P764 - OVERCOMING STRUCTURAL AMBIGUITY IN FOOD IMAGING: PHASOR-BASED MRI FOR QUANTITATIVE NUTRITIONAL CHARACTERIZATION

Linked sessions

P764

OVERCOMING STRUCTURAL AMBIGUITY IN FOOD IMAGING: PHASOR-BASED MRI FOR QUANTITATIVE NUTRITIONAL CHARACTERIZATION

C. Serantoni1,2,*, M. M. D. Giulio2, A. Riente2, D. Hatem2, T. Marchetti1, M. D. Spirito1,2, G. Maulucci1,2 on behalf of Metabolic Intelligence Lab, Rome, Italy

1Fondazione Policlinico Universitario Agostino Gemelli IRCCS, 2Università Cattolica del Sacro Cuore, Rome, Italy

 

Rationale: MRI is a non-invasive, label-free method to analyze the internal structure and composition of food substrates [1]. However, conventional segmentation techniques often struggle to differentiate key nutritional components—such as water, lipids, and proteins—due to overlapping signal features. This study applies phasor transformation to multiparametric MRI data to improve the segmentation and quantification of complex food matrices.

Methods: Forty food samples representing clinical nutrition products were scanned using multi-echo T1- and T2-weighted MRI. Voxel-wise signal decay curves were mapped into phasor space, enabling combined analysis of relaxation properties. Clustering in phasor space identified signal populations corresponding to nutritional components, which were projected back into the spatial domain for segmentation.

Results: The phasor-based method achieved 93% segmentation accuracy, outperforming traditional intensity-based approaches (76%). Combined T1-T2 analysis improved discrimination of overlapping signals. Lipid and aqueous phases were clearly separated, and protein-rich regions showed 22% higher heterogeneity. The method also enabled high-resolution 3D reconstruction of nutritional architecture (0.5 mm³ resolution).

Conclusion: Phasor analysis of T1-T2 MRI data provides a robust tool for non-invasive segmentation and characterization of food substrates. It offers applications in quality control, formulation development, and nutritional profiling, and may support future integration into real-time or in vivo digestive imaging.

References: [1] Capozzi, F., Laghi, L., Peter S., B. (2015). Magnetic Resonance in Food Science. Defining Food by Magnetic Resonance.. Cambridge : Royal Society of Chemistry.

Disclosure of Interest: None declared