Rapid magnetic resonance quantification on the brain: Optimization for clinical usage

Abstract

A method is presented for rapid simultaneous quantification of the longitudinal T 1 relaxation, the transverse T 2 relaxation, the proton density (PD), and the amplitude of the local radio frequency B 1 field. All four parameters are measured in one single scan by means of a multislice, multiecho, and multidelay acquisition. It is based on a previously reported method, which was substantially improved for routine clinical usage. The improvements comprise of the use of a multislice spin-echo technique, a background phase correction, and a spin system simulation to compensate for the slice-selective RF pulse profile effects. The aim of the optimization was to achieve the optimal result for the quantification of magnetic resonance parameters within a clinically acceptable time. One benchmark was high-resolution coverage of the brain within 5 min. In this scan time the measured intersubject standard deviation (SD) in a group of volunteers was 2% to 8%, depending on the tissue (voxel size ‫؍‬ 0.8 ؋ 0.8 ؋ 5 mm). As an example, the method was applied to a patient with multiple sclerosis in whom the diseased tissue could clearly be distinguished from healthy reference values. Additionally it was shown that, using the approach of synthetic MRI, both accurate conventional contrast images as well as quantification maps can be generated based on the same scan. Tissues in the human body can be distinguished with magnetic resonance imaging (MRI) depending on their MR parameters, such as the longitudinal T 1 relaxation, the transverse T 2 relaxation, and the proton density (PD). In clinical routine, the MR scanner settings, such as echo time (T E ), repetition time (T R ), and flip angle (␣), are most often chosen to highlight, or saturate, the image intensity of tissues, resulting in T 1 -weighting or T 2 -weighting in a contrast image. These procedures are well-established and relatively quick. A major disadvantage of using such contrast images is that the absolute intensity has no direct meaning and diagnosis relies on comparison with surrounding tissues in the image. In many cases it is therefore necessary to perform several different contrast scans. A more direct approach is the absolute quantification of the tissue parameters T 1 , T 2 , and PD. In this case, pathology can be examined on a pixel basis to establish the absolute deviation compared to the normal values. Automatic segmentation of such tissue images would be straightforward and the progress of the disease could then be expressed in absolute numbers. An excellent overview of the use of absolute quantification on neurodegenerative diseases is provided in Ref. 1.Although the advantages of absolute quantification are obvious, its clinical use is still limited. At least two major hurdles need to be addressed to stimulate widespread clinical usage. For many methods, the excessive scan time associated with the measurement of the three parameters has so far prohibited its clinical application. However, in recent years there has been substantial progress (see, e.g., R...

Compressed sensing reconstruction for magnetic resonance parameter mapping

“…An overcomplete dictionary that sparsely represents the training data can be designed by the K-SVD method proposed by Aharon et al (15). The K-SVD algorithm works iteratively, applying two steps in each iteration: (1) in the sparse coding step, the dictionary D is fixed, and a sparse representation with respect to that dictionary is obtained; (2) in the dictionary update step, the dictionary columns are updated, one column at a time to minimize the approximation error of the training data. The learned dictionary is optimized for a signal approximation with at most K atoms.…”

“…Direct quantification of the local MR parameters is of interest in a wide range of clinical applications including oncology and neurology (2,3), because it often provides more accurate and reproducible diagnostic information.…”

Synthetic MRI for Clinical Neuroimaging: Results of the Magnetic Resonance Image Compilation (MAGiC) Prospective, Multicenter, Multireader Trial

“…[1][2][3] The speed of diagnostic brain studies can thus be reduced to only about 5 minutes with synthetic MR imaging. 4 This advancement may help improve throughput and reduce rescanning, while also providing quantitative information of research interest. [4][5][6] Clinical studies of synthetic MR imaging are highly heterogeneous in that they examine a variety of conditions with widely varying scan parameters, with a paucity of large, randomized trials to inform clinical usage.…”

“…4 This advancement may help improve throughput and reduce rescanning, while also providing quantitative information of research interest. [4][5][6] Clinical studies of synthetic MR imaging are highly heterogeneous in that they examine a variety of conditions with widely varying scan parameters, with a paucity of large, randomized trials to inform clinical usage. 3,6 Blystad et al 5 (2012) reported that synthetic images had diagnostic utility similar to that of conventional imaging series, though with some quality issues like granulation and contrast particularly apparent in FLAIR views.…”

Synthetic quantitative MRI through relaxometry modelling

“…by generating multiple inversion recovery images covering a range of inversion times. Such a synthetic approach has been proposed as a first step towards the adoption of fully quantitative imaging within a clinical environment,7 and clinical utility has been demonstrated, e.g. in the visualization of tumours 8…”