Harald F. Busse PhD, Presenter: Nothing to Disclose

Nikita Garnov, Abstract Co-Author: Nothing to Disclose

Gregor Thörmer, Abstract Co-Author: Nothing to Disclose

Thomas Kurt Kahn MD, Abstract Co-Author: Nothing to Disclose

Michael Moche MD, Abstract Co-Author: Nothing to Disclose

Stereotactic MRI-guided needle interventions require an accurate patient-to-image registration, for which MR-visible markers are often used. This work presents a technique for simultaneous 3D localization of several custom-made markers and determines the impact of the number of markers on the targeting accuracy of a clinically used navigation system.

In the clinical setup, the surgical instrument is tracked in real time (optically) and navigated using continuously reformatted views of roadmap MRI data (Localite, St. Augustin, Germany). Outside the scanner (1.5 T, 60 cm bore size, Symphony, Siemens). Registration is established by an adjustable reference board with miniature RF coils. These wireless markers are imaged with a balanced SSFP sequence in three orthogonal projections (pixel resolution 0.6 mm, total acquisition time <11 s) and simultaneously localized by matching the independently determined peak coordinates of the 2D images in 3D. The accuracy was assessed by targeting a Gd-DOTA-filled, 1-mm wide capillary in a water bath. Euclidean distances between adjusted needle tip and actual target positions were measured for two marker numbers m (3 and 5) in a total of 20 trials.

Mean needle tip displacements in 3D were 5.9±2.5 mm (m=3) and 3.0±1.5 mm (m=5) (p<0.001, Fig. 5). Sensitivities (positive predictive values) for marker detection in the individual 2D views were 97.8% (89.8%) for m=3 and 96.0% (96.0%) for m=5, respectively. Despite missing (false negative) and false positive peaks in these views, the clinically relevant 3D localization was successful in 20/20 trials due to the effective double sampling of 2D peak coordinates. On average, image analysis of five markers took 50 ms longer (430 vs. 380 ms, +13%, p<0.001, Intel Core2 Duo CPU @ 2.66 GHz) than that of three ones. Complete patient registration (acquisition, transfer and automatic analysis of the marker images) took less than 30 s.

Simultaneous localization of five wireless MR markers could be successfully used for fast, accurate and robust patient registration of a clinical MRI navigation solution. The experimental targeting accuracy was significantly improved over that involving three markers only.

Accurate and easy navigational assistance in a diagnostic MRI environment can be provided by an optimized patient registration based on the simultaneous 3D localization of five wireless MR markers.

Busse, H, Garnov, N, Thörmer, G, Kahn, T, Moche, M, Accurate, Rapid and Automatic Patient Registration of a Clinical MRI Navigation System by Simultaneous 3D Localization of Five Wireless MR Markers.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12028373.html