We define breathomics as the metabolomics study of exhaled air. It is a strongly emerging metabolomics research field that mainly focuses on health-related volatile organic compounds (VOCs). Since the amount of these compounds varies with health status, breathomics holds great promise to deliver non-invasive diagnostic tools. Thus, the main aim of breathomics is to find patterns of VOCs related to abnormal (for instance inflammatory) metabolic processes occurring in the human body. Recently, analytical methods for measuring VOCs in exhaled air with high resolution and high throughput have been extensively developed. Yet, the application of machine learning methods for fingerprinting VOC profiles in the breathomics is still in its infancy. Therefore, in this paper, we describe the current state of the art in data pre-processing and multivariate analysis of breathomics data. We start with the detailed pre-processing pipelines for breathomics data obtained from gas-chromatography mass spectrometry and an ion-mobility spectrometer coupled to multi-capillary columns. The outcome of data pre-processing is a matrix containing the relative abundances of a set of VOCs for a group of patients under different conditions (e.g. disease stage, treatment). Independently of the utilized analytical method, the most important question, 'which VOCs are discriminatory?', remains the same. Answers can be given by several modern machine learning techniques (multivariate statistics) and, therefore, are the focus of this paper. We demonstrate the advantages as well the drawbacks of such techniques. We aim to help the community to understand how to profit from a particular method. In parallel, we hope to make the community aware of the existing data fusion methods, as yet unresearched in breathomics.

Ventilator-associated pneumonia (VAP) is a nosocomial infection occurring in the intensive care unit (ICU). The diagnostic standard is based on clinical criteria and bronchoalveolar lavage (BAL). Exhaled breath analysis is a promising non-invasive method for rapid diagnosis of diseases and contains volatile organic compounds (VOCs) that can differentiate diseased from healthy individuals. The aim of this study was to determine whether analysis of VOCs in exhaled breath can be used as a non-invasive monitoring tool for VAP. One hundred critically ill patients with clinical suspicion of VAP underwent BAL. Before BAL, exhaled air samples were collected and analysed by gas chromatography time-of-flight mass spectrometry (GC-tof-MS). The clinical suspicion of VAP was confirmed by BAL diagnostic criteria in 32 patients [VAP(+)] and rejected in 68 patients [VAP(−)]. Multivariate statistical comparison of VOC profiles between VAP(+) and VAP(−) revealed a subset of 12 VOCs that correctly discriminated between those two patient groups with a sensitivity and specificity of 75.8% ± 13.5% and 73.0% ± 11.8%, respectively. These results suggest that detection of VAP in ICU patients is possible by examining exhaled breath, enabling a simple, safe and non-invasive approach that could diminish diagnostic burden of VAP.

The identification of specific volatile organic compounds (VOCs) produced by microorganisms may assist in developing a fast and accurate methodology for the determination of pulmonary bacterial infections in exhaled air. As a first step, pulmonary bacteria were cultured and their headspace analyzed for the total amount of excreted VOCs to select those compounds which are exclusively associated with specific microorganisms. Development of a rapid, noninvasive methodology for identification of bacterial species may improve diagnostics and antibiotic therapy, ultimately leading to controlling the antibiotic resistance problem. Two hundred bacterial headspace samples from four different microorganisms (Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus and Klebsiella pneumoniae) were analyzed by gas chromatography-mass spectrometry to detect a wide array of VOCs. Statistical analysis of these volatiles enabled the characterization of specific VOC profiles indicative for each microorganism. Differences in VOC abundance between the bacterial types were determined using ANalysis of VAriance-principal component analysis (ANOVA-PCA). These differences were visualized with PCA. Cross validation was applied to validate the results. We identified a large number of different compounds in the various headspaces, thus demonstrating a highly significant difference in VOC occurrence of bacterial cultures compared to the medium and between the cultures themselves. Additionally, a separation between a methicillin-resistant and a methicillin-sensitive isolate of S. aureus could be made due to significant differences between compounds. ANOVA-PCA analysis showed that 25 VOCs were differently profiled across the various microorganisms, whereas a PCA score plot enabled the visualization of these clear differences between the bacterial types. We demonstrated that identification of the studied microorganisms, including an antibiotic susceptible and resistant S. aureus substrain, is possible based on a selected number of compounds measured in the headspace of these cultures. These in vitro results may translate into a breath analysis approach that has the potential to be used as a diagnostic tool in medical microbiology.

Given the impact of health literacy (HL) on patients' outcomes, limited health literacy is a major barrier to improve cancer care globally. HL refers to the degree in which an individual is able to acquire, process, and comprehend information in a way to be actively involved in their health decisions. Previous research found that almost half of the population in developed countries have difficulties in understanding health-related information. With the gradual shift toward the shared decision making process and digital transformation in oncology, the need for addressing low HL issues is crucial. Decision making in oncology is often accompanied by considerable consequences on patients' lives, which requires patients to understand complex information and be able to compare treatment methods by considering their own values. How health information is perceived by patients is influenced by various factors including patients' characteristics and the way information is presented to patients. Currently, identifying patients with low HL and simple data visualizations are the best practice to help patients and clinicians in dealing with limited health literacy. Furthermore, using eHealth, as well as involving HL mediators, supports patients to make sense of complex information.

Shared decision making (SDM) and patient-centered care require patients to actively participate in the decision-making process. Yet with the increasing number and complexity of cancer treatment options, it can be a challenge for patients to evaluate clinical information and make risk–benefit trade-offs to choose the most appropriate treatment. Clinicians face time constraints and communication challenges, which can further hamper the SDM process. In this article, we review patient decision aids (PDAs) as a means of supporting SDM by presenting clinical information and risk data to patients in a format that is accessible and easy to understand. We outline the benefits and limitations of PDAs as well as the challenges in their development, such as a lengthy and complex development process and implementation obstacles. Lastly, we discuss future trends and how change on multiple levels—PDA developers, clinicians, hospital administrators, and health care insurers—can support the use of PDAs and consequently SDM. Through this multipronged approach, patients can be empowered to take an active role in their health and choose treatments that are in line with their values.

Radiomics are quantitative features extracted from medical images. Many radiomic features depend not only on tumor properties, but also on non-tumor related factors such as scanner signal-to-noise ratio (SNR), reconstruction kernel and other image acquisition settings. This causes undesirable value variations in the features and reduces the performance of prediction models. In this paper, we investigate whether we can use phantom measurements to characterize and correct for the scanner SNR dependence. We used a phantom with 17 regions of interest (ROI) to investigate the influence of different SNR values. CT scans were acquired with 9 different exposure settings. We developed an additive correction model to reduce scanner SNR influence. Sixty-two of 92 radiomic features showed high variance due to the scanner SNR. Of these 62 features, 47 showed at least a factor 2 significant standard deviation reduction by using the additive correction model. We assessed the clinical relevance of radiomics instability by using a 221 NSCLC patient cohort measured with the same scanner. Phantom measurements show that roughly two third of the radiomic features depend on the exposure setting of the scanner. The dependence can be modeled and corrected significantly reducing the variation in feature values with at least a factor of 2. More complex models will likely increase the correctability. Scanner SNR correction will result in more reliable radiomics predictions in NSCLC.