Background:Background: Multiple myeloma (MM) is the second most common hematological malignancy of the elderly. The bone marrow is infiltrated by malignant plasma cells. MM may progress into so-called extramedullary disease (EMD) -EMD occurs when a subclone of malignant plasma cells migrates out of the bone marrow and infiltrates soft tissues. Despite recent progress, pathogenesis of EMD still needs to be clarified. Aims:Aims: We focused on the analysis of low molecular weight molecules in peripheral blood of 20 MM and 20 EMD patients using MALDI-TOF mass spectrometry to create a tool for identification of MM and EMD, and discrimination of individual types. Methods:Methods: Matrix-Assisted Laser Desorption/Ionization Time-of Flight Mass Spectrometry (MALDI-TOF MS) has become an indispensable research tool, which is used for analysis of biomolecules and various organic molecules. Artificial Neural Networks (ANN) are components of artificial intelligence inspired by biological neural networks. Using ANN, we can model complex non-linear systems, as previously published. In our previous study, we recorded mass spectra of MM and healthy donor samples. ANN specifically predicted MM samples with high sensitivity, specificity and accuracy. The same approach as applied on MM and EMD. Results:Results: The RStudio was used for statistical analysis, where the data were evaluated using Principal Component Analysis (PCA) and Partial least squares discriminant analysis (PSL-DA). Using MALDI-TOF MS, it was possible to distinguish between samples of MM patients and healthy donors, as well as MM and EMD patients. Informative patterns in mass spectra served as inputs for ANN that specifically distinguished between healthy donors and patients.

No abstract