- Email: [email protected]
Estimation of Dose Distribution in Lung during Total Body Irradiation
I. J. Radiation Oncology d Biology d Physics
S638
2972
Volume 78, Number 3, Supplement, 2010
Estimation of Dose Distribution in Lung during Total Body Irradiation
A. Maciejczyk, M. Janiszewska, M. Raczkowski Lower Silesian Onkology Center, 53-413, Poland Purpose/Objective(s): Purpose of this work is estimation, doses distribution in critical in TBI treatment region -lungs, by using 3D planning methods. Materials/Methods: TBI is a complementary method applied in order to prepare a patient for bone marrow transplantation. At our institution - Radiotherapy Department of Lower-Silesia Center of Oncology in Wroclaw - the dose of 12Gy for the whole body is given in 6 fractions (2Gy per fraction, twice a day with 6-hour-break). Results: The fractions #1 and #3 (so called AP/PA fractions) are delivered from SSD = 125cm and the patient is positioned on the accelerator catch. The fractions #2, #4, #5 and #6 are applied from SSD = 450cm, with the gantry set up on 270 (to the left and the right side of the patient) and the patient is situated on the special table dedicated to this technique. For those lateral exposures the patient is positioned supine centrally in the plexi box filled with rise boluses and irradiation time for these fractions is calculated from 2 variables: appropriate PDD (on the surface it equals 78.0%, depth of max 2.5cm) and the factor of profile uniformity (the flatness of the profile below 5.0% and symmetry 2.0%.). These curves PDD and Profile, were measured for photons 18MeV and square field 180x180cm. The dose 1.0Gy (1/2fraction) is specified at the depth of 26cm, it means in the distance to the centre of the plexi box in which the patient is positioned. PDD at this point equals 60.2%,which means the dose in maximum depth is 1.6611Gy. This, for the dose rate 0.000634Gy/MU, determines time 2620MU. Using calculated treatment time and established distances from CT (plexi - skin of the patient) we evaluate the distributions of output and input doses within reference body levels of: head, neck, shoulders, lungs, navel, genitals, knees and feet. The plan for AP/PA fractions is prepared in 3D planning system Eclips on the basis of image data from CT. The first pair of the fields AP/PA of the size 50x50cm includes the area of head and chest. For these fields portals are prepared allowing the verification of the patient’s geometry and the position of the blocks for the lung. Complementary dose for chest is made with the use of electron beams whose shape, energy and normalization is prepared according to 3D planning rules. The distances between successive pairs of the fields AP/PA are appointed on the basis of dose distribution in vertical reconstruction. Conclusions: Using 3D method for treatment planning in TBI, we can reduce doses in lung for level 8-9 Gy. Author Disclosure: A. Maciejczyk, None; M. Janiszewska, None; M. Raczkowski, None.
2973
GPU-based Iterative CT Reconstruction via Edge-preserving Total Variation Regula
Z. Tian1,2, X. Jia1, T. Pan3, S. Jiang1 1 Department of Radiation Oncology, University of California, San Diego, 92093, CA, 2Department of Biomedical Engineering, Tsinghua University, Beijing, China, 3Department of Imaging Physics, The University of Texas, Houston, 77030, TX
Purpose/Objective(s): High radiation dose in CT scans increases a lifetime risk of cancer and is always a major concern. Two possible ways to decrease the dose are using fewer projections for reconstruction or lower the X-ray tube current. The former may require new hardware while the latter can be readily implemented. However, conventional filtered backprojection (FBP) reconstruction algorithms failed in both ways. Recently, iterative reconstruction algorithm with Total Variation (TV) regularization has been developed to reduce the imaging dose. Nonetheless, CT images reconstructed in this approach are sometimes oversmoothed and the edge information is lost. In this work, we developed an iterative CT reconstruction algorithm with edge preserving TV regularization to accurately reconstruct CT images from highly undersampled data obtained at low dose. Materials/Methods: The CT image is reconstructed iteratively by minimizing TV norm under a constraint posed by few X-ray projections. To avoid over-smoothed edge, an edge-preserving penalty was proposed and added to TV norm to preferentially perform smoothing on those non-edge pixels. Our reconstruction algorithm is implemented on GPU which enables a number of tasks processed in parallel simultaneously to speed up the process. We tested our reconstruction algorithm on two types of data, digital phantoms (Shepp-Logan head phantom and NCAT phantom at thorax region) and sinogram data of a pelvis phantom and a chest phantom obtained at different mAs levels. Sinogram of the digital phantoms were simulated using Siddon’s ray tracing algorithm. Conventional FBP algorithm and conventional TV regularization method without edge preserving term were also studied for comparison purpose. Results: It is found that in our method 40 equi-angle projections are enough for reconstructing the CT images of digital phantoms in digital phantom experiments. For real sinogram data, 160 projections are needed for CT reconstruction at 50mAs and 200 projections are needed at 20mAs. Aliasing artifacts and large image noises appeared in images reconstructed by FBP with few projections at low mAs. The reconstructed images using our approach contain sharper edges and less noise than those using TV method. The algorithm was speeded up by about 20 times with GPU implementation. Conclusions: The edge-preserving TV method enables CT reconstruction with under-sampled projections obtained at low mAs, which implies a possible reduction of exposure dose to patients. Our method outperforms the conventional TV methods in preserving edges and fine structures. Author Disclosure: Z. Tian, None; X. Jia, None; T. Pan, None; S. Jiang, None.
2974
Impact of Caveolin-1 Overexpression on the Radiation Response of TK6 Cells
D. Barzan, P. Maier, F. Wenz, C. Herskind Department of Radiation Oncology, University Medical Centre Mannheim, Mannheim, Germany Purpose/Objective(s): Caveolin-1 (CAV1) is the main structural component of caveolae in the plasma membrane and plays a pivotal role in endocytosis and signal transduction pathways. A radioprotective effect of CAV1 has been reported in tumor cell lines.
S638
2972
Volume 78, Number 3, Supplement, 2010
Estimation of Dose Distribution in Lung during Total Body Irradiation
A. Maciejczyk, M. Janiszewska, M. Raczkowski Lower Silesian Onkology Center, 53-413, Poland Purpose/Objective(s): Purpose of this work is estimation, doses distribution in critical in TBI treatment region -lungs, by using 3D planning methods. Materials/Methods: TBI is a complementary method applied in order to prepare a patient for bone marrow transplantation. At our institution - Radiotherapy Department of Lower-Silesia Center of Oncology in Wroclaw - the dose of 12Gy for the whole body is given in 6 fractions (2Gy per fraction, twice a day with 6-hour-break). Results: The fractions #1 and #3 (so called AP/PA fractions) are delivered from SSD = 125cm and the patient is positioned on the accelerator catch. The fractions #2, #4, #5 and #6 are applied from SSD = 450cm, with the gantry set up on 270 (to the left and the right side of the patient) and the patient is situated on the special table dedicated to this technique. For those lateral exposures the patient is positioned supine centrally in the plexi box filled with rise boluses and irradiation time for these fractions is calculated from 2 variables: appropriate PDD (on the surface it equals 78.0%, depth of max 2.5cm) and the factor of profile uniformity (the flatness of the profile below 5.0% and symmetry 2.0%.). These curves PDD and Profile, were measured for photons 18MeV and square field 180x180cm. The dose 1.0Gy (1/2fraction) is specified at the depth of 26cm, it means in the distance to the centre of the plexi box in which the patient is positioned. PDD at this point equals 60.2%,which means the dose in maximum depth is 1.6611Gy. This, for the dose rate 0.000634Gy/MU, determines time 2620MU. Using calculated treatment time and established distances from CT (plexi - skin of the patient) we evaluate the distributions of output and input doses within reference body levels of: head, neck, shoulders, lungs, navel, genitals, knees and feet. The plan for AP/PA fractions is prepared in 3D planning system Eclips on the basis of image data from CT. The first pair of the fields AP/PA of the size 50x50cm includes the area of head and chest. For these fields portals are prepared allowing the verification of the patient’s geometry and the position of the blocks for the lung. Complementary dose for chest is made with the use of electron beams whose shape, energy and normalization is prepared according to 3D planning rules. The distances between successive pairs of the fields AP/PA are appointed on the basis of dose distribution in vertical reconstruction. Conclusions: Using 3D method for treatment planning in TBI, we can reduce doses in lung for level 8-9 Gy. Author Disclosure: A. Maciejczyk, None; M. Janiszewska, None; M. Raczkowski, None.
2973
GPU-based Iterative CT Reconstruction via Edge-preserving Total Variation Regula
Z. Tian1,2, X. Jia1, T. Pan3, S. Jiang1 1 Department of Radiation Oncology, University of California, San Diego, 92093, CA, 2Department of Biomedical Engineering, Tsinghua University, Beijing, China, 3Department of Imaging Physics, The University of Texas, Houston, 77030, TX
Purpose/Objective(s): High radiation dose in CT scans increases a lifetime risk of cancer and is always a major concern. Two possible ways to decrease the dose are using fewer projections for reconstruction or lower the X-ray tube current. The former may require new hardware while the latter can be readily implemented. However, conventional filtered backprojection (FBP) reconstruction algorithms failed in both ways. Recently, iterative reconstruction algorithm with Total Variation (TV) regularization has been developed to reduce the imaging dose. Nonetheless, CT images reconstructed in this approach are sometimes oversmoothed and the edge information is lost. In this work, we developed an iterative CT reconstruction algorithm with edge preserving TV regularization to accurately reconstruct CT images from highly undersampled data obtained at low dose. Materials/Methods: The CT image is reconstructed iteratively by minimizing TV norm under a constraint posed by few X-ray projections. To avoid over-smoothed edge, an edge-preserving penalty was proposed and added to TV norm to preferentially perform smoothing on those non-edge pixels. Our reconstruction algorithm is implemented on GPU which enables a number of tasks processed in parallel simultaneously to speed up the process. We tested our reconstruction algorithm on two types of data, digital phantoms (Shepp-Logan head phantom and NCAT phantom at thorax region) and sinogram data of a pelvis phantom and a chest phantom obtained at different mAs levels. Sinogram of the digital phantoms were simulated using Siddon’s ray tracing algorithm. Conventional FBP algorithm and conventional TV regularization method without edge preserving term were also studied for comparison purpose. Results: It is found that in our method 40 equi-angle projections are enough for reconstructing the CT images of digital phantoms in digital phantom experiments. For real sinogram data, 160 projections are needed for CT reconstruction at 50mAs and 200 projections are needed at 20mAs. Aliasing artifacts and large image noises appeared in images reconstructed by FBP with few projections at low mAs. The reconstructed images using our approach contain sharper edges and less noise than those using TV method. The algorithm was speeded up by about 20 times with GPU implementation. Conclusions: The edge-preserving TV method enables CT reconstruction with under-sampled projections obtained at low mAs, which implies a possible reduction of exposure dose to patients. Our method outperforms the conventional TV methods in preserving edges and fine structures. Author Disclosure: Z. Tian, None; X. Jia, None; T. Pan, None; S. Jiang, None.
2974
Impact of Caveolin-1 Overexpression on the Radiation Response of TK6 Cells
D. Barzan, P. Maier, F. Wenz, C. Herskind Department of Radiation Oncology, University Medical Centre Mannheim, Mannheim, Germany Purpose/Objective(s): Caveolin-1 (CAV1) is the main structural component of caveolae in the plasma membrane and plays a pivotal role in endocytosis and signal transduction pathways. A radioprotective effect of CAV1 has been reported in tumor cell lines.