Introduction

Magnetic resonance imaging of post-mortem brain samples can reveal valuable information about different brain properties and clinical disease [1-3]. Tissue preservation during prolonged scanning not performed in situ may require substitution of the CSF-fluid by fixation agents. Tissue penetration and cross-linking may be time-consuming, especially for large samples. The rate of fixative diffusion has been previously determined to be 0.55-0.78 [4-5]. A more direct MRI-based measure of this factor is desirable.

Previously, the change in T₂ during immersion fixation has been modelled as contributions from fixed, unfixed and decomposed tissue in varying fractions [2], hence this MR-parameter may putatively serve as a proxy for the fixation process. Precise measurements of transverse relaxation necessitate use of the CPMG technique to avoid influence of diffusion-effects on quantification at long echo-times. CPMG are inherently challenging, since care must be taken to avoid slice crosstalk and requires the use of a train of high SAR 180° pulses, possibly leading to tissue heating. For 3D measurements the effective transverse relaxation time, R₂* , can be used but the relative contributions from R₂ and R₂` need to be investigated.

In the present work we investigate the use of R₂* maps to monitor penetration and fixation with formalin-based agents.

Method

Four formalin-based fixative agents with/without polyvinylpyrrolidone (PVP) and different salt concentrations were formulated to achieve different dielectric properties and to optimize B₁-transmit homogeneity at 9.4T.

Pig brain samples were bought from a slaughter house and placed in fixative solutions with a post-mortem interval of less than 6h and were kept at room temperature at all times. For each fixative a sample was embedded in one of the chambers of a container subdivided into four equal segments, using sterile gauze soaked in the same fixative. Simultaneous MR-scanning of all samples was performed at 12 hours, 1, 2, 3, 4, 13, 19, and 28 days after start of fixation, at 9.4T in a Siemens whole body MR scanner using a 16-channel transmit array in combination with a 31-element receive array [8]. R₂* were obtained from fitting the multi-exponential decay data acquired with a multi-echo GRE (TE= 6.03, 12, 18, 24, 30 ms; TR=34 ms; FA =15°; voxel size=0.4 mm isotropic; FoV=204×165.8×46 mm3). R₂-maps were obtained from CPMG-spin-echo-images of a single slice; (TR=3s ; 32 echoes in steps of 7.2ms; TE¹=7.2,TE³²=227.6ms; voxel size: 0.47x0.47x2mm³; FoV=192x192x2 mm3) analyzed with the extended phase graph model. The data acquired on different days were co-registered using SPM12 and were mapped to the same spatial grid using nearest neighbor interpolation. White (WM) and grey matter (GM) voxels were identified from anatomical T₁-weighted scans at the first time point. In line with previous studies, the kinetics of the changes in R₂* was well-described by a sum of a mono exponential recovery and decay and allowed pixel-wise fitting, R₂* = A.e^-at + B(1-e^-bt), where and are the coefficients for the exponential decay and recovery, respectively. The time-point at which the maximum R₂* was reached, was determined numerically after pixel-wise fitting

Results

Figure 1 shows the R₂*, R₂ and R₂` maps of Fix02 at day 28. Considering all fixatives and all time-points we found that R₂* is dominated by R₂-effects, especially in GM (Figure 2). Figure 3 shows an increase in R₂* values with fixation time, which is sudden in GM, and more prolonged in WM. The R₂* variation rate slows down after day 13 in GM, likely reflecting completion of the fixation procedure. Maximum R₂* values are reached at later time-points for PVP containing fixatives and at larger tissue depths (Figure 4). Fitting of the depths against the square-root-of the timepoint for maximal R₂* yielded coefficients of 0.9mm/sqrt(hour) without PVP in Fix01 and Fix02 and 1.5mm/sqrt(hour) with PVP (Figure 5).

Discussion and conclusion

The multi-faceted kinetics of T₂ during immersion fixation has been modelled previously as contributions from fixed, unfixed and decomposed tissue in varying fractions during fixation [3]. The two exponentials used here for R₂* likely captures the combined effect of cellular swelling and tissue decay, on the one hand, and penetration and molecular cross-linking of the fixative agent on the other hand. Fitting could be obtained voxel-wise, yielding consistent results, with maximal R₂* values reached at later time points with increasing tissue depths. Fitting yielded values that correspond well to previously reported rates of fixative diffusion. [4-5], suggesting the use of quantitative R₂* mapping to monitor formalin penetration and tissue fixation

Acknowledgements

No acknowledgement found.