Palmisani J.1*, Di Gilio A.1, Pizzillo V.1, Tosoni E.2, de Gennaro G.1, Perbellini L.3, Santo A.2
1 Department of Biosciences, Biotechnologies and Environment, University of Bari Aldo Moro, 70125 Bari, Italy
2 Lung Unit, P. Pederzoli Hospital, 37019 Peschiera del Garda, Verona, Italy
3 University of Verona, 37129 Verona, Italy
* presenting author
RATIONALE OF THE STUDY: Lung cancer (LC) is one of the most aggressive tumors and the leading cause of cancer-related death worldwide. The diagnosis of the disease generally occurs at an advanced stage, making radical surgical treatment possible in only less than 20% of cases. Thus, there is an ever-increasing need to equip National Health Systems with a reliable diagnostic tool alternative to traditional diagnostic exams addressed to large-scale screening programs. Chemical characterization of Volatile Organic Compounds (VOCs) in human breath and the resulting identification of a disease-related biomarkers has been recognized as a promising approach for the early detection and follow-up of oncologic diseases such as lung cancer. A prospective-observational study based on the application of complementary methodological approaches and analytical techniques was carried out with the main purpose of identifying a VOCs pattern in human breath as biomarkers of LC. A comparative assessment of the applied methodologies for end-tidal breath sampling (Mistral vs BioVOC) and VOCs chemical characterization (TD-GC/MS vs IMR-MS) was herein performed, highlighting limits and potentialities.
METHODOLOGY AND RESULTS: An overall number of 110 individuals were recruited at the Lung Unit of ‘P. Pederzoli’ Hospital in Verona (Italy): 55 patients affected by LC (average age: 69years), 55 healthy controls (average age: 63 years). The enrollment of volunteers in the clinic trial fulfilled specific criteria, after approval by the Ethical Committee (Prot.n. 45355). The sampling methodology applied in the present study was based on two different approaches: a) end-tidal breath collection directly onto two-beds adsorbent cartridges (Biomonitoring steel tubes, Markes International) by means of the automated sampler Mistral (Predict srl, Italy); b) end-tidal breath collection by means of BioVOC-2 sampler (Markes International). Breath samples collected with approach a) were analysed by thermal desorption (Unity Ultra-xr Markes) and Gas Chromatography/Mass Spectrometry (GC Agilent 7890/MS Agilent 5975) at the University of Bari laboratory whilst breath samples collected by approach b) were analysed by Ion Molecular Reaction-Mass Spectrometry (IMR-MS) at the University of Verona laboratory. Ambient air samples (AA) were simultaneously collected at each sampling session with both the methodological approaches and analysed. Experimental data were statistically processed by non-parametric Wilcoxon signed rank test (software R version 3.5.1) in order to identify the most weighting variables in the discrimination between LC and HC breath samples. Principal Components Analysis (PCA) and Linear Discriminant Analysis (LDA) were also applied to the dataset to validate breath analysis-based methodology in the discrimination among LC and HC subjects. A VOCs pattern resulting from TD-GC/MS and IMR-MS analysis and able for discrimination between LC and HCsubjects was identified. Experimental results resulted to be in line with the existing scientific literature. For each VOC identified as a potential biomarker for LC a metabolic pathway was also speculated. A promising predictive model based on selected variables (Wilcoxon signed rank test p-values lower than 0.05) was developed and the leave-one-out cross-validation approach applied to the dataset provided a prediction accuracy equal to 99% (ROC AUC: 0.99).
