A real-time radiation dose monitoring system for patients and staff during interventional fluoroscopy using a GPU-accelerated Monte Carlo simulator and an automatic 3D localization system based on a depth camera

Abstract

Radiation monitoring systems able to accurately track the radiation dose received by the patient and the medical staff during interventional fluoroscopy can be used to minimize the likelihood and severity of radiation-induced skin injuries and estimate the accumulated organ doses. We describe a method to monitor doses in real time using automatic sensors in the imaging room and a GPU-accelerated computer simulator. The Monte Carlo simulation code MC-GPU is used to estimate patient and staff doses due to primary and scattered radiation, along with the associated statistical uncertainties. The geometrical configuration of the irradiation is automatically determined and updated using data from a depth camera that tracks the location and posture of each person in the imaging room. A virtual x-ray source graphical interface is used to manually trigger the simulations. The implemented computational framework separates the simulation of the x-ray transport through the patient and the operator bodies into two coupled, sequential simulations. The initial simulation uses the patient anatomy and a c-arm source model with a collimated cone beam emitted from a point focal spot. During this simulation a large phase space file with the energy, position and direction of x rays scattered in the direction of the operator is created. The phase space file is then used as the input radiation source for the following simulation with the operator anatomy model. Particle recycling is employed as a variance reduction technique to maximize the information obtained from the limited number of particles scattered towards the operator. For a typical image acquisition, a patient skin dose map can be displayed at the operator's monitor within 10 seconds with a peak skin dose error below 1%. This work demonstrates that a dose monitoring system based on accurate Monte Carlo simulations can be used to estimate in real-time the average and peak organ doses for both the patient and the staff in interventional fluoroscopy, and provide timely information regarding possible overdoses while the imaging procedure is being performed.

Tecnologias em pesquisa: ciências exatas e biológicas

“…4.1 Sistema de controle de dose de radiação usando simulador Monte Carlo acelerado por GPU Badal et al (2013) mostram que sistemas de monitorização de radiação capazes de controlar com precisão a dose de radiação recebida pelo paciente e a equipe médica durante a fluoroscopia intervencionista podem ser utilizados para minimizar a probabilidade e gravidade das lesões de pele e estimar as doses nos órgãos acumuladas por radiação induzida. Além disso, descrevem um método para monitorizar em tempo real as doses usando sensores automáticos na sala de imagem e um simulador de computador acelerado por GPU.…”

“…Badal et al (2013) demonstram que um sistema de monitoramento de dose com base em simulações Monte Carlo pode ser utilizado para estimar em tempo real as médias e picos de doses em ór-gãos, tanto para o paciente quanto para a equipe em fluoroscopia intervencionista, e fornecer informações oportunas sobre possíveis overdoses enquanto o pro-cedimento de imagem está sendo realizado. A Figura 1 apresenta um fluxograma detalhado descrevendo os diferentes elementos da estrutura do software utilizado.…”

“…Quando se trata de simulação computacional, nota-se que esta técnica vem sendo utilizada como auxílio ao desenvolvimento tecnológico no monitoramento de pacientes, como pode ser visto em Badal et al (2013), Chunming e Xuemei (2010), , Sirait et al (2013) e Sloane e Gelhot (2004). A simulação computacional auxilia o processo decisório, tornando possível optar pela melhor decisão após várias experimentações, o que é fundamental quando se quer evitar falhas e permitir assim o melhor bem-estar dos pacientes.…”

See It With Your Own Eyes: Markerless Mobile Augmented Reality for Radiation Awareness in the Hybrid Room

“…For the system to achieve clinical use, the displayed radiation maps should be computed in near realtime while considering the current room configuration. This is feasible by using GPU-accelerated simulations as in [19].…”

“…To this end, computer-based systems combining radiation simulations, person tracking and visualization to show the 3D dose distribution in the OR have been proposed (e.g. [11], [12], [19]). Such information is displayed in an external view by overlaying the computed 3D radiation map over a model of the current OR layout or over the clinicians' tracked shapes.…”