Introduction

Cerebral ischemia-reperfusion (CIR) is characterized by an episode of ischemia followed by restauration of blood flow to the suffering tissue. After CIR, the tissue may undergo physiological changes such as blood-brain barrier (BBB) leakage or tissue water content¹. BBB permeability ranges from abnormal water movement, subtle permeability of small to large molecules to red blood cells extravasation, the latter leading to hemorrhagic transformation (HT), critical for patients’ outcome. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a T1 weighted imaging technique used to study BBB leakage with a paramagnetic contrast agent (CA). Its modeling is influenced by the relationship between signal enhancement and CA concentration dependent of intrinsic tissue parameters such as T1. It is increasingly used in the stroke field to study subtle BBB disruption where the signal enhancement is much smaller than in tumors and multiple sclerosis². Yet, the dynamic of tissue T1 changes in CIR after recanalization therapy remains mainly unknown².Small CA like gadoterate meglumine (Gd-DOTA, 0.5kDa, 1nm diameter) are generally used, but are thought to overestimate vascular permeability³. AGuIX Nanoparticles (NPs) are medium-size CA (8.7kDa, 3nm diameter)⁴ and can be radiolabeled to combine high PET sensitivity and MRI anatomical resolution. These NPs might better quantify BBB leakage, and might also reflect the ability of medium size drug to be delivered to the stroke area. We hypothesized that AGuIX-NPs can quantify BBB permeability after CIR more accurately than small CA. We also evaluated whether local CIR T1 variations due to water movements affect BBB permeability measures.

Material and methods

Six Macaca fascicularis mature males underwent a 110-min transient minimally invasive endovascular middle cerebral artery occlusion. All experiments were approved by ethical committees (C2EA-18; C2EA-42). Simultaneous PET-MR imaging (Siemens Biograph mMR 3T) was performed during ischemia and post-recanalization. Per-ischemia MRI identified the core lesion with diffusion-weighted imaging (DWI) (TR/TE=6650/57ms; voxel size=1.3x1.3x1.3mm; TA=3:00min; b-values=0,1000), followed by post-recanalization DWI and post-contrast fluid attenuated inversion recovery (FLAIR; TR/TE=5000/418ms; voxel size=0.4x0.4x0.8mm; TA=9:15min). Post-recanalization dynamic contrast enhanced MRI (DCE-MRI FLASH; TR/TE=5.2/2.5ms; voxel size=2x2x2mm, bolus of 0.1mmolGd/kg) was done respectively with AGuIX-NPs (at 60min; 100 measurements; r1=8.9 mM-1s-1, at 0.3 mL/sec) and Gd-DOTA (at 90min; 60 measurements; r1=3.5 mM-1s-1). Prior to each CA, T10 was measured (FLASH; TR/TE=5.2/2.5ms; voxel size=2x2x2mm; TA=32sec) using four flip angles (5,10,15,25°). AGuIX-NPs are dual-labeled with [68Ga] (141±18MBq) and a 30-minute dynamic PET (31 frames) was acquired simultaneously with DCE-MRI. Imaging protocol is repeated at day 7 (day+7) after CIR. To compare CA dynamics (PET [68Ga]AGuIX-NPs, DCE-MRI AGuIX-NPs and Gd-DOTA), we extracted signal time curves from blood and ipsilateral and mirror contralateral areas.Apparent diffusion coefficient (ADC) maps (in mm2/s) were computed from DWI imaging. Maps of T10 (T1-maps) were calculated from each pre-contrast acquisitions. For each CA, we computed capillary permeability maps (Ktrans) with the extended Tofts model, placing an input function in the sagittal sinus (Olea Sphere v3.0), i.e. a quantitative Ktrans map (using T1-map to convert signal into gadolinium concentration) and a relative Ktrans map (using only the delta signal enhancement). We manually drew 4-mm circle region-of-interest (ROI) in the per-ischemia lesional and peri-lesional areas and their respective mirrors to assess T1, ADC and Ktrans evolution. ROIs with T1 values above 5s were considered artifactual and were removed. Non-parametric paired Wilcoxon test was used to compare ipsilateral versus contralateral hemispheres values (GraphPad Prism v6).

Results

All animals completed each imaging time-point. AGuIX-NPs input blood curve is representative of large molecule vascular kinetic compare to diffusible Gd-DOTA (Fig.1A). The tissue curve derived from post-recanalization [68Ga]AGuIX PET, DCE-MRI AGuIX and Gd-DOTA sequences (Fig.1B) show similar kinetic patterns, with higher signal in the ischemic hemisphere compared to controlateral. Lesional and perilesional ADC values significantly decreased during ischemia and post-recanalization compared to contralateral mirror values and increased at day+7 after CIR (Fig.1A). We also observe opposite T1 and Ktrans values evolution (Fig.1B-C).We found that post-recanalization Ktrans values were consistent with CA diameter (higher values when diameter is lower; see Table). Illustrative maps in one animal show similar BBB leakage pattern with the two CA (Fig.3). When comparing post-recanalization Ktrans values (Fig.4) obtained with intrinsic T1 values (Ktrans) or without (relative Ktrans), we found excellent (for AGuIX: Fig.4A; r=0.96) and good (for Gd-DOTA; Fig.4B; r=0.76) correlation between the two parameters in the pathological hemisphere, where BBB leakage is visible. Overestimation of relative compare to quantitative Ktrans values is similar with both CA (regression slope of 1.49 for AGuIX and 1.36 for Gd-DOTA), but more dispersed with Gd-DOTA.

Discussion and Conclusion

This study evaluates early BBB damage after CIR, ranging from water distribution changes on ADC and T1 to important leakage of large molecules. With small (Gd-DOTA) and large molecules (AGuIX-NPs), quantitative Ktrans values were consistent with CA size, also suggesting that BBB leakage can be detected with a medium size CA, more representative of drug diameter and molecular weight. This confirm that 60-min post-recanalization is in the therapeutic window for neuroprotective drug administration. Overestimation of relative versus quantitative Ktrans, prevent its use to define a threshold for HT and confirm T1-mapping usefulness in BBB leakage quantification. This multiparametric characterization opens the way for radiomics analysis in stroke after recanalization.

Acknowledgements

This work was supported by the ANR CYCLOPS (ANR-15-CE17-0020) and the RHU MARVELOUS (ANR-16-RHUS-0009) of Université de Lyon, within the program Investissements d’Avenir operated by the French National Research Agency (ANR). The PhD salaries of Oceane Wateau (Cifre, Cynbiose) and Justine Debatisse (Cifre, Siemens) are co-funded by the French Ministry of Higher Education and Research (ANRT).

References

1. K. Villringer, et al, DCE-MRI blood – brain barrier assessment in acute ischemic stroke. Neurology, 433–440 (2016). 2. P. A. Armitage, et al, Use of dynamic contrast-enhanced MRI to measure subtle blood-brain barrier abnormalities, Magn. Reson. Imaging 29, 305–314 (2011).3. Y. Yan, et al, Contrast agents in dynamic contrast-enhanced magnetic resonance imaging, Oncotarget 8, 43491–43505 (2017).4. L. Sancey, et al, Long-term in Vivo clearance of gadolinium-based AGuIX nanoparticles and their biocompatibility after systemic injection, ACS Nano 9, 2477–2488 (2015)