Markerless Tumor Tracking using Fast-kV Switching Dual Energy Imaging with the On-Board Imager of a Commercial Linac

Abstract

Markerless tumor tracking (MTT) using template matching is currently being considered for real-time lung tumor tracking using single energy (SE) fluoroscopy. A limitation of MTT is that the tumor has to be clearly visible on images. In cases where the tumor is obscured by bony anatomy, dual energy (DE) imaging may be used to create a soft tissue only image, improving tumor visibility as well as the accuracy of MTT. The goal of this study is to evaluate MTT with fast-kV switching DE imaging using the on-board imager (OBI) of a commercial linear accelerator. Fast-kV switching DE fluoroscopy was implemented on the OBI of a commercial linac. This technique produces x-ray pulses that alternate between programmed tube voltages (ie, 120 and 60 kVp) at a rate of 15 Hz. To evaluate MTT with DE fluoroscopy, a motion phantom, consisting of a torso with embedded ribs and spine along with a cavity having lung-equivalent density, was used. Simulated tumors (5-25 mm diameter) were placed inside the lung equivalent compartment of the phantom, and then programmed to simulate breathing. Three categories of motion with varying periods were considered: static (ST), slow-motion (SM) - 5 sec period and fast-motion (FM) - 2.5 sec period. While the target was moving, fast-kV switching DE images were acquired over 360° of rotation allowing the tumor projection to overlap with varying amounts of bone. Weighted logarithmic subtraction was performed on consecutive high-low projections to produce soft tissue images. A template-based matching algorithm was then used to track target motion on both DE and SE sequences. Successful matching was defined as any instance in which the algorithm matched the template on the image. Tumor tracking coordinates were also evaluated against ground truth motion using the root-mean-squared error (RMSE). For the 5 mm target, the success rates ranged from 11-17% for SE vs. 36-63% for DE imaging, for the three types of motion considered. Because of the low success rates, RMSE were not evaluated for this target. The 10 mm target demonstrated success rates of 66-73% and 87-90% for SE and DE, respectively. The corresponding RMSE for SE were 2.7, 2.9 and 3.4 mm, while for DE they were 0.2, 1.0 and 0.5 mm, for ST, SM and FM, respectively. For the 15 mm target, SE success rates ranged from 84-89% vs. 90-94% for DE. RMSE for SE was 0.5, 2.0 and 1.4 mm vs. 0.2, 0.5, and 0.2 mm for DE for ST, SM and FM, respectively. For the larger tumors (20 and 25 mm), the success rates were > 90% for both SE and DE. RMSE for SE and DE were <1.1 mm and <0.4 mm, respectively, for all motion categories. Fast-kV switching DE imaging has been implemented on the OBI of a commercial linac. Overall, MTT with DE imaging results in higher success rates and lower RMSE than using SE imaging alone. The approach allows for the accurate tracking of small tumors where MTT fails on SE imaging due to overlapping bony anatomy.