Method for Rapid MRI Needle Tracking Eyal Kochavi, 1 Dorith Goldsher, 2 and Haim Azhari 1 * A new method for MRI needle tracking within a given two-dimen- sional (2D) image slice is presented. The method is based on k-space investigation of the difference image between the current dynamic frame and a reference frame. Using only a few central k-lines of the difference image and a nonlinear optimization pro- cedure, one can resolve the parameters that define the 2D sinc function that best characterizes the needle in k-space. The spatial location and orientation of the needle are determined from these parameters. Rapid needle tracking is obtained by repeated acqui- sitions of the same set of several central k-lines (as in a “keyhole” protocol) and repeated computation of these pa- rameters. The calculated needle tip is depicted on the refer- ence image by means of a graphic overlay. The procedure was tested in computer simulations and in actual MRI scans (the computations were done offline). It was demonstrated that six k-lines out of 128 usually suffice to locate the needle. The refresh rate of the needle location depends on the time required to sample the subset of k-lines, calculate the cur- rent needle location, and refresh the reference image. Magn Reson Med 51:1083–1087, 2004. © 2004 Wiley-Liss, Inc. Key words: interventional MRI; iMRI; rapid imaging; needle tracking; invasive MRI Interventional MRI has recently become available for min- imally invasive body, spinal, and intracranial clinical pro- cedures (1). Such procedures usually involve the insertion of a needle into a target organ. Needle insertion may be used for biopsy (2), fine-needle aspiration (FNA) (3), and thermal treatment or ablation (4,5). It is essential to mon- itor the needle in real time in such cases, and thus rapid imaging techniques are implemented. However, with most rapid imaging techniques, image quality is compromised. Thus, the use of special devices or equipment to increase the needle/tissue contrast or to enable direct needle local- ization has been suggested (6). For example, in one ap- proach (7) a special optical tracking system is used to monitor the spatial coordinates of three light-emitting di- odes located on a special needle holder. From these coor- dinates, and with the assumption that the needle is rigid and perpendicular to the holder, its spatial orientation can be calculated and displayed as a graphic overlay on a reference image. Another approach involves the use of a special miniature coil attached to the needle tip: one cal- culates the location of the needle by measuring the mag- netic field the coil reads from its surroundings (8,9). These techniques add complications to already complex procedures, and therefore direct tracking using only the MRI information can be advantageous. The objective of this study was to develop a method for rapid needle track- ing that does not require the use of complicated tracking equipment. MATERIALS AND METHODS Basic Assumptions The suggested method is based on the following assump- tions: 1) The needle is characterized by a low MR signal and appears as a dark object in the image. 2) The needle appears in the MR image as a long, thin rectangle. 3) The background is quasi-static (i.e., for a short period of time, the only dynamic change noted is the motion of the nee- dle). 4) The width of the needle image is known. 5) The algorithm is activated after the first reference image has been acquired and the needle insertion has started. Fol- lowing these assumptions, subtracting the dynamic image from a reference static image will produce a difference image (DI), which contains mainly the needle information and resembles a rectangle. The main idea behind the cur- rent method is that one can determine the needle location and orientation using partial sampling of the DI k-space (Fourier domain), and find the rectangle that best fits the DI in k-space. Since the time required to sample only a few lines in k-space is short, rapid MR needle tracking can be achieved. Characteristic Needle Parameters The location and orientation of the rectangle that best represents the needle image in two dimensions can be defined by five parameters: 1) a = half the length of the rectangle, 2) b = half the width of the rectangle, 3)  = the inclination angle, and 4) X o and 5) Y o are the coordinates of its center of mass. In the frequency domain (k-space), the rectangle is represented by the 2D sinc function de- scribed in Eq. [1] (up to a constant): Sinc 2 D(k x ,k y ) = sinc{a/(2)  [k x  cos - k y  sin ]}  sinc{b/(2)  [k x  sin + k y  cos ]}  exp{ -j  [k x  X 0 + k y  Y 0 ]} [1] By solving the inverse problem for Eq. [1], i.e., extracting the five needle parameters (a, b, , X o , and Y o ) from the Sinc2D function, the needle location in the image domain can be fully determined. In addition, note that the phase part of the Sinc2D function depends exclusively on the shift parameters (X o and Y o ), while the amplitude part (up to a sign) is determined exclusively by the other three parameters. 1 Department of Biomedical Engineering, Technion, Israel Institute of Technol- ogy, Haifa, Israel. 2 MRI Institute, Rambam Medical Center and Faculty of Medicine, Technion, Israel Institute of Technology, Haifa, Israel. Grant sponsor: Technion V.P.R. *Correspondence to: Dr. Haim Azhari, Dept. of Biomedical Engineering, Tech- nion IIT, Haifa 32000, Israel. E-mail: Haim@Bm.Technion.Ac.Il Received 19 February 2003; revised 24 December 2003; accepted 29 De- cember 2003. DOI 10.1002/mrm.20065 Published online in Wiley InterScience (www.interscience.wiley.com). Magnetic Resonance in Medicine 51:1083–1087 (2004) © 2004 Wiley-Liss, Inc. 1083