Introduction

Brain retention of some gadolinium-based contrast agents (GBCAs) has been reported in patients and laboratory animals after multiple doses¹. However, the current methods for investigation of this phenomenon are mainly based on relative signal intensity changes in T₁-weighted MRI (T₁w). Quantitative relaxometry may provide more precise tool for animal studies. Here, we propose using quantitative T₂ mapping to study brain retention of GBCAs in rats at 7 tesla field strength.

Methods

The animal use protocol was approved in advance by the NCTR IACUC. Male Sprague-Dawley rats (N = 12, 108 ± 3 days old, 480 ± 32 g) were surgically implanted with jugular vein catheters with externalized injection ports (Rat-O-Port, Braintree Scientific, Inc). Six rats received Omniscan injections (0.62 mmol/kg, iv) over 5 weeks (4 injections/week, 20 total) and six rats were injected with saline using the same injection regime and volume (1.24 ml/kg). Brains were imaged before the start of treatment and weekly thereafter (6 scans/subject). MRI was performed using a 7 tesla Bruker Biospec AV III equipped with 12 cm ID gradient insert (440 mT/m) and 38 mm litz-cage quadrature RF coil (Doty Scientific, Inc). The following MRI protocols were used: 1) T₁w fast spin echo sequence, similar to ²: TR = 700 ms, TE = 8.4 ms, MTX = 192 × 192, RARE = 2, FOV = 3.84 × 3.84 cm, 28 slices, 1 mm slice thickness, acquisition time ~ 5 min; 2) T₂ mapping spin echo sequence, as described in 3: TR = 6000 ms, TE = 15 ms, ETL = 12, MTX = 192 × 192, FOV = 3.84 × 3.84 cm, 28 slices, 1 mm slice thickness, acquisition time ~20 min. Both T₁w and T₂ images were acquired using the same geometrical coordinates, so they were intrinsically co-registered. T2 maps were calculated off-line as described in ³. Before each MRI scan, open field locomotor activity was assessed. At the end of the observation, rats were perfused trans-cardially with 4% paraformaldehyde as described in ⁴ for follow-up histopathological evaluation of the brain. The regions of interest (ROIs), which delineated deep cerebellar nuclei (DCN)2 were drawn manually on T₂ maps as they provided better contrast than T₁w for visual boundary detection. As both T₁w and T₂ maps were co-registered for each rat, the same ROIs were used to quantify T₂ values and T₁w image intensities. Additionally, T₁w image intensities in ROIs were normalized against adjacent cerebellum signals (by shifting ROIs upwards by 8 pixels, 1.6 mm – reference ROI) as suggested in ². The ratios of T₁w signal intensities from the ROI of interest to the reference ROI are presented. Statistical analysis was performed using repeated measures ANOVA.

Results

Figure 1 shows an example of manual ROI positioning (red) on T₂ maps and T₁w images and shifted reference ROI (green) on T₁w images. Figure 2 shows that quantitative T₂ mapping could positively detect the accumulation of Omniscan in DCN of rats’ brains, starting from week 3 of dosing (*), whereas the T₁w approach could detect the contrast changes only at 5 weeks, and showed overall larger scattering and less dynamic range. Cohen’s effect size was much higher using the T₂ mapping method (7.53) vs. the T₁w method (2.97) at the 5 week point. There was a moderate negative correlation between T₂ values and T₁w signal intensity ratio (R = -0.574, P= 0.00002). These results suggest that even though Omniscan is predominantly a T₁ agent, T₂ effects are more readily detectable in the current settings. Moreover, T2 mapping provides intrinsically quantitative data, which is less prone to errors related to imaging setup (e.g., RF coil homogeneity among others) as well as to errors due to image analysis (no need for reference ROI for ratio calculations). In addition, the known change in T₂ relaxation due to GBCAs retention could be roughly translated into its concentration in the tissue. Given the R₂ relaxivity of Omniscan in human plasma at 7T of 3.09 mmol/L ⁵, its concentration is estimated at ~0.39 mmol/L in DCN. Omniscan did not alter open field locomotor activity or cause any detectable histological brain abnormalities.

Conclusion

Repeat dosing with Omniscan has been demonstrated to lead to its retention in the DCN of the rat brain and quantitative T₂ mapping provides a robust method for its quantification. Further studies with the additional GBCAs are warranted.

Acknowledgements

This work was supported by the National Center for Toxicological Research (NCTR), US FDA (protocol number P00802).