TY - JOUR
T1 - Application of XMR 2D-3D registration to cardiac interventional guidance
AU - Rhode, Kawal S.
AU - Hill, Derek L.G.
AU - Edwards, Philip J.
AU - Hipwell, John
AU - Rueckert, Daniel
AU - Sanchez-Ortiz, Gerardo I.
AU - Hegde, Sanjeet
AU - Rahunathan, Vithuran
AU - Razavi, Reza
PY - 2003
Y1 - 2003
N2 - We describe a registration and tracking technique to integrate cardiac x-ray images and cardiac magnetic resonance (MR) images acquired from a combined x-ray and MR interventional suite (XMR). Optical tracking is used to determine the transformation matrices relating MR and x-ray image coordinates. Calibration and tracking enable us to combine x-ray projection images with registered projection MR images from a volume acquisition, and to display 3D reconstructions of catheters within the MR volume. Registration errors were assessed using phantom experiments. Errors in the combined projection images (2D root mean square target registration error - RMS TRE) were found to be 2.4 - 4.2mm, and the errors in the integrated volume representation (3D RMS TRE) were found to be 4.6 - 5.1mm. These errors are clinically acceptable for alignment of images of the heart and the great vessels. We demonstrate that our technique can be used to guide catheters and to combine electrical and mechanical information in the context of the patient-specific anatomy for a patient undergoing cardiac radio-frequency (RF) ablation.
AB - We describe a registration and tracking technique to integrate cardiac x-ray images and cardiac magnetic resonance (MR) images acquired from a combined x-ray and MR interventional suite (XMR). Optical tracking is used to determine the transformation matrices relating MR and x-ray image coordinates. Calibration and tracking enable us to combine x-ray projection images with registered projection MR images from a volume acquisition, and to display 3D reconstructions of catheters within the MR volume. Registration errors were assessed using phantom experiments. Errors in the combined projection images (2D root mean square target registration error - RMS TRE) were found to be 2.4 - 4.2mm, and the errors in the integrated volume representation (3D RMS TRE) were found to be 4.6 - 5.1mm. These errors are clinically acceptable for alignment of images of the heart and the great vessels. We demonstrate that our technique can be used to guide catheters and to combine electrical and mechanical information in the context of the patient-specific anatomy for a patient undergoing cardiac radio-frequency (RF) ablation.
UR - http://www.scopus.com/inward/record.url?scp=0345255044&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-39899-8_37
DO - 10.1007/978-3-540-39899-8_37
M3 - Article
AN - SCOPUS:0345255044
SN - 0302-9743
VL - 2878
SP - 295
EP - 302
JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ER -