- Email: [email protected]
Correlation of Clinically Significant Radiation Esophagitis With Dose-Volume Histogram Parameters in Lung Cancer
Proceedings of the 49th Annual ASTRO Meeting whether external motions measured simultaneously at multiple locations by multiple sensors would provide a better surrogate marker of internal motion. Materials/Methods: A prototype pressure monitoring system with 4 independent sensors (Anzai Medical) was used to measure external chest motion. A pneumotach spirometer (ADInstruments) was used to monitor air flow during respiration. A custom system for data acquisition and analysis was developed using modular instruments (NI). 12 test subjects breathed normally into the spirometer while wearing elastic belts containing pressure sensors positioned at 3 locations: 1) the abdomen above the umbilicus, 2) the right breast, and 3) the right chest wall inferior to the axilla. Data was acquired with subjects lying supine normally and with a lordosis board providing increased back arch. Results: Figure 1 shows data acquired from a single test subject. The spirometer signal lagged behind the abdomenal pressure sensor signal by an average of 190 msec. Time delays of up to 450 msec also were detected between the pressure sensors located on the abdomen and those on the upper chest. The correlation coefficient between spirometry and pressure signals was improved by inclusion of all 3 pressure signals as independent variables in the correlation analysis. Conclusions: The measured time delay between the spirometer and pressure signals was large enough to be of potential concern in radiotherapy. The delay between abdomenal signal and lung air flow suggests that diaphragm motion precedes lung air flow. The signal delay was less for the upper chest sensors. For radiotherapy of the upper lung or breast, sensors positioned on the upper chest wall might serve as better surrogates of upper chest motion. The uniqueness of information provided by sensors at multiple locations suggests that a combination of these signals might provide a better surrogate of internal respiratory motion than the current standard of a single abdomen sensor. This work was supported in part by Siemens OCS and Anzai Medical.
Fig. 1. Left Column (top to bottom): (1) Respiratory flowrate measured by spirometry (white) and chest wall motion measured using pressure sensors. (2–4) Air volume change versus pressure signal for sensors 1–3. Right Column: (1) Normalized waveforms (integrated flowrate and pressure signals). (2–4): Pair-wise comparison of pressure signals from the three sensors. Author Disclosure: J.D. Christensen, Equipment, C. Other Research Support; A. Tai, Equipment, C. Other Research Support; X.A. Li, Equipment, C. Other Research Support.
2920
Correlation of Clinically Significant Radiation Esophagitis With Dose-Volume Histogram Parameters in Lung Cancer
G. Rodrigues, J. Rose, D. D’Souza, M. Lock London Regional Cancer Program, London, ON, Canada Purpose/Objective(s): Due to dose escalation and increasing use of concurrent chemotherapy, radiation esophagitis (RE) remains a common treatment-limiting acute side effect. The aim of this review is to systematically assess the value of dosimetric parameters reported in the literature in predicting severity of RE and to provide recommendations for future research in the field. Materials/Methods: Both prospective and retrospective clinical studies assessing the relationship between various dosimetric parameters and RE in the treatment of inoperable lung cancers and thymomas were included in this systematic analysis. Our search strategy included a variety of electronic medical databases, textbooks and bibliographies. Information relating to the relationship between dosimetric parameters, patient demographics, tumor characteristics and radiation/chemotherapy treatment with RE were extracted and analyzed. Results: A total of 18 published studies were found to be suitable for analysis. Eleven of these studies assessed dosimetric parameters contributing to acute RE while the remainder assessed acute and chronic RE together. The overall published prevalence of RE
S717
I. J. Radiation Oncology d Biology d Physics
S718
Volume 69, Number 3, Supplement, 2007
(all grades) ranged from 6.9% to 79.1%. The dosimetric parameters that were frequently studied ($4 papers) included: percent of organ volume receiving greater than a threshold dose (Vdose), mean and maximum point dose delivered to the esophagus, and absolute length of esophagus included in the radiation field. Table 1 describes the dosimetric parameters that were studied and the percentage of those papers that demonstrated a significant correlation with grade 2 or greater RE. Heterogeneity of esophageal contouring practices, reported individual reported information, and RE outcome definitions exists in the literature. Few well-developed models including DVH metrics with or without other relevant prognostic factors to predict the risk of significant RE exist in the literature. Conclusions: We propose that future studies assessing this relationship should focus on a smaller subset of the available parameters (V10, V20, V30, V40, V50 and mean esophageal dose) that have shown consistent correlation between the DVH parameter and RE. A well-developed model would assist in routine radiation therapy planning and design of future clinical trials evaluating novel radiotherapeutic approaches and/or chemotherapeutic agents. Rigorous standardization of dosimetric parameter determination, contouring practices, and RE outcome definition will be critical in rationally improving the therapeutic ratio in this patient population. Table 1: World literature summary of the correlation of esophageal dosimetric parameters
Dosimetric Parameters
Number of Papers assessing relationship between parameter and RE
V50 Mean dose, V60 V40, V55, max dose V45 V65 V20, V30, esophageal length V10
n = 12 n=9 n=8 n=7 n=6 n=5 n=4
Respective Percentages of Statistically Significant Associations 75% 89%, 44% 75%, 63%, 63% 71% 33% 80%, 80%, 40% 75%
Author Disclosure: G. Rodrigues, None; J. Rose, None; D. D’Souza, None; M. Lock, None.
2921
Evaluation of Bone Inhomogeneity Correction by the Analytical Anisotropic Algorithm 1
J. Da Cruz , A. M. M. Vieira1,2, C. P. Lopes1, R. K. Sakuraba1, L. Caprioglio1 1 Hospital Israelita Albert Einstein, Sao Paulo, Brazil, 2Instituto de Pesquisas Energe´ticas e Nucleares, Sao Paulo, Brazil Purpose/Objective(s): A recently commissioned Analytical Anisotropic Algorithm–AAA for photon dose calculation was implemented in the Eclipse (Varian Medical Systems) Treatment Planning System–TPS. In order to evaluate the calculation in inhomogeneous tissue we investigated the dose distribution in a simple geometry containing a bone slab incorporated into a water phantom. The dose prediction, while considering inhomogeneity correction, for this configuration was evaluated in comparison to the reference values suggested by the American Association of Physicists in Medicine–AAPM task group 65 (TG65) and to a preliminary Monte Carlo–MC simulation performed with the Penelope package. The earlier Pencil Beam Convolution–PBC algorithm implemented in the TPS was also included in the comparisons. Materials/Methods: In this study we investigated the 6MV photon beam from the Varian Clinac 600C. The correction factors for bone inhomogeneity reviewed by TG65, based on experimental measurements, were compared to the ones predicted by the Eclipse algorithms for dose calculation. The calculated Percentage Depth Dose–PDD curves were compared to Monte Carlo simulations with the PENELOPE code. The MC method is known as an adequate tool for radiation beam studies. The same slab phantom from TG65 was virtually created in the TPS and in our MC user code. It is composed by a 3 cm thick bone slab immersed in a water cubic phantom at 3 cm depth perpendicular to the beam axis. A single 10 cm 10 cm field was assigned to the phantom surface and the depth dose distribution at the central axis was computed by the PBC algorithm using the modified Batho Power Law correction method, the AAA with the inhomogeneity correction method turned on, and by the Monte Carlo simulation. In order to determine the correction factors a second calculation in a homogeneous water phantom was performed in each case. The calculation grid size for both algorithms in the TPS was set into 2.5 mm. For the MC calculation the dose distribution was tallied into voxels of 1 cm 1 cm 0.5 cm, the radiation source was described by a previously Monte Carlo generated 6MV photon spectrum of the accelerator with a uniform distribution sample, and the delimiting jaws and the phantom were explicitly described. The MC simulation was performed in a 2.8 GHz PC with Pentium 4Ò processor, under WindowsÒ operational system. Results: Average differences between the correction factors for bone inhomogeneity and the reference values of the TG65 evaluated at depths beyond the inhomogeneity (from 6.0 cm to 15 cm depth) and through the central beam axis were (1.2 ± 0.8)% for the AAA and (1.9 ± 1.0)% for the PBC with Modified Batho Power Law correction method. Dose inside the inhomogeneity is overestimated in 5% to 6% by both algorithms. However, the depth dose distributions agree in a 3% level of statistical uncertainty from the preliminary MC calculation. Conclusions: The average differences of the inhomogeneity correction factors indicate that the TPS algorithms predict the dose deposition beyond the inhomogeneity with a satisfactory level of accuracy, but can overestimate the dose inside the inhomogeneity. The results also indicate that AAA improves the dose calculation accuracy. Author Disclosure: J. Da Cruz, None; A.M.M. Vieira, None; C.P. Lopes, None; R.K. Sakuraba, None; L. Caprioglio, None.
Fig. 1. Left Column (top to bottom): (1) Respiratory flowrate measured by spirometry (white) and chest wall motion measured using pressure sensors. (2–4) Air volume change versus pressure signal for sensors 1–3. Right Column: (1) Normalized waveforms (integrated flowrate and pressure signals). (2–4): Pair-wise comparison of pressure signals from the three sensors. Author Disclosure: J.D. Christensen, Equipment, C. Other Research Support; A. Tai, Equipment, C. Other Research Support; X.A. Li, Equipment, C. Other Research Support.
2920
Correlation of Clinically Significant Radiation Esophagitis With Dose-Volume Histogram Parameters in Lung Cancer
G. Rodrigues, J. Rose, D. D’Souza, M. Lock London Regional Cancer Program, London, ON, Canada Purpose/Objective(s): Due to dose escalation and increasing use of concurrent chemotherapy, radiation esophagitis (RE) remains a common treatment-limiting acute side effect. The aim of this review is to systematically assess the value of dosimetric parameters reported in the literature in predicting severity of RE and to provide recommendations for future research in the field. Materials/Methods: Both prospective and retrospective clinical studies assessing the relationship between various dosimetric parameters and RE in the treatment of inoperable lung cancers and thymomas were included in this systematic analysis. Our search strategy included a variety of electronic medical databases, textbooks and bibliographies. Information relating to the relationship between dosimetric parameters, patient demographics, tumor characteristics and radiation/chemotherapy treatment with RE were extracted and analyzed. Results: A total of 18 published studies were found to be suitable for analysis. Eleven of these studies assessed dosimetric parameters contributing to acute RE while the remainder assessed acute and chronic RE together. The overall published prevalence of RE
S717
I. J. Radiation Oncology d Biology d Physics
S718
Volume 69, Number 3, Supplement, 2007
(all grades) ranged from 6.9% to 79.1%. The dosimetric parameters that were frequently studied ($4 papers) included: percent of organ volume receiving greater than a threshold dose (Vdose), mean and maximum point dose delivered to the esophagus, and absolute length of esophagus included in the radiation field. Table 1 describes the dosimetric parameters that were studied and the percentage of those papers that demonstrated a significant correlation with grade 2 or greater RE. Heterogeneity of esophageal contouring practices, reported individual reported information, and RE outcome definitions exists in the literature. Few well-developed models including DVH metrics with or without other relevant prognostic factors to predict the risk of significant RE exist in the literature. Conclusions: We propose that future studies assessing this relationship should focus on a smaller subset of the available parameters (V10, V20, V30, V40, V50 and mean esophageal dose) that have shown consistent correlation between the DVH parameter and RE. A well-developed model would assist in routine radiation therapy planning and design of future clinical trials evaluating novel radiotherapeutic approaches and/or chemotherapeutic agents. Rigorous standardization of dosimetric parameter determination, contouring practices, and RE outcome definition will be critical in rationally improving the therapeutic ratio in this patient population. Table 1: World literature summary of the correlation of esophageal dosimetric parameters
Dosimetric Parameters
Number of Papers assessing relationship between parameter and RE
V50 Mean dose, V60 V40, V55, max dose V45 V65 V20, V30, esophageal length V10
n = 12 n=9 n=8 n=7 n=6 n=5 n=4
Respective Percentages of Statistically Significant Associations 75% 89%, 44% 75%, 63%, 63% 71% 33% 80%, 80%, 40% 75%
Author Disclosure: G. Rodrigues, None; J. Rose, None; D. D’Souza, None; M. Lock, None.
2921
Evaluation of Bone Inhomogeneity Correction by the Analytical Anisotropic Algorithm 1
J. Da Cruz , A. M. M. Vieira1,2, C. P. Lopes1, R. K. Sakuraba1, L. Caprioglio1 1 Hospital Israelita Albert Einstein, Sao Paulo, Brazil, 2Instituto de Pesquisas Energe´ticas e Nucleares, Sao Paulo, Brazil Purpose/Objective(s): A recently commissioned Analytical Anisotropic Algorithm–AAA for photon dose calculation was implemented in the Eclipse (Varian Medical Systems) Treatment Planning System–TPS. In order to evaluate the calculation in inhomogeneous tissue we investigated the dose distribution in a simple geometry containing a bone slab incorporated into a water phantom. The dose prediction, while considering inhomogeneity correction, for this configuration was evaluated in comparison to the reference values suggested by the American Association of Physicists in Medicine–AAPM task group 65 (TG65) and to a preliminary Monte Carlo–MC simulation performed with the Penelope package. The earlier Pencil Beam Convolution–PBC algorithm implemented in the TPS was also included in the comparisons. Materials/Methods: In this study we investigated the 6MV photon beam from the Varian Clinac 600C. The correction factors for bone inhomogeneity reviewed by TG65, based on experimental measurements, were compared to the ones predicted by the Eclipse algorithms for dose calculation. The calculated Percentage Depth Dose–PDD curves were compared to Monte Carlo simulations with the PENELOPE code. The MC method is known as an adequate tool for radiation beam studies. The same slab phantom from TG65 was virtually created in the TPS and in our MC user code. It is composed by a 3 cm thick bone slab immersed in a water cubic phantom at 3 cm depth perpendicular to the beam axis. A single 10 cm 10 cm field was assigned to the phantom surface and the depth dose distribution at the central axis was computed by the PBC algorithm using the modified Batho Power Law correction method, the AAA with the inhomogeneity correction method turned on, and by the Monte Carlo simulation. In order to determine the correction factors a second calculation in a homogeneous water phantom was performed in each case. The calculation grid size for both algorithms in the TPS was set into 2.5 mm. For the MC calculation the dose distribution was tallied into voxels of 1 cm 1 cm 0.5 cm, the radiation source was described by a previously Monte Carlo generated 6MV photon spectrum of the accelerator with a uniform distribution sample, and the delimiting jaws and the phantom were explicitly described. The MC simulation was performed in a 2.8 GHz PC with Pentium 4Ò processor, under WindowsÒ operational system. Results: Average differences between the correction factors for bone inhomogeneity and the reference values of the TG65 evaluated at depths beyond the inhomogeneity (from 6.0 cm to 15 cm depth) and through the central beam axis were (1.2 ± 0.8)% for the AAA and (1.9 ± 1.0)% for the PBC with Modified Batho Power Law correction method. Dose inside the inhomogeneity is overestimated in 5% to 6% by both algorithms. However, the depth dose distributions agree in a 3% level of statistical uncertainty from the preliminary MC calculation. Conclusions: The average differences of the inhomogeneity correction factors indicate that the TPS algorithms predict the dose deposition beyond the inhomogeneity with a satisfactory level of accuracy, but can overestimate the dose inside the inhomogeneity. The results also indicate that AAA improves the dose calculation accuracy. Author Disclosure: J. Da Cruz, None; A.M.M. Vieira, None; C.P. Lopes, None; R.K. Sakuraba, None; L. Caprioglio, None.