A Technical Overview of Quantitative Imaging
Martijn A. Cloos1
1Centre for Advanced Imaging, University of Queensland, St. Lucia, Australia

Synopsis

This talk reviews key principles of quantitative MRI. We will focus on methods to estimate longitudinal relaxation (T1), transverse relaxation (T2), and transmit field variations (B1+) in the brain. We will also briefly touch on other quantifiable properties, such as magnetisation transfer (MT), and discuss some of the limitations associated with model-based methods to quantitative properties of complex organic samples using MRI. Finally, we will briefly reflect on the trade-off between accuracy and precision, and its possible implications for future work.

Most MRI images show a relative contrast between tissues that depends on the specific sequence parameters and scanner hardware used to acquire the data. Quantitative MRI, on the other hand, strives to directly measure objective physical quantities, such as the longitudinal (T1) and transverse (T2) relaxation times. Compared to qualitative images, such quantitative maps enable more straightforward comparisons between scans from different patients, different time points, and ideally, also when using different hardware. [1,2]

Over the years many techniques have been developed to estimate these quantitative parameters using MRI. Historically, relaxometry measurements were performed by fitting exponentials to a series of inversion times (to measure T1) or spin-echo times (to measure T2) [3]. Unfortunately, such measurements are extremely slow. More modern techniques, such as Look-Locker [4] or DESPOT1 and DESPOT2 [5], can be much faster. However, such methods often compound experimental imperfections [5], rely on simplified analytic models [5, 6, 7, among others], or inadvertently mix contrast mechanisms [4,5,8,9,10]. Therefore, fast mapping techniques often compromise accuracy and/or precision.

Not long ago, the simultaneous quantification of multiple parameters was proposed to better balance the trade-off between speed and accuracy [11,12,13]. Most notably, MR Fingerprinting (MRF) [12] which uses a variable flip angle train that produces unique signal evolutions, so-called “fingerprints”, for different tissue parameter combinations. By comparing the measured fingerprints with the entries in a precomputed dictionary, the underlying tissue parameters, such as T1 and T2, can be estimated.


Although MRF side-steps the fitting of (often oversimplified) analytic equations, it is still limited by the accuracy of the model used to generate the dictionary and the encoding power of the sequence itself. For example, most MRF sequences also inadvertently encode B1+ and other experimental imperfections such as the slice-profile. The slice profile is often straightforward to incorporate into the simulation [14,15]. However, B1+ can be trickier. First, it adds an additional dimension to the fitting process. Moreover, without dedicated sequence optimisation, B1+, T1, and T2 can be difficult to decouple [11, 12,16,17]. Alternatively, one can also use separate B1+ measurements to correct the T1 and T2 values [15]. However, care should be taken, imperfections in these B1+ could also bias the final T1 and T2 results [14].

There are many B1+ mapping strategies [18,19,20,21,22,23]. Much like T1 and T2 mapping, faster B1+mapping strategies often trade accuracy for speed. Interestingly, here increased speed often comes at the cost of mixed in T1 effects [18]. For example, the turbo-flash based B1+ mapping strategies [19] typically assume that T1 effects can be neglected during the readout, which is not true for tissues like fat. Therefore, one could argue that it is preferred to jointly estimate tissue properties and experimental imperfections [14].

Finally, we must remember that even our best models have limitations. The microstructural complexity of organic tissue is difficult to captured using basic Bloch equations [24]. For example, motion [25], magnetisation transfer [26], diffusion [27] , and other effects, also play an important role. When these are incorporated into the model, estimation of T1 and T2 tend to change [9,28]. This begs the question: can we ever really get “True” T1 or T2 measurements for organic tissues? With this question in mind, it is tempting to emphasize precision, over accuracy. Perhaps, low accuracy and high precision measurements can already help characterise pathologies and stage disease progression? On the other hand, if such methods become popular, it may become difficult to displace them with new fundamentally better techniques, as their “quantitative” values may not be interchangeable.

Acknowledgements

No acknowledgement found.

References

[1] Margaret Cheng, Hai‐Ling, et al. "Practical medical applications of quantitative MR relaxometry." Journal of Magnetic Resonance Imaging 36.4 (2012): 805-824.

[2] Deoni, Sean CL. "Quantitative relaxometry of the brain." Topics in magnetic resonance imaging: TMRI 21.2 (2010): 101.

[3] Brown, Robert W., et al. Magnetic resonance imaging: physical principles and sequence design. John Wiley & Sons, 2014.

[4] Look, David C., and Donald R. Locker. "Time saving in measurement of NMR and EPR relaxation times." Review of Scientific Instruments 41.2 (1970): 250-251.

[5] Deoni, Sean CL, Brian K. Rutt, and Terry M. Peters. "Rapid combined T1 and T2 mapping using gradient recalled acquisition in the steady state." Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 49.3 (2003): 515-526.

[6] Marques, José P., et al. "MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field." Neuroimage 49.2 (2010): 1271-1281.

[7] Barral, Joëlle K., et al. "A robust methodology for in vivo T1 mapping." Magnetic resonance in medicine 64.4 (2010): 1057-1067.

[8] Malik, Shaihan J., Rui Pedro AG Teixeira, and Joseph V. Hajnal. "Extended phase graph formalism for systems with magnetization transfer and exchange." Magnetic resonance in medicine 80.2 (2018): 767-779.

[9] Hilbert, Tom, et al. "Magnetization transfer in magnetic resonance fingerprinting." Magnetic resonance in medicine84.1 (2020): 128-141.

[10] Gai, Neville D., and John A. Butman. "Fast T1 mapping determined using incomplete inversion recovery look–locker 3D balanced SSFP acquisition and a simple two‐parameter model fit." Journal of Magnetic Resonance Imaging 35.6 (2012): 1437-1444.

[11] Newbould, Rexford D., et al. "Three-dimensional T1, T2 and proton density mapping with inversion recovery balanced SSFP." Magnetic resonance imaging 28.9 (2010): 1374-1382.

[12] Ma, Dan, et al. "Magnetic resonance fingerprinting." Nature495.7440 (2013): 187-192.

[13] Ben‐Eliezer, Noam, Daniel K. Sodickson, and Kai Tobias Block. "Rapid and accurate T2 mapping from multi–spin‐echo data using Bloch‐simulation‐based reconstruction." Magnetic resonance in medicine 73.2 (2015): 809-817.

[14] Cloos, Martijn A., et al. "Multiparametric imaging with heterogeneous radiofrequency fields." Nature communications7.1 (2016): 1-10.

[15] Ma, Dan, et al. "Slice profile and B1 corrections in 2D magnetic resonance fingerprinting." Magnetic resonance in medicine 78.5 (2017): 1781-1789.

[16] Buonincontri, Guido, et al. "Spiral MR fingerprinting at 7 T with simultaneous B1 estimation." Magnetic resonance imaging 41 (2017): 1-6.

[17] Cloos, Martijn A., et al. "Rapid radial T1 and T2 mapping of the hip articular cartilage with magnetic resonance fingerprinting." Journal of Magnetic Resonance Imaging 50.3 (2019): 810-815.

[18] Pohmann, Rolf, and Klaus Scheffler. "A theoretical and experimental comparison of different techniques for B1 mapping at very high fields." NMR in Biomedicine 26.3 (2013): 265-275.

[19] Chung, Sohae, et al. "Rapid B1+ mapping using a preconditioning RF pulse with TurboFLASH readout." Magnetic resonance in medicine 64.2 (2010): 439-446.

[20] Nehrke, Kay, and Peter Börnert. "DREAM—a novel approach for robust, ultrafast, multislice B1 mapping." Magnetic resonance in medicine 68.5 (2012): 1517-1526.

[21] Sacolick, Laura I., et al. "B1 mapping by Bloch‐Siegert shift." Magnetic resonance in medicine 63.5 (2010): 1315-1322.

[22] Yarnykh, Vasily L. "Actual flip‐angle imaging in the pulsed steady state: a method for rapid three‐dimensional mapping of the transmitted radiofrequency field." Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 57.1 (2007): 192-200.

[23] Cunningham, Charles H., John M. Pauly, and Krishna S. Nayak. "Saturated double‐angle method for rapid B1+ mapping." Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine 55.6 (2006): 1326-1333.

[24] Bloch, Felix. "Nuclear induction." Physical review 70.7-8 (1946): 460.

[25] Yu, Zidan, et al. "Exploring the sensitivity of magnetic resonance fingerprinting to motion." Magnetic resonance imaging 54 (2018): 241-248.

[26] Pike, G. Bruce. "Pulsed magnetization transfer contrast in gradient echo imaging: a two‐pool analytic description of signal response." Magnetic resonance in medicine 36.1 (1996): 95-103.

[27] Torrey, Henry C. "Bloch equations with diffusion terms." Physical review 104.3 (1956): 563.

[28] Malik, Shaihan J., Rui Pedro AG Teixeira, and Joseph V. Hajnal. "Extended phase graph formalism for systems with magnetization transfer and exchange." Magnetic resonance in medicine 80.2 (2018): 767-779.

Proc. Intl. Soc. Mag. Reson. Med. 30 (2022)