Comparison Of Respiratory Gate Triggered Kilovoltage (Kv) Portal Images (Gtpi) Versus Kv Cone-Beam Computed Tomography Images (Cbct) For Patient Alignment And Verification In Thoracic Stereotactic Body Radiotherapy (Sbrt) Patients With Implanted Fiducial Markers

Purpose/Objective(s)SBRT is currently being used for the treatment of medically inoperable non-small cell lung cancer. Accurate delivery of radiation to the target with this modality is vital to success. Current on-board imaging methods for target localization such as CBCT frequently are not respiratory gated and do not account for tumor motion with respiration. We investigated target localization differences between gate triggered kV portal images (GTPI) and nongated kV cone-beam computed tomography (CBCT) images.Materials/MethodsA retrospective review was conducted of all SBRT cases at the University of Alabama at Birmingham from December 2007 until current. SBRT cases were defined as 1-5 fractions of highly conformal radiation 8 Gy or greater. Cases selected for analysis had fiducial markers placed by an endobronchial technique utilizing electromagnetic navigation bronchoscopy (superDimension, Inc, Minneapolis, MN) and target alignment using both CBCT and GTPI. Tumor target definitions and treatment plans were developed by a multidisciplinary team of radiation oncology and thoracic surgery physicians. For each fraction of radiation delivered, the radiation oncologist and/or thoracic surgeon registered the CBCT and GTPI with the treatment planning 4DCT prior to treatment. The registration differences between the CBCT and GTPI were retrieved.Results103 cases of SBRT were identified of which, 7 thoracic cases were found to have fiducial marker implantation along with both CBCT and GTPI. The seven patients had 32 sets of imaging. Three patients had kV portal images followed by CBCT and 4 patients had CBCT followed by kV portal images for final alignment prior to treatment. The mean and standard deviation of the absolute values of the differences between the CBCT and the kV corrections were 2.2 ± 2.3, 2.9 ± 3.3, and 3.7 ± 7.7 mm for the anteroposterior, superoinferior, and lateral directions, respectively. Corrections > 5 mm occurred in 11 (34%) of the image sets with 18 (56%) having corrections ≥ 5 mm.ConclusionsCBCT and GTPI using implanted fiducial markers results in similar target localization for most cases; however, 34% of alignments have discrepancies over 5 mm. Our data indicate that while CBCT and GTPI alignment agree in most cases, a sufficient number of cases do not, thus warranting gate triggered onboard imaging to ensure accurate dose delivery to the tumor target. Purpose/Objective(s)SBRT is currently being used for the treatment of medically inoperable non-small cell lung cancer. Accurate delivery of radiation to the target with this modality is vital to success. Current on-board imaging methods for target localization such as CBCT frequently are not respiratory gated and do not account for tumor motion with respiration. We investigated target localization differences between gate triggered kV portal images (GTPI) and nongated kV cone-beam computed tomography (CBCT) images. SBRT is currently being used for the treatment of medically inoperable non-small cell lung cancer. Accurate delivery of radiation to the target with this modality is vital to success. Current on-board imaging methods for target localization such as CBCT frequently are not respiratory gated and do not account for tumor motion with respiration. We investigated target localization differences between gate triggered kV portal images (GTPI) and nongated kV cone-beam computed tomography (CBCT) images. Materials/MethodsA retrospective review was conducted of all SBRT cases at the University of Alabama at Birmingham from December 2007 until current. SBRT cases were defined as 1-5 fractions of highly conformal radiation 8 Gy or greater. Cases selected for analysis had fiducial markers placed by an endobronchial technique utilizing electromagnetic navigation bronchoscopy (superDimension, Inc, Minneapolis, MN) and target alignment using both CBCT and GTPI. Tumor target definitions and treatment plans were developed by a multidisciplinary team of radiation oncology and thoracic surgery physicians. For each fraction of radiation delivered, the radiation oncologist and/or thoracic surgeon registered the CBCT and GTPI with the treatment planning 4DCT prior to treatment. The registration differences between the CBCT and GTPI were retrieved. A retrospective review was conducted of all SBRT cases at the University of Alabama at Birmingham from December 2007 until current. SBRT cases were defined as 1-5 fractions of highly conformal radiation 8 Gy or greater. Cases selected for analysis had fiducial markers placed by an endobronchial technique utilizing electromagnetic navigation bronchoscopy (superDimension, Inc, Minneapolis, MN) and target alignment using both CBCT and GTPI. Tumor target definitions and treatment plans were developed by a multidisciplinary team of radiation oncology and thoracic surgery physicians. For each fraction of radiation delivered, the radiation oncologist and/or thoracic surgeon registered the CBCT and GTPI with the treatment planning 4DCT prior to treatment. The registration differences between the CBCT and GTPI were retrieved. Results103 cases of SBRT were identified of which, 7 thoracic cases were found to have fiducial marker implantation along with both CBCT and GTPI. The seven patients had 32 sets of imaging. Three patients had kV portal images followed by CBCT and 4 patients had CBCT followed by kV portal images for final alignment prior to treatment. The mean and standard deviation of the absolute values of the differences between the CBCT and the kV corrections were 2.2 ± 2.3, 2.9 ± 3.3, and 3.7 ± 7.7 mm for the anteroposterior, superoinferior, and lateral directions, respectively. Corrections > 5 mm occurred in 11 (34%) of the image sets with 18 (56%) having corrections ≥ 5 mm. 103 cases of SBRT were identified of which, 7 thoracic cases were found to have fiducial marker implantation along with both CBCT and GTPI. The seven patients had 32 sets of imaging. Three patients had kV portal images followed by CBCT and 4 patients had CBCT followed by kV portal images for final alignment prior to treatment. The mean and standard deviation of the absolute values of the differences between the CBCT and the kV corrections were 2.2 ± 2.3, 2.9 ± 3.3, and 3.7 ± 7.7 mm for the anteroposterior, superoinferior, and lateral directions, respectively. Corrections > 5 mm occurred in 11 (34%) of the image sets with 18 (56%) having corrections ≥ 5 mm. ConclusionsCBCT and GTPI using implanted fiducial markers results in similar target localization for most cases; however, 34% of alignments have discrepancies over 5 mm. Our data indicate that while CBCT and GTPI alignment agree in most cases, a sufficient number of cases do not, thus warranting gate triggered onboard imaging to ensure accurate dose delivery to the tumor target. CBCT and GTPI using implanted fiducial markers results in similar target localization for most cases; however, 34% of alignments have discrepancies over 5 mm. Our data indicate that while CBCT and GTPI alignment agree in most cases, a sufficient number of cases do not, thus warranting gate triggered onboard imaging to ensure accurate dose delivery to the tumor target.