Introduction

A standard MR imaging protocol as it is currently used in clinical practice relies on the acquisition of multiple images with different levels of T₁- and T₂-weighting. Signal amplitudes in these images are in arbitrary units and tissue types (e.g. gray and white matter) can only be distinguished by relative signal intensity variations within the imag­e. In contrast, quantification of the underlying parameters that determine the observed signal amplitudes puts numbers in physically meaningful units onto the MR images. This is especially helpful in inter-subject and inter-site comparison studies, where the conventional MRI contrasts with their arbitrary signal scale provide a challenge. In longitudinal studies, quantification allows the assessment of disease progression and therapy. Furthermore, the quantitative imaging approach allows for easier tissue segmentation and classification.

Basic quantification approach

The MR signal depends on many different physical tissue properties, such as proton density (PD), the longitudinal (T₁) and transverse (T₂, T₂^*) relaxation times, diffusion, temperature, susceptibility, and many other effects:
$$S(PD, T_1, T_2, T_2^*, D, T, \mu, v, MT, ...)$$
How much each of these properties contributes to the MR signal (its weighting) largely depends on the experimental design, i.e. the MR imaging sequence and its acquisition parameters (TR, TE, TI, FA, b, …). Thus, for quantification it is necessary to design an experiment that shows a high sensitivity to the tissue property of interest. Furthermore, this sensitivity needs to be adjustable, e.g. by variation of one or more acquisition parameters. Quantification can be achieved by acquiring multiple images that each show a different signal contribution (weighting) from the tissue property and by then fitting a known signal model to these multiple data points.

T₁ relaxometry

All imaging sequences are sensitive to T₁ in one way or another and there are many ways how to exploit this sensitivity for quantification.

Steady-state techniques
Steady-state techniques usually rely on rapid imaging with gradient-echo (or gradient-recalled-echo, GRE) acquisitions. The name implies that both longitudinal and transverse magnetization are kept constant, i.e. there is no (or only a negligible) signal evolution during the scan. To ensure the steady-state condition prior to scanning, it is usually necessary to apply a “dummy pulse” period at the start of the sequence. There are many different steady-state techniques that show a different mixture of T₁, T₂, T₂^*, and other weightings (e.g. diffusion). For T₁ mapping, the rf-spoiled GRE sequence is very popular since it shows almost pure T₁-weighting (plus T₂^*-weighting which is constant for constant TE) and is not affected (too much) by T₂. The ideal signal equation for an rf-spoiled GRE is given by:
$$S_{\textit{rf-spoiled GRE}} = M_0 \sin{\alpha} \frac{1-\exp(-TR/T_1)}{1-\exp(-TR/T_1) \cos{\alpha}}\exp(-TE/T_2^*)$$
As we can see, T₁-weighting only depends on the flip angle α and the repetition time T­R. Thus, we can acquire multiple images with variable TR and/or variable flip angle and then fit the image intensity in each pixel to this equation in order to obtain T₁. The variable flip angle technique (VFA; Fram et al.) is by far the most popular option here.

Saturation Recovery
An alternative approach relies on magnetization preparation prior to imaging. The preparation module usually consists of an inversion or saturation pulse (inversion-recovery (IR) / saturation recovery (SR)). The exponential recovery towards the equillibrium magnetization is then sampled at multiple time points TI (“inversion time”) by applying an excitation pulse α. T₁ can then be obtained from a fit to the corresponding signal equation:
$$S_{IR} = M_0 \sin{\alpha} (1 - 2\exp(-TI/T_1))$$
$$S_{SR} = M_0 \sin{\alpha} (1 - \exp(-TI/T_1))$$
As we can see, T₁-weighting only depends on the flip angle α and the repetition time T­R. Thus, we can acquire multiple images with variable TR and/or variable flip angle and then fit the image intensity in each pixel to this equation in order to obtain T₁. The variable flip angle technique (VFA; Fram et al.) is by far the most popular option here.

T2 relaxometry

T₂ relaxometry is most often performed using either a spin-echo sequence with varying T­E or a multi-spin echo sequence (often using the CPMG phase-cycling technique, see Carr and Purcell). Alternatively, it is also possible to prepare the magnetization prior to imaging using a T₂ preparation module that provides variable T₂-weighting (Wong et al.).

Acknowledgements

No acknowledgement found.