Abstract
BACKGROUND: Brain temperature signifies the thermal homeostasis of the tissue, and may serve as a marker for neuroprotective therapy. Currently, it remains challenging to map the human brain temperature with high spatial resolution. The thermal dependence of chemical exchange saturation transfer (CEST) effects of endogenous labile protons may provide a promising mechanism for the absolute brain temperature imaging. In this study, we aimed to investigate the temperature dependency of the CEST effect of creatine (CrCEST), and contemplate its feasibility for brain temperature mapping. METHODS: Creatine (Cr) was selected as the endogenous agent to probe the brain temperature. Proof-of-concept phantom experiments were first conducted using a 400 MHz nuclear magnetic resonance (NMR) spectrometer and a 5.0 T magnetic resonance (MR) scanner at various temperatures. A multi-pool Lorentzian fitting model was utilized to quantify the apparent chemical shift, amplitude, linewidth, and integral of CrCEST peak at around 2 ppm. Regression analysis was performed to estimate the thermal response of these CrCEST parameters. Finally, the temperature mapping of ex vivo swine brain tissues was conducted based on the CEST signal at 2 ppm (CEST@2ppm). RESULTS: A robust linear correlation between the apparent chemical shift of CrCEST and temperature was identified in the in vitro phantom experiments (+0.005 ppm/℃, P<0.001), based on which temperature maps of phantoms were generated. In the subsequent experiments on ex vivo swine brain tissue, a comparable temperature dependency of the apparent chemical shift of CEST@2ppm peak was found (+0.008 ppm/℃), confirming the utility of this approach for mapping brain temperature. CONCLUSIONS: The CEST-based approach proposed in this study suggests a path toward label-free brain thermometry in vivo at high resolution and may be potentially applied in other tissues such as muscle and kidney.