3D motion estimation for on-line MR temperature mapping

Magnetic resonance (MR) temperature mapping can be used to monitor temperature changes during minimally invasive thermal therapies during the procedure. Robust 3D estimation of organ displacement during the intervention is hardly feasible due to technical limitations (spatial and temporal resolution are not sufficient to perform classic 3D registration methods on anatomical images). However, organ displacements due to physiological activity (heart and respiration) may induce important artifacts on apparent temperature maps and prevent the treatment of a tumor with an external heating device. Recent development has allowed increasing external information for 3D motion estimation. This paper presents a new method that exploits image information for robust 3D motion estimation from MR images, in order to make possible the treatment of organs such as the kidney, and to improve the precision of temperature estimation using the proton resonance frequency (PRF) shift. The motion estimation described in this paper consists of two steps: a reference volume is initially computed in a preparation phase; during the intervention a 3D displacement vector is estimated on-line for each pixel of the acquired images by using this reference volume. This method can be applied to any type of image sequences and thus is not restricted to MR images.