Compensating for distortion caused by marker susceptibility in MR images

Ning Xu; Duane A. Yoder; J. Michael Fitzpatrick; Cynthia B. Paschal

doi:10.1117/12.537307

Translator Disclaimer

12 May 2004 Compensating for distortion caused by marker susceptibility in MR images

Ning Xu, Duane A. Yoder, J. Michael Fitzpatrick, Cynthia B. Paschal

Author Affiliations +

Proceedings Volume 5370, Medical Imaging 2004: Image Processing; (2004) https://doi.org/10.1117/12.537307
Event: Medical Imaging 2004, 2004, San Diego, California, United States

Abstract

Fiducial markers are used in image-guided surgery to register images to physical space. Submillimetric accuracy is achievable with CT, but with MR geometrical distortions may cause substantial error. Some anatomical regions may suffer minimal distortion, and the markers can be placed in areas of low distortion, but the marker's own magnetic susceptibility causes distortions of its shape and centroid, compromising the accuracy of its localization. General methods for correcting MR distortion require a second image acquisition. We show that it is possible to provide an accurate localization of marker position without a second image. Our method is to perform a simulation of MR imaging based on the known shape and contents of the marker and the known parameters of the imaging protocol. We compare simulated images at multiple candidate angles and positions to the acquired image. The centroid associated with the most similar image is the improved localization. We use a second simulator to provide ground truth, a binary marker model, and a 1-mm resolution for the candidate positions. For three orientations, the method recovered the correct centroid for signal-to-noise ratios as low as 10. For ratios of 5 and 7, we found an improvement in localization accuracy of 1.0±0.4 mm.

Citation Download Citation

Ning Xu, Duane A. Yoder, J. Michael Fitzpatrick, and Cynthia B. Paschal "Compensating for distortion caused by marker susceptibility in MR images", Proc. SPIE 5370, Medical Imaging 2004: Image Processing, (12 May 2004); https://doi.org/10.1117/12.537307

ACCESS THE FULL ARTICLE