Diagnostic prediction of renal failure from blood serum analysis by FTIR spectrometry and chemometrics

• Fast and reliable diagnostic method for renal failure.
• Avoiding the consumption of chemical reagents in medical testing.
• Minimally-invasive medical approach.

A new diagnostic approach based on Attenuated Total Reflectance–Fourier Transform Infrared (ATR–FTIR) spectrometry and classification algorithm has been introduced which provides a rapid, reliable, and easy way to perform blood test for the diagnosis of renal failure. Blood serum samples from 35 renal failure patients and 40 healthy persons were analyzed by ATR–FTIR spectrometry. The resulting data was processed by Quadratic Discriminant Analysis (QDA) and QDA combined with simple filtered method. Spectroscopic studies were performed in 900–2000 cm−1 spectral region with 3.85 cm−1 data space. Results showed 93.33% and 100% of accuracy for QDA and filter-QDA models, respectively. In the first step, 30 samples were applied to construct the model. In order to modify the capability of QDA in prediction of test samples, filter-based feature selection methods were applied. It was found that the filtered spectra coupled with QDA could correctly predict the test samples in most of the cases.

Figure optionsDownload as PowerPoint slide