Portal de Periódicos da CAPES

Effect Of ITV-based Treatment Planning And Delivery For Moving Lung Tumors In Comparison To Real-time Target Tracking

2011; Elsevier BV; Volume: 81; Issue: 2 Linguagem: Inglês

10.1016/j.ijrobp.2011.06.1407

1879-355X

Georges Hobeika, Z. Wang, Elizabeth Bossart, Xiaodong Wu,

Lung Cancer Diagnosis and Treatment

Purpose/Objective(s)It is established that real-time target tracking is an ideal method for treating moving lung tumors. In its absence, 4D-CT simulation with generation of an ITV has been considered an acceptable alternative. However there are no data regarding the choice of an optimal density to the ITV for dose planning that would result in identical in situ dose deposition. The purpose of this study is to compare in situ target coverage for lung tumors treated with real-time tracking vs ITV-based plan and treatment delivery.Materials/MethodsA dynamic lung phantom by CIRS, Norfolk, VA was used with a moving spherical target of 25mm in diameter, and of soft-tissue density with a periodic (5 sec) cranio-caudal movement of 3 cm. For the ITV-based approach, an ITV was generated using 4D-CT images. Two conformal arc treatment plans were created using Pinnacle-3 treatment planning system with a superposition convolution algorithm, referred to as CR1 and CR2, calculated with and without density correction for ITV respectively. A Gafchromic EB-2 film by ISP was cut to fit the mid sagittal plane of the target to record the dose. This plane was contoured as ITV1 for DVH comparison. Dosimetric analysis was performed using a calibrated H_D curve. 700cGy was arbitrarily prescribed to the isodose line encompassing 86% of the ITV1 for CR1 and 88% of the ITV1 of CR2. 10 fractions of each plan were delivered using a Varian Trilogy LINAC with 6 MV beams. The real-time tracking approach was done using CyberKnife system with Monte Carlo dose computation algorithm for planning and the Synchrony tracking system delivery. Similar dose was prescribed to encompass100% of the GTV and 10 fractions were delivered. An in-house MatLab code was used to analyze the scanned films and generate DVHs.ResultsFor the ITV-based treatment, the measured average ± standard deviation for the V100 was 90.2 ± 1% for CR1 compared to a calculated 86% and 98.2 ± 0.5% for CR2 compared to a calculated 88%. The measured data showed a difference of 4% in dose heterogeneity compared to calculated. For the CyberKnife-based treatment, the measured V100 was 100%±1% and was consistent with calculated coverage of 100%.ConclusionsWhile an ITV-based treatment of moving lung tumors is conceptually sound, there is apparent uncertainty in the determination of the ITV density. This could result in significant discrepancy between calculated and measured coverage. Our study confirmed this and revealed more significant difference when the ITV density correction was not applied. These data highlight the need for better density modeling algorithms for the ITV-based treatment and reinforce the superiority of real-time tracking in the treatment of moving targets, especially in a highly heterogeneous environment. Purpose/Objective(s)It is established that real-time target tracking is an ideal method for treating moving lung tumors. In its absence, 4D-CT simulation with generation of an ITV has been considered an acceptable alternative. However there are no data regarding the choice of an optimal density to the ITV for dose planning that would result in identical in situ dose deposition. The purpose of this study is to compare in situ target coverage for lung tumors treated with real-time tracking vs ITV-based plan and treatment delivery. It is established that real-time target tracking is an ideal method for treating moving lung tumors. In its absence, 4D-CT simulation with generation of an ITV has been considered an acceptable alternative. However there are no data regarding the choice of an optimal density to the ITV for dose planning that would result in identical in situ dose deposition. The purpose of this study is to compare in situ target coverage for lung tumors treated with real-time tracking vs ITV-based plan and treatment delivery. Materials/MethodsA dynamic lung phantom by CIRS, Norfolk, VA was used with a moving spherical target of 25mm in diameter, and of soft-tissue density with a periodic (5 sec) cranio-caudal movement of 3 cm. For the ITV-based approach, an ITV was generated using 4D-CT images. Two conformal arc treatment plans were created using Pinnacle-3 treatment planning system with a superposition convolution algorithm, referred to as CR1 and CR2, calculated with and without density correction for ITV respectively. A Gafchromic EB-2 film by ISP was cut to fit the mid sagittal plane of the target to record the dose. This plane was contoured as ITV1 for DVH comparison. Dosimetric analysis was performed using a calibrated H_D curve. 700cGy was arbitrarily prescribed to the isodose line encompassing 86% of the ITV1 for CR1 and 88% of the ITV1 of CR2. 10 fractions of each plan were delivered using a Varian Trilogy LINAC with 6 MV beams. The real-time tracking approach was done using CyberKnife system with Monte Carlo dose computation algorithm for planning and the Synchrony tracking system delivery. Similar dose was prescribed to encompass100% of the GTV and 10 fractions were delivered. An in-house MatLab code was used to analyze the scanned films and generate DVHs. A dynamic lung phantom by CIRS, Norfolk, VA was used with a moving spherical target of 25mm in diameter, and of soft-tissue density with a periodic (5 sec) cranio-caudal movement of 3 cm. For the ITV-based approach, an ITV was generated using 4D-CT images. Two conformal arc treatment plans were created using Pinnacle-3 treatment planning system with a superposition convolution algorithm, referred to as CR1 and CR2, calculated with and without density correction for ITV respectively. A Gafchromic EB-2 film by ISP was cut to fit the mid sagittal plane of the target to record the dose. This plane was contoured as ITV1 for DVH comparison. Dosimetric analysis was performed using a calibrated H_D curve. 700cGy was arbitrarily prescribed to the isodose line encompassing 86% of the ITV1 for CR1 and 88% of the ITV1 of CR2. 10 fractions of each plan were delivered using a Varian Trilogy LINAC with 6 MV beams. The real-time tracking approach was done using CyberKnife system with Monte Carlo dose computation algorithm for planning and the Synchrony tracking system delivery. Similar dose was prescribed to encompass100% of the GTV and 10 fractions were delivered. An in-house MatLab code was used to analyze the scanned films and generate DVHs. ResultsFor the ITV-based treatment, the measured average ± standard deviation for the V100 was 90.2 ± 1% for CR1 compared to a calculated 86% and 98.2 ± 0.5% for CR2 compared to a calculated 88%. The measured data showed a difference of 4% in dose heterogeneity compared to calculated. For the CyberKnife-based treatment, the measured V100 was 100%±1% and was consistent with calculated coverage of 100%. For the ITV-based treatment, the measured average ± standard deviation for the V100 was 90.2 ± 1% for CR1 compared to a calculated 86% and 98.2 ± 0.5% for CR2 compared to a calculated 88%. The measured data showed a difference of 4% in dose heterogeneity compared to calculated. For the CyberKnife-based treatment, the measured V100 was 100%±1% and was consistent with calculated coverage of 100%. ConclusionsWhile an ITV-based treatment of moving lung tumors is conceptually sound, there is apparent uncertainty in the determination of the ITV density. This could result in significant discrepancy between calculated and measured coverage. Our study confirmed this and revealed more significant difference when the ITV density correction was not applied. These data highlight the need for better density modeling algorithms for the ITV-based treatment and reinforce the superiority of real-time tracking in the treatment of moving targets, especially in a highly heterogeneous environment. While an ITV-based treatment of moving lung tumors is conceptually sound, there is apparent uncertainty in the determination of the ITV density. This could result in significant discrepancy between calculated and measured coverage. Our study confirmed this and revealed more significant difference when the ITV density correction was not applied. These data highlight the need for better density modeling algorithms for the ITV-based treatment and reinforce the superiority of real-time tracking in the treatment of moving targets, especially in a highly heterogeneous environment.