Spectral biomarkers of antiplatelet therapy effectiveness revealed by SERS

The paper performs results of the use of machine learning methods to differentiate SERS spectra in patients with and without cardiovascular pathologies. Approaches were applied to processing spectral data arrays consisting of 1266 spectra for various groups of patients: healthy patients, patients with pathology of cardiovascular diseases, healthy patients receiving therapy, and patients with pathology of cardiovascular diseases receiving therapy. The applicability of the random forest algorithm for classification problems were shown. Potential spectral biomarkers of differences between the groups of patients on whom these algorithms were tested were identified. The achieved classification accuracy using the random forest spectra algorithm for the groups of healthy patients without therapy and patients with cardiovascular pathology without therapy was 83.4%. When classifying the presence of therapy in healthy patients (control), the accuracy was 76.26%; in patients with cardiovascular pathology, the accuracy was 70%.