Dr. Agnieszka Smolinska - The Use of Volatile Biomarkers in Monitoring of Health and Disease (Watch Now)
Dr Agnieszka Smolinska - Department Pharmacology & Toxicology, Maastricht University
Dr. Smolinska studied chemistry at the Faculty of Mathematic, Physics and Chemistry of the Silesian University in Katowice, Poland. She obtained her PhD in 2012 at Radboud University in Nijmegen, the Netherlands, where she was working in the metabolomics and proteomics biomarker discovery research field. In particular, she carried out several studies in neurological disorders using Nuclear Magnetic Resonance and Mass Spectrometry in combination with sophisticated machine learning and biostatistic tools. In 2012, directly after her PhD, she moved as postdoctoral researcher to Pharmacology and Toxicology department in Maastricht, the Netherlands. As the researcher at Maastricht University she extended her knowledge towards metabolomics of exhaled air in clinical applications of various disease and its relation to gut microbiome and metabolism. In 2015 she received Niels Stensen Fellowship from Dutch organization to conduct her research at Dartmouth College in Hanover, USA for period of one year. Since 2017 as appointed assistant professor, she has been working on her personal grant focusing on the early detection of rare liver disease, Primary Sclerosing Cholangitis.
She is in the board of Dutch chemometrics society (role: treasurer). She is teaching and co-coordinating in Bachelor and Master educational programs in Molecular Life Sciences and supervised four PhD students. She published 40 papers in peer-reviewed journals (H-index is 16). Her main areas of research interests are biomarker discovery, metabolomics, diagnostic tool and advance machine learning and biostatistics in clinical applications.
Metabolomics of exhaled air (breathomics) is an emerging field that mainly focuses on the detection, identification, and quantification of volatile organic compounds (VOCs) in human breath, and its use to monitor a range of medical complications. A majority of breathomics studies are aiming at finding changing patterns of VOCs associated with abnormal metabolic and inflammatory processes occurring in the human body. Nowadays in its effort to diagnose disease and follow health status, breathomics utilizes the latest technological innovations including sophisticated analytical instrumentations and learning algorithm softwares. In the past, we established a Gas Chromatography-Mass Spectrometry (GC-MS) based platform to cover a large range of VOCs (volatome) in exhaled breath, combined with machine learning to extract the maximal information from the wealth of GC-MS data. In our multivariate statistical approach, we select first the most significant VOCs and subsequently use them to construct a final classification model. In the statistical approach, validation is twofold i.e. bootstrapping or cross-validation and using an independent internal validation set. If available, the findings are validated again using a new set of samples obtained at a different time point than the discovery samples.
As first example, we showed previously that VOCs could be indicative of pulmonary malfunctions, as for instance in groups of patients with COPD or asthma. Asthma is an inflammatory condition and we found that oxidative stress related VOCs are able to classify correctly 80% of patients from healthy or wheezing individuals. A second example is the feasibility of exhaled VOCs to monitor different stages of disease in Inflammatory Bowel Disease (IBD) patients. We followed 191 Crohn’s disease (CD) patients into a prospective, diagnostic study to find VOCs indicative of disease activity; a limited set of compounds predicted disease severity with 81% of correct classification. In addition, using the same set of CD data we found several VOCs in exhaled breath that related to the gut microbiome. A second illustration of monitoring the gut is a dietary intervention study with healthy, male volunteers consuming two different infant formulas milk. Their exhaled breath was sampled six times over a period of 4 hours and the number of VOCs changed significantly between the two formulas during this time window. The statistically significant compounds were related to specific differences in the digestion, absorption of lipids and proteins and absorption of the products in the gut. A final example is the application to monitor chorioamnionitis, known as intra-amniotic infection, in sheep. In this study, exhaled breath was collected before and after an installation with Ureaplasma parvum, every day up to six days. The exhaled breath profile from infected sheep changed significantly after five days and these findings were successfully validated in an independent set of animals.
To conclude, breath analysis as non-invasive method holds great promise to be of importance in monitoring purposes and clinical practices. Non-invasive monitoring technologies have the potential to revolutionize modern medicine moving from a reactive treatment mode into a preventive and personalized mode facilitated by point-of-care and home-based diagnostic tools.