MRI technique detects penumbra following stroke by assessing O2 metabolism: validation with [14C] 2-deoxyglucose autoradiography

Craig Robertson1, Chris McCabe1, William Holmes1, Lindsay Gallagher1, Rosario Lopez-Gonzalez1,2, David Brennan1,2, Barrie Condon1,2, Keith Muir1,2, Celestine Santosh1,2, Mhairi Macrae1. 1University of Glasgow, Glasgow Experimental MRI Centre (GEMRIC), Glasgow, United Kingdom, 2Institute of Neurological Sciences, Southern General Hospital, Glasgow, United Kingdom.

Objectives: Following stroke, the ischaemic penumbra represents potentially salvageable tissue surrounding the irreversibly-damaged ischaemic core. Diffusion-perfusion mismatch (DWI-PWI), the current MRI method for identifying penumbra, cannot detect ongoing penumbral metabolism. We have developed a novel MRI technique that should enable a more precise definition of penumbra where 100% oxygen is used as a metabolic biotracer to detect tissue metabolism. Delineating a metabolically active tissue compartment bordering ischaemic core may accurately define viable, candidate penumbral tissue capable of recovery. Penumbral tissue, which has an increased oxygen extraction fraction, should display an increased T2* signal change during oxygen challenge (OC), which we believe equates with persisting tissue metabolism. This assumption was tested by co-registering T2* OC brain images with [14C]-2-deoxyglucose (14C 2-DG) autoradiographic sections, for determination of local cerebral glucose utilisation (LCMRglu).

Methods: Male Sprague Dawley rats (280-320g, n=6) were anaesthetised (2% isofluorane), ventilated (air) and the middle cerebral artery permanently occluded (MCAO) by an intraluminal filament. Imaging (8 coronal slices throughout MCA territory) was performed on a Bruker Biospec 7T/30cm MRI scanner. T2-weighted images provided a neuroanatomical template for co-registration of MR images and 2-DG autoradiograms. Diffusion-weighted imaging (DWI) produced apparent diffusion coefficient (ADC) maps of ischaemic damage and perfusion-weighted imaging (PWI, by continuous ASL) determined the perfusion deficit. A T2*-weighted MRI sequence was run during OC: 5 mins air, 5 mins 100% oxygen & 15 mins air inhalation. Following OC, rats were received an iv bolus of 125 μCi/kg 14C 2-DG and LCMRglu was determined in MRI-defined regions of interest: cortical ischaemic core, contralateral cortex and penumbra, as defined by PWI/DWI mismatch.

Results: Mean (±SD) time from MCAO to start of OC was 171±27.7min. OC induced a T2* signal increase of 3.7±1.2% in contralateral cortex. Within ischaemic core, LCMRglu was reduced by 96% (relative to contralateral cortex) and OC induced a T2* change of

-2.5±2.3%. In penumbra, a 40% increase in LCMRglu co-registered with the greatest (8.34±3.1%) increase in T2* signal change. Interestingly, hyperglycolysis (75% increase in LCMRglu) at the ADC lesion boundary, corresponded with a 1.53% T2* signal increase.

Conclusions: Co-registration of MRI and 2-DG autoradiograms validates the detection of penumbral tissue by the OC technique. The area of greatest T2* signal increase, co-registered with penumbra as defined by PWI/DWI and was metabolically active, as defined by LCMRglu. Negligible T2* signal change and glucose utilisation was recorded within ischaemic core. Hyperglycolysis which bordered the ADC lesion and coregistered with a lower T2* increase, may be indicative of anaerobic glycolysis.