Methodological comparison of fetal brain ¹H-MRS data analysis techniques and gestational age analysis using LCModel...

Proton (¹H) magnetic resonance spectroscopy (MRS) is a non-invasive imaging technique that can be used to assess brain metabolism. Establishing reference levels of fetal brain metabolites in a translatable preclinical model is crucial for identifying subtle deviations from normal cerebral development. Herein, we developed a ¹H-MRS protocol for fetal sheep across three gestational periods (109 ± 2, 120 ± 5 and 139 ± 2 days gestation; term is 150 days gestation). Using this protocol, we established reference metabolic levels using recommended ¹H-MRS data analysis techniques. Objectives included adapting protocols for reliable metabolite detection despite the challenges of fetal ¹H-MRS, applying the protocol at multiple gestational ages, using recommended processing techniques and comparing two analysis software packages for robustness. MRI scans were performed on a 3 T Siemens clinical system (Magnetom Skyra, Siemens Healthineers, Germany) while the ewe was ventilated. ¹H-MRS was performed using a point-resolved spectroscopy sequence at an echo time of 135 ms and a voxel size of 15 mm × 15 mm × 15 mm, gated to maternal respiration. We showed that TARQUIN metabolite level estimates had a significantly higher Cramér-Rao lower bound uncertainty compared with LCModel in N-acetyl aspartate and choline, with differences in fit quality, while also underestimating metabolites. LCModel was then used to determine reference levels for N-acetyl aspartate, choline, creatine and lactate across three gestational time points in late gestation. Our findings demonstrate the feasibility and reproducibility of non-invasive fetal sheep brain ¹H-MRS. The study highlights the importance of objective criteria for accurate data interpretation, providing insights into fetal brain development and potential non-invasive applications for earlier detection of poor neurodevelopmental outcomes.