?-Muscle Analysis Matrix – ChiMAlMa

An in-depth knowledge of the microscopic fine ultrastructure of muscles is imperative for the accurate diagnosis of various diseases characterized by muscle dysfunction. At present, a biopsy is the only reliable clinical method that offers a conclusive diagnosis of the pathology. Duchenne dystrophy, inclusion body myositis or peripheral neuropathies i.e. Charcot-Marie-Tooth disease are few examples of muscle pathologies that require a muscle biopsy. But, a muscle biopsy is an invasive procedure that can be distressing for both the muscle tissue and the patient.

Optical solutions offer label-free information compatible with in vivo clinical imaging. But the current solutions probe a limited
number of biomedical substances of interest, not enough for a reliable diagnostic of muscle pathologies.

The ?-MAlMa project introduces an innovative instrumental approach for characterizing biomedical targets based on their contrast in non-linearities characterized by four nonlinear parameters. A supercontinuum device serves as laser excitation, allowing the selective stimulation of each them. The emitted spectra are recorded across five spectral zones, constituting the ?-matrix of the sample under test. Hyperspectral data processing is undertaken to identify distinguishing features within this ?-matrix (15 spectra for each image pixel) utilizing first artificial intelligence (AI) and then chemometric methods.

The solution of AI aims to identify the presence of discriminating biomarkers using a training database. The chemometrics solution aims to identify the discriminating spectral signatures from one or two ?ij parameters. This step plays a significant role for the test of the use of an optical fibre for the delivery of excitation and emission beams.

This innovative approach would lead to a new medical device and practice that could improve the diagnostic reliability, orientation of therapy, patients care and healthy life expectancy.