Meeting Banner
Abstract #1558

Kalman Filtering for Undersampeled Continuous Volumetric MR-Temperature Imaging

Baudouin Denis de Senneville1, 2, Sbastien Roujol, 23, Silke Hey2, 4, Chrit Moonen1, 2, Mario Ries1, 2

1Imaging Division, UMC Utrecht, Utrecht, Netherlands; 2CNRS / University of Bordeaux 2, IMF, Bordeaux, France; 3Cardiovascular Division, Beth Israel Deaconess, Medical Center, Harvard Medical School, Boston, United States; 4Philips Healthcare, Best, Netherlands

Volumetric real time MR-thermometry on moving organs for therapy guidance is challenging. As a result, recent efforts focused on the possibility to exploit the physical knowledge of the heating process for the artefact free reconstruction of 3D MR-temperature maps from under-sampled MR-data. This study investigates the possibility to reconstruct continuous volumetric temperature data from spatio-temporally under-sampled 3D MR-temperature maps using an Extended Kalman Filter with the bio-heat transfer equation as the model predictor. The proposed method was characterized with in-vivo HIFU experiments on porcine kidney.

Keywords

ablation accuracy achievable achieve acoustic adapts adjusting affected allows applications applied approaches available basic basis best beth cardiovascular coefficient compatible compromise computation continuous continuously corrected covariance coverage currently curve depicting depicts deposition described desired determined division duration dynamic dynamically efforts ellipsoid employed employs energy entire equation evaluated evaluating evolution evolved exactly exploit extended field filter filtered filtering focal free full gain generally heat heating held in vivo includes interventions introduce iteratively journal kidney knowledge known limitations limited local located loss maps measured medical model models newly noise note observation observe often originally parallel physiology placed porcine position power practical practice precision predictor preferably press pressure prevents priory process processes processing proposed real recent reconstructed reconstructing reconstruction reconstructs reduced reduces reduction relies removal resolution role sampled school shot since slice slices sliding space spatial spatially static stronger suboptimal system systematic target temperature temporal temporally thermometry transcribed transfer update variance volume volumetric watts width window