Colien Hazelaar1, Wilko F A R Verbakel2, Hassan Mostafavi3, Lineke van der Weide2, Ben J Slotman2, Max Dahele2. 1. Department of Radiation Oncology, Cancer Center Amsterdam, VU University Medical Center, Amsterdam, The Netherlands. Electronic address: c.hazelaar@vumc.nl. 2. Department of Radiation Oncology, Cancer Center Amsterdam, VU University Medical Center, Amsterdam, The Netherlands. 3. Varian Medical Systems, Palo Alto, California.
Abstract
PURPOSE: The purpose was to report our initial experience with online markerless 3-dimensional (3D) spine position monitoring. We used template matching plus triangulation of fluoroscopic kilovoltage images acquired with the gantry-mounted imager during flattening filter-free volumetric modulated arc spine stereotactic body radiation therapy delivery on a conventional linear accelerator. METHODS AND MATERIALS: Kilovoltage images were acquired at 7 frames per second and streamed to a stand-alone computer. Two-dimensional templates (1/°) containing the clinical target volume were generated from planning computed tomography (CT) data before the first fraction and matched to the (prefiltered) kilovoltage images during treatment. Each 2-dimensional registration was triangulated with multiple previous registrations, resulting in the 3D spine position offset from the planned position in real time during treatment. If the offset was more than a certain threshold, the treatment was manually stopped and a cone beam CT scan was acquired to reposition the patient. RESULTS: During irradiation of 10 fractions in 3 patients, images were analyzed at an average rate of 1.0 to 1.3 frames per second; all other frames were excluded from the analysis because of limitations in processing speed. As a result of the start-up period of triangulation and poorer image quality at the start of treatment (lateral imaging angles), the first 3D position was determined after an average of 4.9 seconds. On the basis of the position results, we interrupted the treatment beam 2 times for different patients. In all cases the spine position results corresponded well with the CT-cone beam CT match values used for subsequent repositioning. CONCLUSIONS: For the first time, we have determined the spine position during stereotactic body radiation therapy delivery on a standard linear accelerator using the gantry-mounted kilovoltage imager. This has the potential to increase confidence in the treatment, and the need for 2 treatment interruptions demonstrates the benefit of monitoring during irradiation. However, software improvements are needed to increase processing speed.
PURPOSE: The purpose was to report our initial experience with online markerless 3-dimensional (3D) spine position monitoring. We used template matching plus triangulation of fluoroscopic kilovoltage images acquired with the gantry-mounted imager during flattening filter-free volumetric modulated arc spine stereotactic body radiation therapy delivery on a conventional linear accelerator. METHODS AND MATERIALS: Kilovoltage images were acquired at 7 frames per second and streamed to a stand-alone computer. Two-dimensional templates (1/°) containing the clinical target volume were generated from planning computed tomography (CT) data before the first fraction and matched to the (prefiltered) kilovoltage images during treatment. Each 2-dimensional registration was triangulated with multiple previous registrations, resulting in the 3D spine position offset from the planned position in real time during treatment. If the offset was more than a certain threshold, the treatment was manually stopped and a cone beam CT scan was acquired to reposition the patient. RESULTS: During irradiation of 10 fractions in 3 patients, images were analyzed at an average rate of 1.0 to 1.3 frames per second; all other frames were excluded from the analysis because of limitations in processing speed. As a result of the start-up period of triangulation and poorer image quality at the start of treatment (lateral imaging angles), the first 3D position was determined after an average of 4.9 seconds. On the basis of the position results, we interrupted the treatment beam 2 times for different patients. In all cases the spine position results corresponded well with the CT-cone beam CT match values used for subsequent repositioning. CONCLUSIONS: For the first time, we have determined the spine position during stereotactic body radiation therapy delivery on a standard linear accelerator using the gantry-mounted kilovoltage imager. This has the potential to increase confidence in the treatment, and the need for 2 treatment interruptions demonstrates the benefit of monitoring during irradiation. However, software improvements are needed to increase processing speed.
Authors: Mahmud Mossa-Basha; Peter C Gerszten; Sten Myrehaug; Nina A Mayr; William Tc Yuh; Pejman Jabehdar Maralani; Arjun Sahgal; Simon S Lo Journal: Br J Radiol Date: 2019-07-25 Impact factor: 3.039
Authors: Wei Zhao; Bin Han; Yong Yang; Mark Buyyounouski; Steven L Hancock; Hilary Bagshaw; Lei Xing Journal: Radiother Oncol Date: 2019-07-11 Impact factor: 6.280
Authors: David Y Hu; Yiwen Xu; Yu-Hui Chen; Marjan Khosravi; Yulia Lyatskaya; Jeremy S Bredfeldt; Fred L Hacker; Tracy A Balboni; Alexander Spektor; Daniel Cagney; Raymond Mak; Mai Anh Huynh Journal: Adv Radiat Oncol Date: 2022-04-08