Measurement of Neonatal Body Composition using Dual-Energy X-ray Absorptiometry (DXA) Magnetic Resonance Imaging and Spectroscopy (MRI/MRS) in Piglets. Part II: Validation of MRI/MRS by Chemical Analysis (CA)† 1516

Purpose: Animal study to evaluate DXA and MRI/MRS as tools to assess body composition (BC) in newborn infants. DXA data are presented in a first abstract. Methods: 21 piglets (age: 1-37d, B: 848-7550g). After measuring total body water (TBW) (0.4 ml 99.8% D₂O/kg BW, correction factor for isotope sequestration 1.04), the animal was sacrificed. BC was assessed within 2h by DXA and MR (1.5T, Signa, GE, Milwaukee). The carcass was homogoenized, CA was done using standard methods. To obtain the fat/water ratio (FWR) by CA, body water was calculated as BW minus weight of lyophilisate plus its residual water content. MRI: contiguous 5-mm T1-weighted scans (TR 340ms, TE 16ms); volume determination by point-counting, pixels were assigned to either subcutaneous fat, visceral fat, bone, vessel, or lean mass. MRS: body coil, 3D 8×16×16 FID-CSI sequence (TR 300ms), correction for shifts of resonance frequencies, fit of water and fat resonances in the sum of the corrected spectra. MRS data acquisition took 20.4min, data analysis 5min. In 4 piglets, saturation factors(SF) of water and fat were measured (0.55 and 1.0) to correct signal intensities accordingly. MRS-FWR were calculated assuming proton densities per molecule of 0.111 and 0.118 for water and trigycerides, resp. Absolute fat content by MRS was obtained from the FWR using in-vivo TBW determination by D₂O. Approvement by the cantonal and federal commissions for animal studies. Results: The alignment of water resonances of all voxels(typical voxel size of 8.0 ± 1.8 ml) improved the signal shape significantly allowing reliable signal fitting. 1H-3D-CSI MRS and CA showed a correlation for the fat/water ratio of r² = 0.984, the slope however was 0.772, the offset was 0.0025. The correlation was slightly higher for the fat content in absolute terms, (r² = 0.988, with a slope of 0.841 and an intercept of 8.5 g). Correlation of body fat as measured by MRI (volume) and CA (weight) was lower (r² = 0.913), MRI overestimated fat volume by a factor of 2.46 with an offset of 69.4 ml. Discussion: 1H-3D-CSI-MRS is a fast, non-invasive method with minimal operator interaction and measures neonatal fat/water ratios surprisingly precisely and accurately. It seems to be a promising tool to assess neonatal body fat provided that TBW is additionally measured.