4733

Inversion Recovery Ultrashort Echo Time Imaging of Short T2 Tissue Components in Ovine Brain: A Sequential D2O Exchange Study

Shu-Juan Fan¹, Yajun Ma¹, Graeme M. Bydder¹, and Jiang Du¹

¹Dept. of Radiology, University of California, San Diego, San Diego, CA, United States

Synopsis

Myelin produces ultrashort-lived MRI signals, and cannot be directly imaged using conventional MRI sequences that typically have TEs of several milliseconds or longer. This study explored the potential of inversion recovery ultrashort echo time (IR-UTE) sequences in direct myelin imaging in an ovine brain D₂O exchange model. The IR-UTE signals survived D₂O exchange. Myelin T2* was measured to be 200-300 µs both before and after exchange. These results support myelin to be the major source of the ultrashort T2* signals seen on IR-UTE images, and IR-UTE sequence as a tool for assessing myelin loss in multiple sclerosis and other diseases.

Purpose

Myelin produces extremely short-lived signals that cannot be detected using conventional MRI sequences. Inversion recovery ultrashort echo time (IR-UTE) imaging holds the potential to directly detect and measure these signals. This study aimed to explore this potential in ovine brain.

Methods

Sample preparation: Five unfixed ~3-mm thick ovine brain specimens containing cerebral hemisphere (n=4) and cerebellum (n=1) were examined before and after exchange with D₂O (99.9%, Sigma-Aldrich, St. Louis, MO). Each specimen was immersed in 10 mL D₂O, which was refreshed daily.

IR-UTE MRI: All specimens were imaged using a GE 3T MRI scanner with a 3-inch receive-only coil. Long T2 signals were measured using a clinical proton density-weighted fast spin echo (PD-FSE) sequence with TR/TE = 8000/13.5 ms. Ultrashort T2 signals were measured using IR-UTE sequences with a minimum TE of 10 µs (Figure 1A).¹ To determine the optimal inversion time (TI_null) for completely nulling white matter (WM) or grey matter (GM) long T2 signals (WM_L or GM_L, Figures 1B-C), IR-UTE images were first acquired with variable inversion times with TR/TE = 1000/2.2 ms to measure WM_L and GM_L T1s (Figure 2A). Then at WM_L or GM_L TI_null, images with TEs of 0.01-4 ms were acquired for measuring the T2*s of WM or GM short T2 tissue components (WM_S or GM_S).

Data analysis: WM_L and GM_L T1s and TI_nulls, and WM_S and GM_S T2*s were calculated in each user-defined ROI offline using single-component models in Matlab. Changes in PD-FSE signals, T1 and T2* values were plotted against D₂O exchange time.

Results

Figures 2-4 show representative results from one cerebral hemisphere specimen. The contrast between GM and WM changed significantly with inversion time (Figure 2B). Before exchange with D₂O, WM_L and GM_L TI_null values were ~300 ms and 370 ms, respectively (Figure 2D), and the WM_S T2* was 200-300 ms at WM_L TInull = 300 ms (Figure 3A, lower panel, right corner). Subtracting the magnitude image of TE_0.60ms from that of TE_10µs provided positive signal for WM and negative signal for GM (Figure 3B). After 24-hr to 7-day exchange, the specimen showed near-zero PD-FSE signal (Figure 4A&C), but its IR-UTE signal, especially that of the WM, was still higher than background noise (Figure 4B). WM_L and GM_L T1s first decreased then increased, resulting in different TI_null values. The WM_S T2* showed minimal change (Figure 4D), ranging from 197 to 248 µs in all studied cerebral hemisphere specimens (data not shown) at all time points.

Figure 5 shows results from the cerebellar specimen. At WM_L TI_null (260 ms), the subtraction image (TE_10ms–TE_0.6ms) provided positive signal in WM and negative signal in GM (Figure 5A). At GM_L TI_null (350 ms), the subtraction image (TE_10ms–TE_0.6ms) showed positive signals for both WM and GM (Figure 5B). Both the WM_S and GM_S T2* values remained constant (209 ± 9 µs and 258 ± 4 µs, respectively) before and after exchange.

Discussion

Ultrashort T2 signals (50 µs<T2<1 ms) in myelinated nerves and excised rat spinal cord are thought to primarily originated from myelin lipid protons.^2,3 In this study, the source of ultrashort T2 signals in ovine brain was explored. Instead of empirically choosing TI_null, TI_nulls herein were estimated through measuring WM_L and GM_L T1s using the same IR-UTE sequences that were used for subsequent ultrashort T2* quantification. This enables accurate nulling of long T2 signals in conditions of tissue T1 change due to different pathological conditions. T1 change after D₂O exchange might be caused first by a decrease of long T2 proton fraction, followed by an increase of D₂O concentration in tissue. D₂O tends to hamper proton longitudinal magnetization relaxation.⁴ Tissue degradation may also cause T1 increase,⁵ which is not considered here and needs further investigation.

D2O is not detectable using conventional ¹H MRI techniques.⁶ After prolonged exchange with D₂O, the specimens had negligible long T2 PD-FSE signals, but were still detectable on the IR-UTE images. The obtained T2* values (200-300 µs) were constant for both WM_S and GM_S before and after D₂O exchange, and were comparable with those measured in myelin extract powder as well as in mixtures with D₂O.⁷ These results support the view that the IR-UTE signal source is from non-aqueous protons.

Conclusion

The ultrashort T2* signals seen on IR-UTE images are unlikely to be from water protons exchangeable with deuterons in D₂O. Their source is more likely to be myelin itself. The IR-UTE sequences may serve as a promising tool for directly imaging myelin loss in WM, as well as in GM.

Acknowledgements

The authors acknowledge grant support from the NIH (1R01 NS092650). The authors would like to thank Dr. G.M. Bydder for his valuable inputs on data interpretation and proofreading.

References

1. Du J, Ma G, Li S, Carl M, Szeverenyi NM, VandenBerg S, Corey-Bloom J, Bydder GM. Ultrashort echo time (UTE) magnetic resonance imaging of the short T2 components in white matter of the brain using a clinical 3T scanner. NeuroImage 2014;87:32-41.

2. Horch RA, Gore JC, Does MD. Origins of the ultrashort-T2 1H NMR signals in myelinated nerve: a direct measure of myelin content? Magn Reson Med 2011;66(1):24-31.

3. Wilhelm MJ, Ong HH, Wehrli SL, Li C, Tsai PH, Hackney DB, Wehrli FW. Direct magnetic resonance detection of myelin and prospects for quantitative imaging of myelin density. Proc Natl Acad Sci U S A 2012;109(24):9605-9610.

4. Zhong JH, Gore JC, Armitage IM. Quantitative studies of hydrodynamic effects and cross-relaxation in protein solutions and tissues with proton and deuteron longitudinal relaxation times. Magn Reson Med 1990;13(2):192-203.

5. De Guio F, Reyes S, Duering M, Pirpamer L, Chabriat H, Jouvent E. Decreased T1 Contrast between Gray Matter and Normal-Appearing White Matter in CADASIL. Am J Neuroradiol 2014;35(1):72-76.

6. Muller S, Seelig J. In vivo NMR Imaging of Deuterium. J Magn Reson 1987;72(3):456-466.

7. Sheth V, Shao H, Chen J, Vandenberg S, Corey-Bloom J, Bydder GM, Du J. Magnetic resonance imaging of myelin using ultrashort Echo time (UTE) pulse sequences: Phantom, specimen, volunteer and multiple sclerosis patient studies. Neuroimage 2016;136:37-44.

Figures

Figure 1 (A) Diagram of the 2D IR-UTE pulse sequence with a minimum TE of 10 µs. The slice-selective gradients were turned off because the specimens were ~3 mm thin. (B-C) Expected magnetization changes of short and long T2 components (WM_S, GM_S, WM_L and GM_L), with an inversion time that is chosen to completely null the WM_L (B) or GM_L (C) using an adiabatic IR pulse. The ultrashort T2 components experience significant transverse relaxation during the long adiabatic inversion process, and are not inverted but partly saturated.

Figure 2 (A) Contribution of long T2 components to IR-UTE signals is dependent on the inversion time (TI) used. (B) IR-UTE images of the cerebral hemisphere specimen before D₂O exchange with variable TIs and TR/TE = 1000/2.2 ms. In these images, contribution from WM_S and GM_S was negligible due to their ultrashort T2*s. The contrast between WM and GM changed significantly with TI used. (C) T1 map calculated from images in Figure (B). Three WM ROIs were defined to quantify the average WM T1. (D) T1 fitting curves of a WM ROI (left) and a GM ROI (right).

Figure 3 (A) IR-UTE images acquired from one cerebral specimen (no.1) before exchange with D₂O at a series of TE with TR/TI_null (for WM_L) = 1000/300 ms. The calculated T2* map (lower panel, right corner) demonstrates a T2* value ranging from 200 to 300 ms in WM. (B) The magnitude subtraction image (TE_10ms – TE_0.6ms) provides a positive contrast for WM. (C) Mono-exponential fitting of the IR-UTE images in Figure A gives a T2* value of 206 µs in the defined ROI (yellow box, Figure B).

Figure 4 (A) PD-FSE images of one cerebral hemisphere specimen (no.1) acquired before (0 min) and after sequential immersion in D₂O for 90 min, 150 min, 24 hrs, 4 days and 7 days. PD signals progressively decrease with increased immersion time. (B) Corresponding IR-UTE images (TE = 10 µs) of the same specimen. (C) Quantitative sequential signal intensity changes of the PD-FSE images in Figure (A), as measured in in three WM ROIs (A, yellow boxes, 0-min PD-FSE image). (D) Corresponding average T1 changes in these WM ROIs. (E) Fluctuation of T2* values sequentially measured in these ROIs.

Figure 5 (A) PD-FSE image of the cerebellar specimen before exchange with D₂O. (B) Selected corresponding IR-UTE images with variable TEs. At TI/TR = 260/1000 ms (top panel), the magnitude subtraction image of TE_10µs–TE_0.6ms highlights WMS. At TI/TR = 350/1000 ms, the subtraction image both WMS and GMS, with WMS showing higher signal. (C) T2* fitting curves in a WM ROI and a GM ROI. (D) Sequential changes of average PD-FSE signal intensity, T1 of WML and GML, and T2* of WM_S and GM_S in the defined ROIs in Figure C after immersion in D₂O.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)

4733