|ECR 2018 / C-1615||
|A global radiation awareness system using augmented reality and Monte Carlo simulations|
Demonstrations of our prototype radiation awareness system have been performed to our collaborating clinical and industrial partners. Initial feedback has been positive since the system provides in-situ feedback about radiation exposure in an interventional room in an intuitive manner. The potential of the system to be employed as a tool to teach about radiation's diffusion effects has also been acknowledged in the demonstrations. Indeed, using such a system would enable to teach trainees in real clinical conditions. The provided visual feedback about radiation exposure can render more intuitive the current training sessions, as these are based on students gathering and comparing measurements from dosimeters for different room lay-outs and C-arm configurations.
We also performed a set of dose measurements in an interventional room (IHU Strasbourg) using an Artis Zeego robotized X-ray imaging device and a set of RaySafe wireless active personal dosimeters to evaluate our radiation simulation approach. A plexiglas phantom of 20x20x24 cm, with 10 mm thick plexiglas walls and filled with water was irradiated under different imaging protocols. Eight dosimeters were either placed over the operating table or taped to drip rods and placed around the work area. The evaluation setup is shown below in figure 7.
References: Loy Rodas N., 2018. Context-aware radiation protection for the hybrid operating room. PhD Dissertation, University of Strasbourg, France.
Different imaging protocols were performed using the same tube tension and filtration values (100 kVp and 0.4 mm Al respectively), while varying the C-arm projection angles in the two rotation planes i.e. LAO/RAO and CAUD/CRAN. The same experimental conditions were simulated with our GPU-accelerated approach and the results were corrected using the measurements from two calibration dosimeters. The relative errors between the simulated dose values and the ones measured by the validation dosimeters were computed to verify the accuracy of the simulated dose distribution. A total of 15 C-arm configurations were considered. A simulation with 108 particles (uncertainty below 2%) was performed for each of them. The energy values deposited at the dosimeters' positions could be simulated in approximately 30 s with our GPU-accelerated approach. The errors calculated using the measurements from the 6 remaining dosimeters are reported in the table below. The mean error obtained for all experiments was of 18.7%. We consider this error acceptable since it is within the range of possible measurement inaccuracies from the type of dosimeters we use. As reported in , the intrinsic response error from active personal dosimeters can go up to 30%. Moreover, approximations performed in the X-ray source simulation model may introduce further errors to the results. Hence, the simulation approach presents promising performances in terms of simulation time while remaining accurate when compared to real dose measurements.
|X-ray device angulation||Simulation Error (%)|
Table 1. Mean simulation errors per evaluated X-ray imaging protocol.
To qualitatively assess the performance of our system, we also provide below a video demonstrating the use of our radiation awareness system in an OR at IHU Strasbourg. The video shows live recordings illustrating the different visualizations modes which are provided to improve radiation safety.
Thematically related posters
ECR 2018 / C-0511
Comparative analysis of effective risk for fixed tube current (FTC) and automatic tube current modulation (ACTM) during abdominal CT imaging
ECR 2018 / C-0144
Methodology for remote and automated QC in planar imaging: guidelines of an IAEA working group