Temperature effects in post mortem structural MRI of human brain in situ

Gunther Helms^1,2, Arne Wrede³, Peter Dechent², and Walter Schulz-Schaeffer³

¹Medical Radiation Physics, Lund University, Lund, Sweden, ²Cognitive Neurology, Göttingen University Medical Center, Göttingen, Germany, ³Neuropathology, Göttingen University Medical Center, Göttingen, Germany

Synopsis

Structural 3D MRI and FLASH-based mapping of T₁ and magnetization transfer (MT) at 3T was performed in situ on 11 subjects with probable Creutzfeldt-Jacob disease prior to autopsy. The mean diffusivity in ventricles yielded temperatures between 6°C and 29°C. T₁ contrast in the deep brain decreased with temperature and vanished under refrigerated conditions. T₁ of gray matter lowered towards the normal white matter’s T₁ which did not change. The MT saturation was generally independent of temperature, except in normal WM below 13°C. Above 13°C, MT maps yield a high contrast and can be used for quantitative assessment of structural changes.

Target audience

MR physicist, neuroradiologists, forensic experts

Purpose

Post mortem (p.m.) MRI is increasingly performed in situ to bridge the gap between clinical examinations and MRI on the fixated brain [1]. Relaxation and thus MRI contrast depends on the molecular motion of water, and hence on temperature. This study aims to access the influence of the in situ temperature on structural p.m. MRI.

Methods

Patients with clinically probable Creutzfeldt-Jacob disease (CJD confirmed in 7/11 subjects) were scanned before autopsy at 3T (Siemens TimTrio) within 36 hours p.m. Four elderly subjects (56/65/79/81 years) were included as in vivo controls. Informed written consent had been obtained from patients or next of kin. Temperature was determined from mean diffusivity in ventricles [2] (b=0/500/1000 mm²/s, 38 slices, 1.9+0.6 mm thickness, 1.51 mm resolution). 3D structural MRI were performed as in vivo at 1 mm³: MP-RAGE (TI/TR/TE = 900/2250/3.26 ms, α=9° [3]) TSE (TR/TE= 2900/419 ms with variable refocusing angle), and FLAIR (TR/TE= 6000/403 ms with variable refocusing angle). 3D FLAIR was run with TI = 2100/1600/1400 ms to cover temperatures down to refrigeration [2]. FLASH-based parameter mapping [4] at 1.25 mm³ yielded T₁, signal amplitude and magnetization transfer saturation (MTsat). MTsat, the percent signal loss imposed by a single MT-pulse [4], is independent of underlying T₁ and flip angle bias, and thus particularly suited for pm studies. Volumes were co-registered and analyzed in bilateral regions-of-interest in deep brain (frontal lateral ventricle, adjacent caudate head and genu).

Results

5/7 CJD brains appeared structurally normal, except for age and disease-related atrophy; 5 brains had vascular lesions and atrophy of varying degree; one was normal (abuse-related dementia).

Temperature in ventricles was between 6°C and 29°C. A progressive loss of contrast in the T₁-weighted MP-RAGE images with decreasing temperature was observed, especially at low temperatures and in the deep brain region (Figure 1). The T₂-weighted contrast created by the spin echo train did not change much, but FLAIR contrast was strongly influenced by TI (Figure 2). Too short TI hampered the visibility of periventricular vascular lesions. Cortical contrast was improved at TI longer than for optimal CSF suppression (approx. TI=1750 ms at room temperature).

T₁ changed most strongly in cerebro-spinal fluid (CSF; 80ms/°C), slightly in gray matter (GM; 11.5 ms/°C), while no change was detectable in white matter (WM) (Figure 3). MT saturation, however, did not change in GM, and in WM only below 13°C (Figure 4). Above 13°C, the MTsat maps provided consistent structural contrast independent of temperature (Figure 5).

Discussion

Image contrast on structural MRI was mainly influenced by T₁, which is decreasing with temperature. After refrigeration, T₁ of GM was very similar to T₁ of WM, which hardly changed with temperature. Hence, “structural” T₁-weighted contrast cannot be established by conventional means. MT saturation maps, however, were nearly independent of temperature and thus provided consistent structural contrast for pm MRI in situ when the temperature could not be controlled.

The contribution to T₁ relaxation from cross-relaxation with “invisible” macromolecules (as measured by MTsat) was found to be largely independent of temperature. This explains the smaller influence of temperature in parenchyma compared to CSF. An explanation of T₁ constancy in WM can be derived from a 4-pool model, where much of the MT and T₁ relaxation is exerted on the water within the myelin sheaths and then transferred to bulk water by exchange across the phosphor-lipid bi-layers. As these get progressively impermeable at low temperatures, the rapidly-decaying myelin water contributes less to the observed T₁. This effect counteracts the shortening of bulk T₁. Myelin permeability also explains the drop of MTsat in WM below 13°C.

The correlation with temperature was found to be affected by individual pathological changes. Widespread vascular lesions or old age resulted in increased T₁ and decreased MT in the genu. CJD may have influenced MTsat in the caudate [5]. The specific behavior of T₁ and MT in GM and WM was experimentally confirmed in freshly excised bovine brain at 21°C and 6°C (4 and 8 hours p.m.; data not shown).

Conclusion

Post mortem structural MRI in situ is feasible with in vivo scan protocols, unless scanning below 13°C. For quantitative comparisons, FLASH-based MT-mapping is recommended because it is not affected by temperature-induced changes in relaxation.

Acknowledgements

Dr. K. Kallenberg and P. Holz are thanked for help with organization and scanning.

References

1: Grinberg LT, Amaro Junior E, da Silva AV, et al. Improved detection of incipient vascular changes by a biotechnological platform combining post mortem MRI in situ with neuropathology. J Neurol Sci. 2009;283:2–8.

2: Tofts PS, Jackson JS, Tozer DJ, et al. Imaging Cadavers: Cold FLAIR and Noninvasive Brain Thermometry Using CSF Diffusion. Magn Reson Med. 2008;59(1):190–195.

3: Jack CJ, Bernstein M, Fox N, et al. The Alzheimer’s Disease Neuroimaging Initiative (ADNI): MRI methods. J Magn Reson Imaging 2008;27:685–691.

4: Helms G, Dathe H, Kallenberg K, et al. High-Resolution Maps of Magnetization Transfer with Inherent Correction for RF Inhomogeneity and T1 Relaxation Obtained from 3D FLASH MRI. Magn Reson Med. 2008;60(6):1396–1407.

5: Helms G, Matros M, Kallenberg K, et al. Correlation of magnetization transfer (MT) and diffusion MRI in sporadic Creutzfeldt-Jacobs disease. Proc. ISMRM. 2014:22;7350.

Figures

Figure 1: T₁-weighted MP-RAGE of subjects at different temperature. Contrast between WM and thalamus, striate and cortex is progressively lost with decreasing temperature. The CJD patient at 29°C also showed vascular lesions (arrow).

Figure 2: FLAIR of a subject with vascular lesions at different TI. TI(0) for signal zeroing was determined by interpolation. At TI<TI(0) the lesion signal is attenuated. At TI>TI(0), improved GM-WM contrast is traded in against incomplete CSF suppression. The body bag’s zipper caused the frontal artifact.

Figure 3: T₁ over temperature; in vivo measurements are included at 37°C. ROI averages from right (solid) and left (open) hemisphere are plotted with the standard deviation as error. In genu, T₁> 900 ms represent subjects with extensive WM lesions.

Figure 4: MT saturation over temperature. ROI averages from right (solid) and left (open) hemisphere are plotted with the standard deviation as error. In genu, the open symbols of low MTsat represent subjects with extensive WM lesions.

Figure 5: MT saturation maps co-localized to Figure 1. When compared to MP-RAGE, contrast in the deep brain region and within WM is improved. The contrast is less influenced by temperature, with a minor reduction observed at 9°. The arrow indicates a vascular lesion.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

4040