IEEE TRANSACTIONS ON MAGNETICS, VOL. 48, NO. 2, FEBRUARY 2012 395 Magnetic Tracking Inside Conducting Bores for Radiotherapy Tumor Localization Systems Zubiao Xiong , Shi Feng , John E. McGary , and Ji Chen Department of Electrical and Computer Engineering, University of Houston, Houston, TX 77204 USA Baylor College of Medicine, Houston, TX 77030 USA Magnetic tracking is sensitive to the eddy currents induced on the conducting gantry during radiotherapy tumor localization. We propose a unique magnetic tracking method that can resist the eddy current distortion. Our method uses a specific sensor array to estimate the secondary magnetic field generated from the eddy currents. An optimization procedure is adopted to reduce the distortion effects. A technique of providing the initial guess is also presented to accelerate the convergence of the algorithm. Numerical results demonstrate the effectiveness of our method for tumor localization inside radiotherapy gantry bores. Index Terms—Eddy currents, magnetic tracking, optimization, tumor localization. I. INTRODUCTION M AGNETIC tracking is one of the main technologies for capturing tumor motion during radiation therapy [1], [2]. It has attracted wide attention, since it does not use ionizing radiation that may interfere with the therapy. In addition, it can also operate in real time. These algorithms use one transponder to generate a weak magnetic field that can be detected by an array of sensors. Based on the received signals by the array sen- sors, various algorithms can be used to track the location of the transponder. However, the accuracy of these tracking algorithms starts to degrade when applied to track tumors motion inside gantry bores [3]. This is caused by the secondary magnetic field gener- ated by the induced currents on the nearby conducting gantry. Such secondary magnetic field disturbs the original magnetic field generated by the transponder and causes tracking errors. Since most conventional magnetic tracking methods [4]–[6] do not consider the interferences from the eddy currents; the tracking error can be over 10 cm. Even with some calibration techniques [7]–[9] that measure a reference target at a set of predefined positions, it is difficult for them to adapt to environ- ment changes. So far, no published methods could handle the dynamic distortions without fusing other tracking devices. In this paper, a novel magnetic tracking method is proposed to minimize the tracking error caused by gantry eddy currents. Since tracking and correcting are processed simultaneously, dynamic changes of distortion could also be handled. Sensor arrangement and distortion minimization algorithm are dis- cussed. Numerical simulations are performed to evaluate tracking performance. II. SYSTEM DESCRIPTION The magnetic tracking system used here is composed of an energizer, a transponder, and an array of sensors. Fig. 1 shows Manuscript received July 05, 2011; revised October 18, 2011; accepted November 04, 2011. Date of current version January 25, 2012. Corresponding author: J. Chen (e-mail: jchen18@uh.edu). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TMAG.2011.2175439 Fig. 1. Magnetic tracking system used for tumor localization. a schematic diagram of the magnetic tracking system. The pa- tient lies on a sliding bed inside the CT bore during treatment. A wireless transponder is implanted into or near the patient’s tumor mass, serving as the target to be tracked. An external en- ergizer powers up the transponder by electromagnetic induction during the exciting stage. After energized, the transponder emits an oscillating magnetic field, which can then be detected by ex- ternal magnetic sensors. Using carefully designed tracking al- gorithms, the instantaneous position of the transponder may be obtained. Since the detection distance is much smaller than the wave- length of the operating frequency and much larger than the transponder’s dimension, the magnetic field generated by the transponder can be approximated by the magnetic dipole as (1) where and are the positions of the transponder and the sensor respectively. is the magnetic moment of the transponder, is the permeability in the free space, and defines the Euclidean norm. Solving the transponder position from the measurements of the magnetic field is a typical inverse problem. Because the number of unknowns is six (typically is also unknown), at least six sensors are required. To improve the tracking accuracy, the number of sensors is usually larger than six. The proposed sensor array consists of four groups of sensors. This configuration is different from traditional planar array [10] 0018-9464/$31.00 © 2012 IEEE