236 Non-Invasive Cerebral Temperature Mapping by Proton Spectroscopic Imaging

Abstract

Background: Cerebral hypothermia shows promise as a neuroprotective strategy following perinatal hypoxia-ischaemia. A non-invasive technique for the quantification of regional brain temperature is urgently required to assess the cerebral effects of different cooling strategies. We aimed to test the hypothesis that thermometry by proton magnetic resonance spectroscopic imaging (1H-MRSI) is as accurate as invasive monitoring during whole body hypothermia and selective head cooling.Methods: Cerebral magnetic resonance (MR) data was acquired from 6 newborn piglets using a 7 Tesla Bruker Biospec MR system. Mild whole body hypothermia was induced in 3 piglets using a thermally regulated water mattress. For each animal measurements were collected at rectal temperatures (TR) of 38±1°C and 34±1°C. Mild hypothermia was induced in 3 piglets using a water-filled plastic “cooling cap” applied to the animals' scalp. In these animals measurements were obtained at (i) TR 38±1°C; cap temperature (TC) 20°C (ii) TR 34±1°C; TC10±2°C. 1H-MRSI data were acquired from a coronal slice (thickness 4mm; TE 45ms; TR 3000ms, 16x16 voxels each 3.75 x 3.75mm2) and processed using in-house software: for each voxel, separate water and metabolite spectra were obtained from a single non-water suppressed spectrum. The chemical shift difference between the water peak and n-acetylaspartate (NAA) was obtained by a cross-correlation method. This chemical shift difference was converted to absolute temperature using a previously validated calibration curve. A two dimensional cerebral temperature map of the selected slice was obtained in a single measurement. Invasive brain temperature measurements were obtained concurrently using a surgically inserted fibre-optic thermometer, which yielded 4 temperature values at successive 5mm intervals from the probe tip.Results: The graphs demonstrate 4 data points for each subject at each TR.Conclusion: 1H-MRSI is an accurate and robust non-invasive method of cerebral temperature measurement in both total body and selective head cooling models. The determination of regional brain temperature distributions in infants undergoing hypothermia treatment may provide new insights into the local effects of brain cooling and enable its neuroprotective potential to be optimised.