z-logo
Premium
Liver 4DMRI: A retrospective image‐based sorting method
Author(s) -
Paganelli Chiara,
Summers Paul,
Bellomi Massimo,
Baroni Guido,
Riboldi Marco
Publication year - 2015
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4927252
Subject(s) - image quality , computer science , artificial intelligence , medical imaging , image registration , computer vision , iterative reconstruction , interquartile range , sorting , gold standard (test) , magnetic resonance imaging , sorting algorithm , nuclear medicine , image (mathematics) , medicine , radiology , mathematics , algorithm , statistics
Purpose: Four‐dimensional magnetic resonance imaging (4DMRI) is an emerging technique in radiotherapy treatment planning for organ motion quantification. In this paper, the authors present a novel 4DMRI retrospective image‐based sorting method, providing reduced motion artifacts than using a standard monodimensional external respiratory surrogate. Methods: Serial interleaved 2D multislice MRI data were acquired from 24 liver cases (6 volunteers + 18 patients) to test the proposed 4DMRI sorting. Image similarity based on mutual information was applied to automatically identify a stable reference phase and sort the image sequence retrospectively, without the use of additional image or surrogate data to describe breathing motion. Results: The image‐based 4DMRI provided a smoother liver profile than that obtained from standard resorting based on an external surrogate. Reduced motion artifacts were observed in image‐based 4DMRI datasets with a fitting error of the liver profile measuring 1.2 ± 0.9 mm (median ± interquartile range) vs 2.1 ± 1.7 mm of the standard method. Conclusions: The authors present a novel methodology to derive a patient‐specific 4DMRI model to describe organ motion due to breathing, with improved image quality in 4D reconstruction.

This content is not available in your region!

Continue researching here.

Having issues? You can contact us here