3T BOLD MRI with low intrascan variability and high reproducibility of limb oxygenation measurements

Background Current imaging methods, including BOLD MR imaging, cannot reliably quantify muscle oxygenation in patients with limb ischaemia. The BOLD response is complex and measurements have so far been performed in relatively low-resolution EPI-readout images where it may be difficult to exclude oxygenation changes from partial volume effects. The reproducibility of BOLD imaging has been low and consequently its usefulness as a clinical test has been limited. We propose a high-resolution BOLD sequence whereby edge artefacts and vessels may be excluded from measurements, analysed using a Maximum Likelihood Estimate with Rician noise correction, where noise is read in k-space, for a precise and unbiased estimation of T2* from magnitude images. We test the hypothesis that this gives high reproducibility at baseline, and during and after transient ischaemia.