Quantification of Cerebral Perfusion using Laser Speckle Imaging and Infarct Volume using MRI in a Pre-clinical Model of Posterior Circulation Stroke

Summary

A canine model of LVO stroke was utilized to develop laser speckle imaging to monitor cerebral perfusion in real-time.  Diffusion-weighted MRI was optimized to image infarct volume utilizing a high b-value, enabling ADC and MRA, correlated with DSA at the time of stroke.  Finally, ADC reconstructions correlated with histological findings.

Abstract

Background: Basilar artery occlusion (BAO) is a subset of posterior circulation stroke that carries a mortality as high as 90%.  The current clinical standard to diagnose ischemic stroke include computerized tomography (CT), CT angiography and perfusion and magnetic resonance imaging (MRI). Large animal pre-clinical models to accurately reflect the clinical disease as well as methods to assess stroke burden and evaluate treatments are lacking.

Methods: We describe a canine model of large vessel occlusion (LVO) stroke in the posterior circulation, and developed a laser speckle imaging (LSI) protocol to monitor perfusion changes in real time.  We then utilized high b-value DWI (b=1800s/mm²) MRI to increase detection sensitivity. We also evaluated the ability of magnetic resonance angiography (MRA) to assess arterial occlusion and correlate with DSA. Finally, we verified infarct size from apparent diffusion coefficient (ADC) mapping with histology. 

Results:  Administration of thromboembolism occluded the basilar artery as tracked by DSA (n=7).   LSI correlated with DSA, demonstrating a reduction in perfusion after stroke onset that persisted throughout the experiment, allowing us to monitor perfusion in real time.  DWI with an optimized b-value for dogs illustrated the stroke volume and allowed us to derive ADC and magnetic resonance angiography (MRA) images. The MRA performed at the end of the experiment correlated with DSA performed after occlusion. Finally, stroke burden on MRI correlated with histology.

Conclusions: Our studies demonstrate real time perfusion imaging using LSI of a canine thromboembolic LVO model of posterior circulation stroke, which utilizes multimodal imaging important in the diagnosis and treatment of ischemic stroke.

Introduction

The prevalence of stroke worldwide is almost 25.7 million, the majority of which are ischemic¹.  Posterior circulation stroke accounts for 20% of all strokes of which basilar artery occlusion is the most severe, approaching 90% mortality^1,2.  In 1995, recombinant tissue plasminogen activator (rtPA) was the first acute therapy developed for ischemic stroke in patients who presented within 3 hours from stroke onset³. More recently, mechanical thrombectomy has demonstrated benefit in treating acute ischemic stroke in patients who present with large vessel occlusion (LVO), which includes the intracranial portion of the internal carotid artery or the first segment of the anterior and middle cerebral arteries⁴.  None of the recent clinical trials included posterior circulation stroke and its outcomes remain dismal despite utilizing mechanical thrombectomy for basilar artery occlusion^5,6.  

Advances in assessment techniques in stroke patients have an impact on predicting the chance of functional recovery and survival⁷. Pre-clinical models of posterior circulation stroke have been previously described^8,9,10, however assessing stroke burden and revascularization remain suboptimal.  Smaller species such as rodents offer several advantages including ease of genetic manipulation, inexpensive animal purchase, and low per diem housing costs^11,12. However, small animal experiments sometimes do not fully represent large animal and human vasculature, physiological conditions, or related inflammatory responses⁷. Large animals more closely mimic human stroke^2,7,13,14.  Moreover, serial blood sampling can be performed for blood analysis of thrombotic and inflammatory markers.

In this study, we describe a canine model of basilar artery occlusion verified by digital subtraction angiography (DSA) from the onset of stroke.  We utilize laser speckle perfusion imaging (LSI) to monitor perfusion in real time.  We then utilize a novel microvascular enhancement algorithm  based on laser speckle perfusion imaging (LSI) acquisition as well as a high b-value magnetic resonance imaging (MRI) technique to optimize infarct imaging¹⁵. These techniques allow us to monitor and quantify local and global ischemia. Finally, we correlate these imaging findings to histology. Understanding prognosis and the need to study posterior circulation stroke in pre-clinical models is critical in order to improve therapies. 

Protocol

All procedures were performed in compliance with the Animal Welfare Act and the Guide for Care and Use of Laboratory Animals (NRC 2011), as approved by the Ohio State University’s Institutional Animal Care and Use Committee (IACUC).

1. Step 1 Animal Preparation and Surgical Protocol A canine model of basilar artery occlusion (BAO) stroke was used as previously described^9,10.

1. Fast adult beagles (8-13 kg, 14-21 months old) overnight with free access to water.
2. Inject,  pre-anesthetic, an intramuscular administration of acepromazine (0.2 mg/kg) 
3. Introduce a 20 gauge catheter into a cephalic vein. 
4. Induce anesthesia with intravenous administration of ketamine (10 mg/kg) and midazolam (0.025 mg/kg). 
5. Following anesthetic induction, intubate dogs and mechanically ventilate using constant inhaled anesthesia (2-3% isoflurane). 
6. Create a 1 cm² craniotomy window for laser speckle imaging. 
7. Introduce a 7F arterial sheath into the right femoral artery for access and blood pressure measurement.  
8. Introduce a 16 gauge angiocatheter into the right femoral vein for blood draws.
9. Prepare thromboembolus (blood clot) as previously described¹⁶. Briefly, draw and mix 5 mL of canine whole blood with 0.5 g of barium sulfate (Ba₂SO₄) in a plastic serum blood collection tube while rolling for 30 seconds. Rest the mixture undisturbed for 60 minutes at room temperature before catheter administration.
10. Begin recording baseline digital subtraction angiography (DSA) prior to accessing the middle basilar artery. Advance a 4F guiding catheter under fluoroscopic guidance, using a retrograde trans-aortic approach, into the 7F arterial sheath previously placed into the right femoral artery through a vertebral artery to the base of the basilar artery. Inject two milliliters of contrast agent with normal saline to identify the basilar artery.
11. Using a surgical scalpel, resect the clot into small pieces with both fibrin-rich and erythrocyte-rich layers¹⁶ to load into a 3 mL syringe and inject through the microcatheter into the middle of basilar artery. Allow the clot to stabilize for 10 minutes. Perform a follow up angiogram to verify your desired clot location. Arterial occlusion can be verified by DSA and decreasing cerebral perfusion by laser speckle imaging (LSI).

2. Step 2 Laser Speckle Imaging 

1. Focus the laser speckle perfusion imaging (LSI) camera on the cranial window. Configure the high resolution laser speckle imaging (LSI) camera system as previously described¹⁵.  
2. Record perfusion with interruptions during performance of angiogram at desired time points. Acquire data from a 1.5 cm x 1.5 cm field of view using a 785 nm wavelength and 80 mW lasers with a sampling rate of 60 Hz at a working distance of 10 cm in this canine model.
3. From the real-time perfusion graphs, choose the time-of-interest (TOI) to include lower peaks only to exclude the respiratory motion related artifacts. Average relative perfusion units over a 10 s sampling period using PimSoft v1.4 software. Perform laser speckle contrast analysis (LASCA) as previously described¹⁵.  
4. To optimize the quantification of brain microvasculature in this canine model, record images at 15 frames per second and perform intensity and variance calculations with spatiotemporal averaging over a 5 x 5 pixel area with 5 frames. The overall frame rate for the intensity and variance data was 3 frames per second. Choose the median value of perfusion for each pixel to reduce the effects on the mean of large sudden changes in perfusion readings due to motion from canine respiration. Convert raw data into binary files and process the data into meaningful imaging of the vasculature. Utilize the program re-tooled LASCA algorithm (rt-LASCA) to use the variance of the contrast data over time to determine the locations of vasculature as previously described¹⁵.   

3. Step 3 Magnetic Resonance Imaging (MRI) and Magnetic Resonance Angiography

1. Perform MRI the day before surgery for comparison if desired, then repeat to confirm BAO and again before sacrifice if a therapeutic is to be evaluated.
2. Place continuously anesthetized canines head-first in a supine position as previously described  in a Siemens Prisma 3 Tesla field strength and 60 cm-diameter bore MRI scanner including a 32 channel head coil as a receiver with enhanced parallel imaging performance to obtain brain images¹⁷.  
3. Perform localizer scans to acquire pilot images of each canine brain before the anatomical imaging begins.  The system utilized to obtain the presented data has an integrated imaging system which allows faster scanning in optimal spatial and temporal resolutions. The 80 mT/m gradients generate high-quality T2-weighted, diffusion-weighted images and MR angiograms. Diffusion-weighted imaging (DWI) is sensitive enough and can show more anatomic sub-structure than by conventional structural MRI methods such as T2-weighted images. In this study, MRI was performed 4h after BAO.
4. After proper localization, perform T2-weighted gradient echo imaging (Parameters: FOV = 130 mm, Matrix size = 320 x 320, pixel size = 0.3 x 0.3 mm, Slice thickness =3 mm, TR= 4s, FA= 180 degrees, BW =255 Hz/pixel, NEX= 2, TE=75ms, Resolution= 2.4615 pixels per mm) followed by a flow attenuated inversion recovery (FLAIR) imaging to visualize the structure of the brain anatomy.
5. Perform magnetic resonance angiography (MRA) to visualize the vascular anatomy and blood circulation measurement. Acquire MRA of the brain covering the head and neck with a time-of-flight-3D (TOF) sequence in transverse view (Parameters: FOV = 129x129 mm, Matrix size = 768 x 768, pixel size = 0.3 x 0.3 mm, slice thickness = 81.59 mm, TR= 25 ms, FA= 18 degrees, BW =185 Hz/pixel, NEX= 1, TE=4.22ms, Resolution = 5.91 pixels per mm). Perform maximum intensity projection (MIP) with 3D color-coded visualization to maximize the signal intensity in the blood vessels.  Post-process acquired DICOM images to visualize the blood vessels and to confirm that the basilar artery was occluded.

4. Step 4 Diffusion weighted imaging and stroke volume calculation

1. Perform diffusion weighted imaging sequence to detect acute ischemic strokes (Parameters: FOV = 149mm x149 mm, Matrix size = 132 x0x0x 100, pixel size = 0.30mm x 0.30 mm, slice thickness = 4 mm, TR = 4.6s, FA = 90 degrees, BW = 255 Hz/pixel, NEX= 1, TE = 86ms, Resolution= 0.93 pixels per mm). Transfer DICOM images for post-processing. 
2. Generate apparent diffusion maps (ADC) from DWI images and calculate infarct volumes using OsiriX MD v.5.0 software. 
3. Trace both the brain hemispheres and infarct areas per slice and multiply by slice thickness to acquire infarct volumes. 
4. Convert the absolute whole volume to 100 units to calculate the percent stroke volume of each canine.

5. Step 5 Hematoxylin and Eosin staining brain histology 

1. At time of sacrifice in anesthetized canine, harvest the brain and cut two medial sections 4 mm thick with a sharp scalpel, one section will be used for TTC staining below.
2. Fix the 4 mm section in 10% formalin for a minimum of 7 days to allow infiltration throughout the entire section.
3. Embed the fixed brain section in paraffin following our protocol¹⁷. 
4. Trim and level each paraffin block (multiple blocks can be stored and processed at the same time). 
5. Section each paraffin block at 4μm and place cut tissue on a 2" x 3" inch slide.
6. Process each slide in Hematoxylin 560 for 8 min, differentiate with 1% acid alcohol for 1s three times with rinsing in tap water. 
7. Blue each slide with 1% ammonium hydroxide for 1s and rinse for 2s with tap water.
8. Dehydrate in 70% ethanol for 1s twelve times, counterstained in eosin for 1 min.
9. Dehydrate in 95% for 1s twelve times followed by 100% ethanol. 
10. Clear in xylene and apply a 2" x 3" inch coverslip with mounting media, removing air bubbles.

6. Step 6 2% 2,3,5-triphenyltetrazolium chloride brain staining

1. Place the second 4 mm section which was harvested beside the H&E section into a previously prepared solution containing with 100 mL 2% 2,3,5-triphenyl-2H-tetrazolium chloride (TTC) in pH 7.4 PBS warmed to 37 °C in the dark.
2. Incubate in the dark at 37 °C for at least 20 min, flipping the brain section over gently every 5 minutes. 
3. When the section turns cherry red on both sides, remove the TTC solution and replace with 4% paraformaldehyde in PBS, pH 7.4, to optimize the contrast overnight. 
4. When the contrast is optimal between white and red staining in the brain (1-3 days), place between clear plastic sheets, dry excess fluid, and scan at high resolution.
5. Trace the ischemic regions and whole brain slide to obtain percent infarction in each section as previously described¹⁷. 

Results

Laser Speckle Perfusion Recording and Imaging: Perfusion recording was performed continuously until the animal was transported to MRI, and again at sacrifice (Figure 1A). Data showed that cerebral perfusion decreased by ~15% to 83 ± 10% at the time point before basilar artery occlusion (pre-BAO). This nominal decline is likely the result of a microcatheter insertion in the distal vertebral artery. After injecting the prepare thromboembolus, the post-BAO perfusion d...

Discussion

The most common causes of posterior circulation stroke include embolism, large-artery atherosclerosis, and small artery disease⁵. Basilar arterial occlusion (BAO) represents a subset of posterior circulation strokes, carrying significant morbidity and mortality¹³. In this context, a canine model of acute posterior stroke was utilized and we developed an LSI protocol to monitor perfusion of the occluded region in real time. Laser speckle perfusion imaging was performed throu...

Materials

Name	Company	Catalog Number	Comments
2% 2,3,5-triphenyltetrazolium chloride (TTC in PBS, pH 7.4)	Sigma Aldrich	T8877
EDTA K3 vacutainers	Becton Dickinson	BD455036
Eosin	Surgipath	3801602
Formalin, neutral buffered, 10%	Richard-Allan Scientific	5701
Hematoxylin 560	Surgipath	3801570
HUG-U-VAC positioning system	DRE Veterinary	1320
LabChart Software	ADInstruments Inc.
Laser Speckle Imaging camera	Perimed Inc., Jarfalla, Sweden	PeriCam PSI HR System
Lithium heparin vacutainer, 4.5%	Becton Dickinson	BD 368056
Matlab	The MathWorks, Inc., Natick, MA
OsiriX MD v.5.0 software	Pixmeo Inc, Geneva
Paraformaldehyde 4% in PBS	Alfa Aesar	AAJ61899AP
PimSoft v1.4 software	Perimed Inc.	software that accompanies LSI equipment
Prisma Fit 3 tesla (3T) magnet	Siemen's Diagnostics
Sodium heparin for injection (to coat blood gas syringe)	NovaPlus	402525D