Urinary volatile fingerprint based on mass spectrometry for the discrimination of patients with lung cancer and controls

Profile signals of urine samples corresponding to patients with lung cancer and controls were obtained using a non-separative methodology. The method is based on the coupling of a headspace sampler, a programed temperature vaporizer and a mass spectrometer (HS-PTV-MS). With only a centrifugation step as prior sample treatment, the samples were subjected to the headspace generation process and the volatiles generated were introduced into the PTV where they were trapped in the Tenaxo (R) packed liner while the solvent was purged. Finally, the analytes were introduced directly, without separation, into the mass spectrometer which allows obtaining the fingerprint of the analyzed sample. The mass spectrum corresponding to the mass/charge ratios (m/z) ranging between 35 and 120 amu (amu) contains the information related to the composition of the headspace and is used as the analytical signal for the characterization of the samples. Samples of 14 patients with some type of cancer and 24 healthy volunteers were analyzed and the profile signals were subjected to different chemometric techniques, including support vector machines (SVM), linear discriminant analysis (LDA) and partial least squares-discriminant analysis (PLS-DA), with the aim of differentiating the samples of patients with cancer from those of control. Values of 100% were obtained both in sensitivity and specificity in most cases. This methodology has been used previously, as described later, for the analysis of the fingerprint corresponding to saliva samples of patients and controls. However, up to date, the method has not been used in urine samples with the aim of fast discrimination between patients with cancer and controls. The advantages and disadvantages of using urine versus other types of matrices such as saliva are stated. In view of the results obtained in this work, the use of pattern recognition techniques with data corresponding to HS-PTV-MS profile signals is highly suitable as a first screening step to differentiate samples. In addition, it could be applied to a high number of samples in a relatively short period of time due to its high throughput.