file_474313_l8dzpz.doc Page 7 of 15 9/7/2010

8 applicator:compression paddle and compression paddle:breast, breast tissues at d = 3.0 mm would receive the dose at d = 4.0 mm. Based on these dosimetric data and breast 175 dimensions from section I, the optimal focus depth was determined to be f = 5.0 mm. With the optimal 6C applicator defined as r = 28.0 mm, h = 19.5 mm, and f = 5.0 mm, center D max for various S values were compared (Fig. 5) to the 6R /center dose ratios. Over the full range of applicator diameters, optimal cone applicators reduced center D max from % 32% compared to the round applicators, with center D max < 2.0 for all S values considered. In other words, D D t max < umor max for 4-field arrangements with the current height of parallel-opposed AccuBoost applicators. III.A.2. Height reduced applicator The optimal cone applicators reduced t umor D max by 3% and increased treatment time by 185 3% for a constant t umord avg. Comparisons of center D max and treatment time ratios for 6C applicator for variable H as normalized to 6R applicator (with H = 31 mm) are tabulated in Tables II and III. Due to proximity of source dwell plane to the surface, D max increased as H decreased. With large H reductions, the frustum shield used in the cone applicators could not diminish D max in comparison to the round applicators. Though the 190 6C applicator with H = 13 mm diminished treatment time by 40%, D max increased by 42%. To balance the benefits of reduced D max and treatment time, H = 26 mm was chosen for uniform application to all size applicators. With H = 26 mm and S = 60 mm for the 6C applicator, center D max reduced by 16% and treatment time reduced by 13% in comparison to 6R applicator. Similar results were observed for the other diameter conical applicators _3_art_file_474313_l8dzpz.doc Page 8 of 15 9/7/2010

9 195 III.A.3. Dosimetric variability For the maximum and minimum changes in source shielding due to source positioning within the catheter and catheter guide, large differences were observed in the shadow of the frustum shield (45% at d ~ 0 mm on the central axis) and also outside the applicator aperture 200 (19% at large depths) due to the frustum shielding angle where the relative dose rates are very low. Otherwise, the dose differences due to source positioning were within 7% and explained by the inverse-square law. Cylindrical symmetry of the applicator dose distribution was broken upon shifting the frustum. MC mesh tally results for 1 mm and 2 mm frustum shifts relative to the unshifted 205 frustum results are shown in Fig. 2c for the optimal 6C applicator. On the d = 0 mm plane at any one point, a maximum increase of 48% and maximum decrease of 53% was observed for the 1 mm frustum shift, with a maximum increase of 67% and maximum decrease of 86% for the 2 mm frustum shift. However, D max increased by 1.3% and 2.6% for frustum shifts of 1 mm and 2 mm, respectively. On the d = 30 mm plane at any one point, the maximum 210 change was 10% and 15% for 1 mm and 2 mm frustum shifts, respectively. tumor D max increased by only 0.25% and 0.50% for frustum shifts of 1 mm and 2 mm, respectively. III.B. Radiochromic film Radiochromic film results are shown in Fig. 6. Depth-dose measurements along the 215 central axis in PS from the radiochromic film were compared (Fig. 7a) to MC simulations for the 6R and optimal 6C applicators. Agreement typically within 2% was observed over the 7 < d < 84782_3_art_file_474313_l8dzpz.doc Page 9 of 15 9/7/2010

10 70 mm range for the 6R and 6C applicators. Differences exceeding 3% were observed for the 6C applicator for d < 7 mm, and may be explained due to differences between the idealized MC computational environment with the source centrally positioned within the catheter guide and 220 the measurement environment where source positioning may have varied (section III.A.3). The film and MC results on the central axis differed by 8% at d = 7 mm and 16% at d = 0 mm, and were well within the potential 45% difference possible due to variable positioning of the source within the catheter guide. Dose profiles at d = 0 mm and d = 30 mm for the 6R and optimal 6C applicators (Fig. 7b) show agreement between film and MC results typically within 225 5% for both the 6R and 6C applicators. Larger differences were observed substantially outside the field edge where the dose rates were lowest due to the previously mentioned influence of variable source positioning (section III.A.3). III.C. Ion chamber Depth-dose measurements along the central axis in PS from the ion chamber were 230 compared (Fig. 7a) to MC simulations for the 6R and optimal 6C applicators. Agreement typically within 3% was observed over the 0 < d < 70 mm range for the 6R applicator. Differences exceeding 5% were observed for the 6C applicator for d < 8 mm. Based on the prior film analysis, it was evident that volume averaging within the ionization chamber led to the disparity. Upon volume averaging the MC results to mimic the spatial volume averaging of the 235 ion chamber, the chamber:mc agreement improved to within 3% for 3 < d < 70 mm. IV. SUMMARY Through design optimization using Monte Carlo methods for radiation transport simulations, 84782_3_art_file_474313_l8dzpz.doc Page 10 of 15 9/7/2010

11 the current round AccuBoost applicator was altered and its dosimetric properties were studied. 240 The main design change included an internal frustum positioned centrally within the round applicator. The dimensions and resulting dosimetric influence of the frustum were studied individually. This new cone applicator exhibited substantial reduction (up to 32%) in maximum dose while not perturbing tumor dose uniformity and only minor reduction in total tumor dose for a fixed treatment time. Consequently, an additional parameter (applicator height) 245 was also examined. The final design balanced the benefits of reduced dose ( 10%) and treatment time ( 30%) in comparison to the round applicators. Other combinations for reducing dose and treatment time are easily attainable should the clinical rationale change. Comparisons of measured results with MC simulations indicated good agreement, and revealed the importance of considering variations between measurements and an idealized 250 MC environment. The optimal cone applicators have been dosimetrically characterized, and preparations are underway for treatment planning commissioning and clinic use. ACKNOWLEDGEMENTS We extend our thanks to Christopher Melhus and Martin Fraser of Tufts Medical Center in 255 Boston for assistance with some of the measurements and simulations. Advanced Radiation Therapy, LLC provided the AccuBoost cone applicator and technical data on the applicator design. Rivard is a stakeholder of Advanced Radiation Therapy, LLC _3_art_file_474313_l8dzpz.doc Page 11 of 15 9/7/2010

12 260 Table I. Dosimetric uncertainty analysis of the Monte Carlo simulations in breast, and ion chamber measurements, and radiochromic film measurements in polystyrene at a depth of d = 30 mm on the central axis. Monte Carlo simulations uncertainties Uncertainty component Type A Type B Source geometry 0.46% Capsule geometry 0.01% Source positioning (± 0.45 mm in all directions) 3.34% Frustum shift (± 1.0 mm laterally) 0.49% Source radiation spectrum 1 % Breast phantom composition 1.08% Physics of Monte Carlo code 0.05% μ en/ ρ for dose calculation 0.61% Cross-sections in breast 0.13% Tally volume averaging 0.002% Tally statistics 0.69% Total standard uncertainty (k = 1) 3.8% Ion chamber measurement uncertainties Uncertainty component Type A Type B 192 Ir air-kerma strength calibration 1.3 % Ionization chamber ADCL calibration coefficient 0.7 % Energy correction ( 60 Co to 192 Ir) 2.0 % Electrometer ADCL calibration coefficient 0.2 % Correction for water-to-polystyrene conversion factor (0.4 MeV) 0.16 % Applicator:ionization chamber positioning (± 0.5 mm vertically) 1.39 % Applicator:ionization chamber positioning (± 1.0 mm laterally) 0.49 % Source positioning (± 0.45 mm in all directions) 3.34 % Frustum shift (± 1.0 mm laterally) 0.49 % Irradiation time (± 0.1 s) 0.17 % Temperature (± 1 K) and pressure (± 10 mbar) correction 1.04 % Ion chamber collecting volume averaging (0.2 mm) % Repetitive measurements for same experimental setup 1.61% Total standard uncertainty (k = 1) 4.9% Radiochromic film measurement uncertainties Uncertainty component Type A Type B 192 Ir air-kerma strength calibration 1.3 % Energy correction (6 MV to 192 Ir) 0.5 % Radiochromic film calibration (from 6 MV linac) 0.62 % Film digitizer density:dose calibration 0.94 % Applicator:radiochromic film positioning (± 0.5 mm vertically) 1.39 % Applicator:radiochromic film positioning (± 1.0 mm laterally) 0.49 % Source positioning (± 0.45 mm in all directions) 3.34 % Frustum shift (± 1.0 mm laterally) 0.49 % Irradiation time (± 0.1 s) 0.17% Film/digitizer volume averaging (356 μm pixels) 0.005% Repetitive measurements for same film/irradiation 2.10% Total standard uncertainty (k = 1) 4.6% 84782_3_art_file_474313_l8dzpz.doc Page 12 of 15 9/7/2010

13 265 Table II. Maximum /center dose ratios for parallel-opposed arrangements of height-reduced 6 cm cone applicators normalized to 6R_31 applicator for breast separations S ranging from 30 mm to 80 mm. ICRU 44 Breast S/mm 6C_13 6C_16 6C_19 6C_22 6C_25 6C_28 6C_31 6R_ Table III. Treatment time ratios of height-reduced 6 cm cone applicators normalized to 6R_31 applicator for breast separations S ranging from 30 mm to 80 mm. ICRU 44 Breast S/mm 6C_13 6C_16 6C_19 6C_22 6C_25 6C_28 6C_31 6R_ _3_art_file_474313_l8dzpz.doc Page 13 of 15 9/7/2010

14 FIGURE CAPTIONS Fig. 1 Cone applicator structure used in Monte Carlo simulations and experimental measurements. Fig. 2 a) Applicator variables: frustum projected radius r (relative to the HDR 192 Ir source plane), frustum height h (inside the applicator), apical depth f (into the breast tissue relative to the compression paddle, applicator height H, and frustum shift s; b) enlarged HDR 192 Ir source with different position (lower); and c) relative Monte Carlo-estimated breast dose distributions for the optimal 6 cm cone (6C) applicator with 0 mm, 1 mm, and 2 mm frustum shift. In each image set, dose profiles are depicted in X-Y plane at Z = 40 mm (left) and in Y-Z plane at X = 0 mm (right). Starting with the outermost region, the isodose values vary in 10% increments. Fig. 3 Dose profiles at d = 0 mm and 30 mm for 6 cm diameter round (6R) applicator and 6 cm diameter cone (6C) applicators: a) frustum with f = 5mm, h = 20.5 mm and r = 27 to 30 mm with 0.5 mm increment; b) frustum with f = 5mm, r = 28 mm and h = 5.5 to 20.5 mm with 5 mm increment; and c) frustum with h = 20.5 mm, r = 28 mm and f = 2, 3, 4, 5 and 10 mm. Fig. 4 Dose profiles for 6 cm diameter cone (6C) applicators with different apical depth at depth of 4.0 mm (upper) and 4.5 mm (lower) in comparison to the 6 cm diameter round (6R) applicator. Fig. 5 Maximum /center dose ratios normalized to the center of the breast for the round applicators and optimal cone applicators as a function of breast separation S. The 4R, 5R, 6R, 7R, 8R and 4C, 5C, 6C, 7C, 8C are round and optimal cone applicators with internal diameters of 4 8 cm respectively. Fig. 6 Depth dose profiles and dose profiles in polystyrene determined using radiochromic EBT film for the 6 cm diameter round (6R) applicator and optimal 6 cm diameter cone (6C) applicator. In each image set, dose profiles are depicted through the X-Y plane (upper) at Z = 40 mm and Y-Z plane (lower) at X = 0 mm. Fig. 7 a) Comparison of depth dose profiles determined using ion chamber and radiochromic film in polystyrene to Monte Carlo simulation results for the 6 cm diameter round (6R) applicator and optimal 6 cm diameter cone (6C) applicator; and b) Dose profiles in polystyrene at d = 0 mm and 30 mm depth obtained using Monte Carlo methods and radiochromic film for the optimal 6R applicator and 6C applicator _3_art_file_474313_l8dzpz.doc Page 14 of 15 9/7/2010

15 a) author to whom all correspondence should be addressed: 1 Y. Yang and M. J. Rivard, Monte Carlo simulations and radiation dosimetry measurements of peripherally-applied HDR 192 Ir breast brachytherapy D-shaped applicators, Med. Phys. 36, (2009). 2 M. J. Rivard, C. S. Melhus, D. E. Wazer, and R. J. Bricault, Dosimetric characterization of round HDR 192 Ir AccuBoost applicators, Med. Phys. 36, (2009). 3 M. J. Rivard, R. J. Bricault, J. R. Hiatt, C. S. Melhus, P. Sioshansi, and D. E. Wazer, Interactive image guided peripheral brachytherapy for the boost dose in breast irradiation, Int. J. Radiat. Oncol., Biol., Phys. 69, S659 abstract (2007). 4 S. Sioshansi, M. J. Rivard, J. R. Hiatt, A. A. Hurley, Y. Lee, and D. E. Wazer, Dose modeling of the AccuBoost brachytherapy system with comparison to standard external beam partial breast irradiation techniques, Int. J. Radiat. Oncol., Biol., Phys. (in press) DOI: /j.ijrobp (2010). 5 American Cancer Society, Guidelines for the cancer-related checkup: recommendations and rationale, CA Cancer J. Clin. 30, (1980). 6 X. Wu, G. T. Barnes, and D. M. Tucker, Spectral dependence of glandular tissue dose in screen-film mammography, Radiol. 179, (1991). 7 S. A. Willson, E. J. Adam, and A. K. Tucker, Patterns of breast thickness in normal mammograms, Clin. Radiol. 33, (1982). 8 S. Huang, J. M. Boone, K. Yang, A. L. C. Kwan, and N. J. Packard, The effect of thickness determined using breast CT on mammographic dosimetry, Med. Phys. 35, (2008). 9 X-5 Monte Carlo Team, MCNP A General Monte Carlo N-Particle Transport Code, Version 5, Los Alamos National Laboratory, Los Alamos, NM, S. Hamid, D. E. Wazer, S. Ackerman, D. Arthur, R. Benda, S. Cavanaugh, R. Kuske, B. Prestidge, C. A. Quiet, S. Sha, et al., A multi-institutional assessment of the feasibility, implementation, and early clinical results with noninvasive image-guided breast brachytherapy (NIIGBB) for tumor bed boost, Cancer Research (accepted) abstract (2010). 11 D. Baltas, K. Geramani, G. T. Ioannidis, K. Hierholz, B. Rogge, C. Kolotas, K. Muller-Sievers, N. Milickovic, B. Kober, and N. Zamboglou, Comparison of calibration procedures for 192 Ir high-dose rate brachytherapy sources, Int. J. Radiat. Oncol., Biol., Phys. 43, (1999). 12 C. M. Ma and A. E. Nahum, Bragg-Gray theory and ion chamber dosimetry for photon beams, Phys Med Biol. 36, (1991). 13 M. J. Rivard, Brachytherapy dosimetry parameters calculated for a 131 Cs source, Med. Phys. 34, (2007) _3_art_file_474313_l8dzpz.doc Page 15 of 15 9/7/2010

16

17

18

19

20

21

22

23

24