Development of a realistic in vitro digestion model (RGM) coupled UV-VIS-SWNIR fibre optics spectroscopy

Abstract

Background: The development of realistic gastric models unlocked the possibility of studying important digestion phenomena occurring during the digestion of food (e.g., retropulsion). Understanding the dynamics of food digestion in real-time, without sample manipulation, is still a challenge, but brings a huge potential in providing important insights regarding the dynamic process of food digestion (e.g., real time nutrient release kinetics) This study presents a realistic 3D printed in vitro gastric model coupled with ultraviolet-visibleshort- wave-near-infrared (UV-VIS-SWNIR) spectroscope that can be used for real time quantification of nutrients/bioactive compounds. Methods: The INFOGEST semi-dynamic in vitro protocol was used to simulate the digestion of rice (model food). The spectroscope was calibrated for glucose analysis, and the spectra were pre-processed and both chemometric and machine learning techniques were used for glucose quantification using the correlation coefficient as assessment metric. Results: The machine learning algorithms showed to be more accurate at predicting glucose release during the in vitro gastric digestion. Conclusions: The gastric compartment development techniques provide the opportunity to develop a potential standard dynamic in vitro gastric model. Furthermore, it was possible to accurately measure and quantify glucose release during the in vitro digestion process, in real time, using UV-VIS-SWNIR fibre optics spectroscopic.