The aim of this work was to assess chemical shift-encoded MRI (CSE-MRI) method to quantify content and composition changes of fat storage in healthy volunteers during a 31 days overfeeding protocol while comparing CSE-MRI results with DEXA, MRS and Gas chromatography measurements. A total of 21 volunteers underwent a NMR protocol at 3T with an axial 3D CSE-MRI on abdominal region and spectroscopy acquired on subcutaneous adipose tissue, visceral adipose tissue and liver. The NMR protocol was used to analyze the volume and the fatty acid composition of abdominal adipose tissues, and the fat content in the liver.
A total of 21 non-obese healthy male volunteers underwent 31 days of high-sucrose and high-fat overfeeding and two NMR examinations (before and after overfeeding). The overfeeding protocol consisted of eating the equivalent of +50% of their daily energy requirements while maintaining their usual eating and sport activity habits. This protocol was registered on the clinicaltrials.gov site (NCT02145780). Only results from subjects with a placebo complement are reported.
All volunteers underwent an imaging and spectroscopy protocol on a 3T Ingenia Philips system composed of an axial 3D CSE-MRI on abdominal region encompassing lumbar L1 to L5 and the liver (parameters table 1) and a set of three STEAM sequences acquired on SAT, VAT and liver (parameters in the caption of the Figure 1). This NMR protocol was used to assess the proton density fat fraction (PDFF) in the liver and the FA composition2,5 in SAT and VAT and volumes. Gas chromatography-mass spectrometry analyses were performed on 13 samples (seven before and six after the overfeeding intervention) of SAT to validate the FA composition. The FA composition was characterized by two parameters (the number of double bound: ndb and the number of methylene-interrupted double bound: nmidb)6. Indexes of the relative proportion of polyunsaturated FA (PUFAindx), monounsaturated FA (MUFAindx) and saturated FA (SFAindx) were derived from ndb and nmidb with the following relations7:
$$PUFA_{indx}=\frac{nmidb}{3}*100$$
$$MUFA_{indx}=\frac{(ndb-2*nmidb)}{3}*100$$
$$SFA_{indx}=100-PUFA_{indx}-MUFA_{indx}$$
Volume of adipose tissues was automatically segmented on CSE-MRI between L2 to L4 lumbar vertebrae and compared to the DEXA measurement in the corresponding region.
An excellent correlation between CSE-MRI measurement of total adipose tissue TAT (SAT + VAT) and DEXA measurement of the android region (r = 0.98, P < 0.001, n = 18) was found. SAT and VAT volumes significantly increased between the two examinations (paired t-test: +91 cm3, p < 0.001; +129 cm3, p < 0.001). The ratio VAT/SAT also significantly increased after overfeeding (+0.06, p = 0.020).
The PDFF significantly increased in the liver (MRS: +1.74, paired t-test, p = 0.002; CSE-MRI: +1.35 paired t-test, p = 0.002) after the overfeeding. Assessment of PDFF variations measured by MRS or CSE-MRI were found to be similar and non-significantly different (p = 0.85, figure 2).
Regarding CSE-MRI results from the whole segmented volume versus the localized MRS results, the comparison of FA composition of SAT and VAT appeared to be substantially equivalent in average (figure 3). In both cases, paired t-tests showed significant difference between SAT and VAT FA composition (p<0.001 for CSE-MRI measurements and p=0.001 for MRS measurements). Good correlations were found between GC-MS and CSE-MRI measurements (for ndb: r = 0.84, p = 0.001; for nmidb: r = 0.66, p = 0.020) and between GC-MS and MRS measurements for ndb: r = 0.83, p < 0.001; for nmidb: r = 0.74, p = 0.006). The CSE-MRI method permitted to detect a significant decrease of PUFAindx (p = 0.012) and an increase of MUFAindx (p < 0.001) in VAT (figure 4).
A significant increase in fat distribution in VAT, SAT, and liver were established by DEXA, MRS, and CSE-MRI measurements after an experimental overfeeding. FA composition changes (more precisely the MUFA proportion) of the adipose tissue were also observed by CSE-MRI measurements. The use of a single 3D CSE-MRI sequence associated with a dedicated post-processing method is a suitable protocol to follow several parameters linked to adipose tissue changes. This NMR protocol could be used in future work to noninvasively evaluate the impact of nutrition on adipose tissues and liver.
More details could be found in the journal article 8.