Implementation of a low-cost reliable non-contact respiratory motion monitoring system in radiotherapy

Target motion during treatment must be minimized to optimize photon and proton therapy. For example during treatment of the upper torso, the target is known to move substantially due to patient respiration. This motion has also an effect on images acquired before as well as during treatment, causing a reduced precision of the planned dose. As a result, the delivered dose distribution deteriorates. To overcome these problems, respiratory monitoring systems are available to track and reduce motion errors. The visionRT body contour system has been proposed for this purpose. A low-cost alternative to these solutions is the Kinect surface sensor from Microsoft. It is a real-time, non-invasive surface sensing system that is available with open source software and its potential has been proven in a number of pre-clinical studies [1,2,3,4]

Recently, we compared the performance of the Kinect surface sensor with that of a series of cone-beam CT acquisitions in our clinic. For 11 patients we calibrated the Kinect system to a 4D- cone beam CT and analyzed its performance by matching the phase of the breathing cycle based on Kinect with that based on CT. A respiratory signal extraction algorithm was used to detect local surface motion. In this presentation I will discuss the results of this study and the performance of the Kinect sensor in our institute.

(1) Martinez et al. 2012 “Breath Rate Monitoring During Sleep using Near-IR Imagery and PCA”, ICPR 2012.

(2)  Xia et al. 2012 “A real-time respiratory motion monitoring system using KINECT: Proof of concept” Medical Physics Volume 39

(3) Alnowami et al. 2012 “A quantitative assessment of using the Kinect for Xbox360 for respiratory surface motion tracking” SPIE Medical Imaging 2012

(4) Wasza et al.” Sparse Principal Axes Statistical Surface Deformation Models for Respiration Analysis and Classification” book: Bildverarbeitung fur die Medizin 2012