RESEARCH LINE
Metabolic-driven Medical Physics
Medical physics is where a physical measurement becomes a clinical decision. In this research line we make clinical and preclinical imaging quantitative — positron emission tomography and multispectral optoacoustic tomography — and turn the raw signal into parameters a physician can act on.
01 · NUCLEAR MEDICINE
Quantitative PET and machine learning
Dynamic PET carries more information than a static uptake map: the time course of tracer accumulation in each voxel reflects transport and metabolism.
Together with nuclear medicine, we showed that machine-learning models trained on dynamic PET data can detect and classify tumour histotypes automatically — from the kinetics of each voxel, rather than from a single frame.
Bianchetti, G.; Taralli, S.; Vaccaro, M.; Indovina, L.; Mattoli, M.V.; Capotosti, A.; Scolozzi, V.; Calcagni, M.L.; Giordano, A.; De Spirito, M.; Maulucci, G. Automated detection and classification of tumor histotypes on dynamic PET imaging data through machine-learning driven voxel classification. Computers in Biology and Medicine 2022, 145, 105423. (DOI: 10.1016/j.compbiomed.2022.105423)

Dynamic PET: kinetics in every voxel.

MSOT: real-time spectral unmixing.
02 · TOMOGRAPHIC IMAGING
Multispectral optoacoustic tomography
MSOT combines the contrast of optical absorption with the penetration depth of ultrasound. Its quantitative use depends on separating overlapping spectral signatures: oxygenated and deoxygenated haemoglobin, lipids, contrast agents.
We introduced a phasor-based method that performs this unmixing in real time, without assuming in advance how many components are present.
Di Giacinto, F.; Riente, A.; Mignini, I.; Ainora, M.E.; Esposto, G.; Borriello, R.; Zocco, M.A.; Minordi, L.M.; Sala, E.; Scaldaferri, F.; Gasbarrini, A.; De Spirito, M.; Maulucci, G. Advancing multispectral optoacoustic tomography (MSOT): phasor analysis for real-time spectral unmixing. Computers in Biology and Medicine 2025, 195, 110586. (DOI: 10.1016/j.compbiomed.2025.110586)
The complete list of publications behind this research line is on the Publications page.
