Quantification of the postprandial intragastric fat volume and distribution during gastric fat emptying is important in food science in order to asses the in vivo performance of engineered lipid emulsions. The IDEAL method has been recently validated for this purpose [1]. The presence of fat in the small bowel stimulates gastric acid secretion, which is required for efficient intragastric emulsification and subsequent absorption of the ingested fat. The postprandial intragastric accumulation and mixing of gastric secretion has been previously quantified by rapid T₁ mapping [2]. A simultaneous assessment of both intragastric secretion and fat has hitherto been hampered by the bi-exponential relaxation behavior of fat emulsions. In order to allow the simultaneous measurement of intragastric secretion and fat, this work aims to introduce a fat correction for the previously established T₁ mapping method.

Hierarchical IDEAL [3] was implemented incorporating a spectral fat model of rapeseed oil with the relative amplitude $$$\alpha_p$$$ and chemical shift $$$f_p$$$ of each peak $$$p$$$. Multi-echo data were acquired with a 6-point gradient echo sequence. Using IDEAL, the complex field map $$$\hat{\psi}$$$ was estimated and the signal separated into a water and fat signal, $$$W$$$ and $$$F$$$, respectively, from which quantitative FF maps were obtained. T₁ measurements were based on the dual flip angle (DFA) acquisition, where two measurements at different flip angles $$$\alpha_{1,2}$$$ are performed. The bi-exponential signal component was removed by substracting $$$F$$$ from the DFA signal, according to

$$S_w = S - F \frac{\sin\beta_{1,2} (1-E_{1f})}{1-E_{1f}\,\cos\beta_{1,2}} \sum_p \alpha_p \operatorname{e}^{i2\pi f_p}\, \operatorname{e}^{2\pi\hat{\psi}}\,,$$

where $$$S_w$$$ is the water signal, $$$S$$$ the overall DFA signal and $$$E_{1f}$$$=200 ms the predefined T₁ decay of fat. A T₁ fitting procedure for the remaining mono-exponential signal was then applied according to [4]. The result was corrected for nonrectangular slice profiles and for transmit field inhomogeneity using the rapid B₁ mapping technique DREAM [5]. The final T₁ maps were interrelated to secretion concentration by an in vitro calibration curve (Fig. 1). They also allowed for correcting the T₁ bias in the FF maps.

Meal: An acid stable emulsion prepared with water, rapeseed oil, polysorbate 80 and contrast agent Gd-DOTA (Laboratoire Guerbet, France) was investigated (200 mL, 20 wt% fat, droplet size 0.6 μm).

In vitro experiments: T₁ and FF maps were acquired using different diluted samples of the emulsion. Results were validated with spectroscopic inversion recovery measurements as a reference and compared to T₁ mapping without fat correction.

In vivo experiments: Ten healthy volunteers were imaged in right decubitus position on a 1.5 T MR system (Philips Healthcare, Best, The Netherlands). After drinking the emulsion, IDEAL, DFA and DREAM scans were performed covering the complete stomach every 20-30 minutes for 3 hours in total. IDEAL scan parameters were: TR 10 ms, TE/dTE 1.25/1.54 ms, flip angle 10°, FOV 360x292x156 mm³, voxel size 2.8x2.8 mm², slice thickness 7.5 mm, scan duration 1.4 s/slice. DFA scan parameters were: TR 9 ms, TE 2.1 ms, flip angles 2°/20°, FOV 360x292x147 mm³, voxel size 2.8x2.8mm², slice thickness 15 mm, scan duration 2.0 s/slice. DREAM scan parameters were: FID first, TE 1.3/1.7 ms, Ts 3.0 ms, STEAM flip angle 60°, imaging flip angle 10°, TFE factor 52, FOV 360x292x147 mm³, voxel size 5.6x5.6 mm², slice thickness 15 mm, scan duration 1 s/slice. The resulting images were analyzed by semi-automatic segmentation of gastric content to assess the intragastric dynamics of secretion and fat.

Fig. 2 shows the agreement of in vitro T₁ values with the reference. 3D visualizations of intragastric FF and secretion maps acquired at different time points are depicted in Fig. 3a. Fig. 3b displays the projection of the secretion concentrations along the direction of gravity over time. The initial accumulation of secretion in the distal stomach, continuous dilution and mixing of the emulsion by secretion and subsequent accumulation of secretion as a layer are clearly visible. In contrast to a consistent linear fat emptying, intragastric secretion showed considerably inter-subject variability, as seen in Fig. 4.Combining rapid T₁ mapping with IDEAL enabled the simultaneous quantification of the intragastric distribution and temporal development of gastric secretion and fat in emulsions revealing the interaction of the two components by dilution and mixing. In vitro validation measurements highlighted the need to correct for the bi-exponential relaxation behavior of fat, especially at higher fat fractions. Simultaneous MR imaging of intragastric secretion and fat seems a promising tool to non-invasively assess the emulsification and emptying of ingested fat.