Perspective: Breathomics in medicine

The present procedures for assessment of obstructive airway diseases include spirometry, the response to inhaled bronchodilators and/or challenge with bronchoconstrictive stimuli. As an alternative empirical approach, omics techniques allow analysis of the overall molecular pattern of complex samples such as exhaled breath in a single measurement. This perspective describes the recent advances, current status and future developments of breathomics using gas chromatography mass spectrometry and electronic noses both from a medical and technical point of view. Examples of medical advantages of breathomics are the noninvasiveness of the test, the possibility to store breath samples, the availability of portable devices and the applicability of the technique in many diseases. Disadvantages are the inability to distinguish between endogenous and exogenous compounds and the influence of many factors, such as medication and diet, on the composition and quality of air samples. Technical advantages of eNose measurements are the limited costs, reversibility and short response time of sensors, and the use of pattern recognition algorithms. The latter enables a single device to be used in different diseases by simply changing the diagnostic algorithm. Technical drawbacks are the necessity of training the device for each application, incompatibility between devices and limited stability of sensors. This perspective focuses on advances made in the field of pulmonology, but recommendations for application can be easily translated to other fields of medicine. Steps towards introduction of metabolomic breath tests into daily clinical practice are currently undertaken and require joined forces between medical and technical task forces. Perspective: breathomics in medicine 19 Ch ap te r 2 INTRODUCTION: OMICS IN PULMONARY MEDICINE The present procedures for the clinical assessment of airway diseases are rather complex. These include spirometry, the response to inhaled bronchodilators and/or challenge with bronchoconstrictive stimuli [1,2]. These tests have internationally been standardized, and are generally considered to be reliable. However, they are cumbersome, time consuming and not widely applicable. This has seriously limited the implementation of these techniques at all necessary levels of medical care. The strength of current diagnostic procedures is that they are measuring a core component of both diseases, namely, (variable) airways obstruction. Hence, the tests are derived from pathophysiological reasoning, similar to many other diagnostic tests in medicine. This fits into a tradition in which diagnostic tests are developed based on understanding of disease mechanisms. An increasingly successful alternative of pathophysiological reasoning in diagnosis, monitoring and pathogenetic research is the empirical approach, in which the tests are selected based on probabilistic evidence only. Such procedures are fully evidence-based, whilst being essentially hypothesis-free. The best examples for this are the recently developed high-throughput methods, the so-called ‘omics’ techniques, which allow analysis of the overall molecular pattern of complex biological samples by a single set of measurements [3]. This can for instance be done at the level of RNA (transcriptomics), proteins (proteomics) or metabolites (metabolomics). The information provided by these techniques is not based on detecting single and separate molecular signals, but is exclusively derived from pattern recognition among an array of signals by using powerful bioinformatics. This is based on cluster analysis and learning algorithms (artificial intelligence), which can be trained and applied for classification and identification of disease. The omics techniques have the potential to be applied in almost every field of medicine. This can be done by an integrative systems medicine approach [4], providing ‘fingerprints’ of disease. In other disease areas, such as cardiology and neurology, this has already led to improved prediction of disease outcome [5,6]. The procedures to evaluate their diagnostic accuracy are essentially the same as for any other diagnostic test. After determining their test accuracy in diagnostic research (including patients with the disease, healthy subjects and subjects suspected for the disease) such tests can be implemented in the clinical or epidemiological setting [7-10]. METABOLOMICS IN ExHALED AIR Can omics techniques be used in the diagnosis and monitoring of asthma and COPD? Application of proteomics in blood [11] and sputum [12] has provided promising results. As a non-invasive alternative, metabolomics in exhaled breath (breathomics) is potentially suitable for molecular diagnosis in asthma and COPD. volatile organic compounds Already back in 1971 Pauling et al demonstrated that exhaled breath contains a complex mixture of several hundreds of volatile organic compounds (VOCs) [13]. The VOCs in human