A Unified Maximum Likelihood Framework for Simultaneous Motion and T Estimation in Quantitative MR T Mapping

In quantitative MR <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> mapping, the spin-lattice relaxation time <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> of tissues is estimated from a series of <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula>-weighted images. As the <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> estimation is a voxel-wise estimation procedure, correct spatial alignment of the <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula>-weighted images is crucial. Conventionally, the <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula>-weighted images are first registered based on a general-purpose registration metric, after which the <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> map is estimated. However, as demonstrated in this paper, such a two-step approach leads to a bias in the final <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> map. In our work, instead of considering motion correction as a preprocessing step, we recover the motion-free <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> map using a unified estimation approach. In particular, we propose a unified framework where the motion parameters and the <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> map are simultaneously estimated with a Maximum Likelihood (ML) estimator. With our framework, the relaxation model, the motion model as well as the data statistics are jointly incorporated to provide substantially more accurate motion and <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> parameter estimates. Experiments with realistic Monte Carlo simulations show that the proposed unified ML framework outperforms the conventional two-step approach as well as state-of-the-art model-based approaches, in terms of both motion and <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula> map accuracy and mean-square error. Furthermore, the proposed method was additionally validated in a controlled experiment with real <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula>-weighted data and with two in vivo human brain <inline-formula> <tex-math notation="LaTeX">T_{1} </tex-math></inline-formula>-weighted data sets, showing its applicability in real-life scenarios..

Medienart:	Artikel

Erscheinungsjahr:	2017
Erschienen:	2017

Enthalten in:	Zur Gesamtaufnahme - volume:36
Enthalten in:	IEEE transactions on medical imaging - 36(2017), 2, Seite 433-446

Sprache:	Englisch

Beteiligte Personen:	Ramos-Llorden, Gabriel [VerfasserIn] den Dekker, Arnold J [Sonstige Person] Van Steenkiste, Gwendolyn [Sonstige Person] Jeurissen, Ben [Sonstige Person] Vanhevel, Floris [Sonstige Person] Van Audekerke, Johan [Sonstige Person] Verhoye, Marleen [Sonstige Person] Sijbers, Jan [Sonstige Person]

Links:	Volltext ieeexplore.ieee.org

BKL:	44.09
Themen:	T ₁ mapping Brain modeling Data models Dynamic MRI Magnetic resonance imaging Mathematical model Maximum likelihood Maximum likelihood estimation Motion correction Registration Three-dimensional displays

RVK:	RVK Klassifikation XA 48667

doi:	10.1109/TMI.2016.2611653

funding:
Förderinstitution / Projekttitel:

PPN (Katalog-ID):	OLC1990934595