[Use of artificial intelligence for image reconstruction]. / Einsatz künstlicher Intelligenz für die Bildrekonstruktion.

CLINICAL/METHODOLOGICAL PROBLEM: In the reconstruction of three-dimensional image data, artifacts that interfere with the appraisal often occur as a result of trying to minimize the dose or due to missing data. Used iterative reconstruction methods are time-consuming and have disadvantages. STANDARD RADIOLOGICAL METHODS: These problems are known to occur in computed tomography (CT), cone beam CT, interventional imaging, magnetic resonance imaging (MRI) and nuclear medicine imaging (PET and SPECT). METHODOLOGICAL INNOVATIONS: Using techniques based on the use of artificial intelligence (AI) in data analysis and data supplementation, a number of problems can be solved up to a certain extent. PERFORMANCE: The performance of the methods varies greatly. Since the generated image data usually look very good using the AI-based methods presented here while their results depend strongly on the study design, reliable comparable quantitative statements on the performance are not yet available in broad terms. EVALUATION: In principle, the methods of image reconstruction based on AI algorithms offer many possibilities for improving and optimizing three-dimensional image datasets. However, the validity strongly depends on the design of the respective study in the structure of the individual procedure. It is therefore essential to have a suitable test prior to use in clinical practice. PRACTICAL RECOMMENDATIONS: Before the widespread use of AI-based reconstruction methods can be recommended, it is necessary to establish meaningful test procedures that can characterize the actual performance and applicability in terms of information content and a meaningful study design during the learning phase of the algorithms.

Age-dependent inhalation doses to members of the public from indoor short-lived radon progeny.

The main contribution of radiation dose to the human lungs from natural exposure originates from short-lived radon progeny. In the present work, the inhalation doses from indoor short-lived radon progeny, i.e., (218)Po, (214)Pb, (214)Bi, and (214)Po, to different age groups of members of the public were calculated. In the calculations, the age-dependent systemic biokinetic models of polonium, bismuth, and lead published by the International Commission on Radiological Protection (ICRP) were adopted. In addition, the ICRP human respiratory tract and gastrointestinal tract models were applied to determine the deposition fractions in different regions of the lungs during inhalation and exhalation, and the absorption fractions of radon progeny in the alimentary tract. Based on the calculated contribution of each progeny to equivalent dose and effective dose, the dose conversion factor was estimated, taking into account the unattached fraction of aerosols, attached aerosols in the nucleation, accumulation and coarse modes, and the potential alpha energy concentration fraction in indoor air. It turned out that for each progeny, the equivalent doses to extrathoracic airways and the lungs are greater than those to other organs. The contribution of (214)Po to effective dose is much smaller compared to that of the other short-lived radon progeny and can thus be neglected in the dose assessment. In fact, 90 % of the effective dose from short-lived radon progeny arises from (214)Pb and (214)Bi, while the rest is from (218)Po. The dose conversion factors obtained in the present study are 17 and 18 mSv per working level month (WLM) for adult female and male, respectively. This compares to values ranging from 6 to 20 mSv WLM(-1) calculated by other investigators. The dose coefficients of each radon progeny calculated in the present study can be used to estimate the radiation doses for the population, especially for small children and women, in specific regions of the world exposed to radon progeny by measuring their concentrations, aerosol sizes, and unattached fractions.

Headspace measurements of irradiated in vitro cultured cells using PTR-MS.

A pilot study was performed to evaluate a new concept for a radiation biodosimetry method. Proton transfer reaction-mass spectrometry (PTR-MS) was used to find out whether radiation induces changes in the composition of volatile organic compounds (VOCs) in the headspace of in vitro cultured cells. Two different cell lines, retinal pigment epithelium cells hTERT-RPE1 and lung epithelium cells A-549, were irradiated with gamma radiation at doses of 4 Gy and 8 Gy. For measuring the cell-specific effects, the VOC concentrations in the headspace of flasks containing cells plus medium, as well as of flasks containing pure medium were analyzed for changes before and after irradiation. No significant radiation-induced alterations in VOC concentrations in the headspace could be observed after irradiation.

Discrimination of cancerous and non-cancerous cell lines by headspace-analysis with PTR-MS.

Proton transfer reaction mass spectrometry (PTR-MS) has been used to analyze the volatile organic compounds (VOCs) emitted by in-vitro cultured human cells. For this purpose, two pairs of cancerous and non-cancerous human cell lines were selected:1. lung epithelium cells A-549 and retinal pigment epithelium cells hTERT-RPE1, cultured in different growth media; and 2. squamous lung carcinoma cells EPLC and immortalized human bronchial epithelial cells BEAS2B, cultured in identical growth medium. The VOCs in the headspace of the cell cultures were sampled: 1. online by drawing off the gas directly from the culture flask; and 2. by accumulation of the VOCs in PTFE bags connected to the flask for at least 12 h. The pure media were analyzed in the same way as the corresponding cells in order to provide a reference. Direct comparison of headspace VOCs from flasks with cells plus medium and from flasks with pure medium enabled the characterization of cell-line-specific production or consumption of VOCs. Among all identified VOCs in this respect, the most outstanding compound was m/z = 45 (acetaldehyde) revealing significant consumption by the cancerous cell lines but not by the non-cancerous cells. By applying multivariate statistical analysis using 42 selected marker VOCs, it was possible to clearly separate the cancerous and non-cancerous cell lines from each other.

Only therapeutic ICOS:ICOSL blockade alleviates acute graft versus host disease.

Inducible co-stimulator (ICOS) interaction with its ligand (ICOSL) is involved in several T cell effector functions. While blockade of ICOS:ICOSL interaction in chronic graft versus host disease (GVHD) seems beneficial, results for acute GVHD remain controversial. To further elucidate its role in acute GVHD, C57BL/6 mice were reconstituted with allogeneic spleen cells in the absence or presence of ICOSL-blocking mAb. Mice reconstituted with allogeneic spleen cells experienced severe GVHD and died untreated within 6-9 days after transplantation. Mice treated with an anti-ICOSL mAb starting from day 3 after transplantation gained weight again and survived for at least additional 12 days, although the treatment was already stopped at day 11 after transplantation. In contrast, the anti-ICOSL treatment starting from day 0 did not prevent GVHD. The difference between therapeutic (day 3) and prophylactic (day 0) anti-ICOSL treatment was independent of CD25+CD4+ regulatory T cells since their depletion did not abrogate the therapeutic effect of ICOSL blockade. Microarray analysis revealed IFN-gamma and chemokine up-regulation in spleen cells of prophylactically treated mice, emphasizing kinetic dependence of acute GVHD modulation via blockade of ICOS:ICOSL interaction.

EURAMED's vision on medical radiation protection (research).

While many areas of radiation protection have formed so-called 'platforms' in Europe which provide strategic research agendas for their areas of interest, this did not happen for a long while for medical exposure, which is the application of ionising radiation that causes the greatest man-made exposure, at least in first world countries. Finally, in 2015, a European medical radiation protection strategic research agenda was set up, and a corresponding platform was launched in 2016. This was named 'EURAMED' - the European Alliance for Medical Radiation Protection Research. In its strategic research agenda, EURAMED defined its vision for medical radiation protection and the corresponding research needed. Five major topics were identified, ranging from measurements of medical application-related parameters such as exposures and image quality and radiation biology aspects relevant for medical applications to individual optimisation strategies, to optimal use of techniques and harmonisation of practises, and finally to justification of the use of ionising radiation in medicine, all based on sufficient infrastructures for quality assurance. The ultimate goal is to reduce radiation exposure and risk individually for patients and staff by interdisciplinary research between clinicians, physicists, and engineers. Therefore, it is essential that the results are translated into clinical practice.

Local organ dose conversion coefficients for angiographic examinations of coronary arteries.

New organ dose conversion coefficients for coronary angiographic interventions are presented, as well as dose distributions and resulting maximal local dose conversion coefficients in the relevant organs. For the Monte Carlo based simulations, voxel models of the human anatomy were employed which represent the average Caucasian adult man and woman as defined by the International Commission on Radiological Protection. In the 21 investigated projections, the mean organ dose conversion coefficients vary from a few 0.01 to 2 mGy(Gy cm(2))(-1), depending on the projections. However, especially in portions of the lungs and the active bone marrow, the conversion coefficients can locally amount up to 10 mGy(Gy cm(2))(-1), which is half the average conversion coefficient of the skin at the field entrance. In addition to the dose conversion coefficients, the dependence of the patient dose on the projection has been estimated. It could be shown that the patient doses are highest for left anterior oblique views with strong caudal or cranial orientation. Nevertheless, for a large range of image-intensifier positions no significant dose differences could be found.

[Comparison of four digital and one conventional radiographic image systems for the chest in a patient study with subsequent system optimization]. / Vergleich von vier digitalen und einem konventionellen Röntgenaufnahmesystem für den Thorax mittels einer Patientenstudie mit nachfolgender Systemoptimierung.

PURPOSE: Using a patient study to prove the clinical relevance of a comparison of five different radiographic systems for the chest conducted with an anthropomorphic chest phantom. Depending on the results, it was tested whether the performance of a modern digital system with a transparent imaging plate can be improved by changing the post-processing of the image. METHOD: Chest radiographs of patients were taken with a CsI/aSi-flat panel detector (FDR), transparent imaging plate (tDLR), selenium drum detector (DSR), conventional storage phosphor plate (DLR) and asymmetrical screen-film-system (aFFS), and compared using image criteria scoring (ICS) and visual grading analysis (VGA) for anatomical structures (modified criterions of the EUR 16 260 EN guidelines). After optimizing the post processing, the images of the tDLR-system were evaluated once more in a phantom ROC study and patient VGA study. RESULTS: The flat panel detector-system proved to meet best the anatomical image quality criteria, followed by DSR, tDLR, aFFS and DLR. The modified post processing of the tDLR-images resulted in a significantly better detection of simulated pathological lung-structures, but improved the perceptibility of anatomical structures only slightly. CONCLUSIONS: The results of the patient VGA study and the phantom ROC study are similar and considered valid. The new digital imaging systems with flat panel detector and transparent imaging plate provide the best image quality of the tested radiographic devices for chest imaging, assuming that all system components are attuned and optimized for the type of structure to be detected. Image processing is of primary importance for system optimization.

Assessment and optimisation of the image quality of chest-radiography systems.

A complete evaluation strategy had been developed for thoracic X-ray imaging. It has been validated by investigating five chest-radiography systems, two of these systems after optimising image processing. The systems were a screen-film combination, a selenium drum, a conventional and a transparent imaging plate and a Cs/I-based flat panel detector (the two latter ones have been optimised using different post processing). At first all detectors have been characterised using physical parameters like DQE and MTF. After that all systems have been evaluated by human observer studies using anatomy in clinical images (VGA, ICS) and added pathological structures in thoracic phantom images (ROC). The ranking of the image quality of the systems was nearly the same in all studies. There was a similar assessment of main image quality parameters like spatial resolution, dynamic range and MTF. The modification of image post processing changed the visibility of pathological structures more than the visualisation of the anatomical criteria. The assessment of the clinical image quality has to be done for anatomical structures, and the recognition of pathological structures has to be evaluated.

Average glandular dose conversion coefficients for segmented breast voxel models.

For 8 voxel models of a compressed breast (4-7 cm thickness and two orientations for each thickness) and 14 radiation qualities commonly used in mammography (HVL 0.28-0.50 mm Al), tissue dose conversion coefficients were calculated for a focus-to-film distance of 60 cm using Monte Carlo methods. The voxel models were segmented from a high-resolution (slice thickness of 1 mm) computed tomography data set of an ablated breast specimen fixated while being compressed. The contents of glandular tissues amounted to 2.6%, and were asymmetrically distributed with regard to the midplane of the model. The calculated tissue dose conversion coefficients were compared with the recent literature values. These earlier tissue dose conversion coefficients were also calculated using Monte Carlo methods and breast models of various thickness, but these consist of homogeneous mixtures of glandular and adipose tissues embedded in 5 mm pure adipose tissue both at the entrance and exit sides. The results show that the new glandular tissue dose conversion coefficients agree well with the literature values for those cases where the glandular tissue is predominantly concentrated in the upper part of the model. In the opposite case, they were lower by up to 40%. These findings reveal a basic problem in patient dosimetry for mammography: glandular dose is not only governed by the average breast composition, which could be derived from the breast thickness, but also by the local distribution of glandular tissue within the breast, which is not known.

RADIUS--closing the circle on the assessment of imaging performance.

The RADIUS (Radiological Imaging Unification Strategy) project addresses the assessment of image quality in terms of both physical and clinically relevant measures. The aim is to unify our understanding of both types of measure as well as the numerous underlying factors that play a key role in the assessments of imaging performance. In this way it is expected to provide a solid basis for the improvement in radiological safety management, where not only radiation risks are considered but also diagnostic risks of incorrect clinical outcomes (i.e. false positive/false negative). The project has applied a variety of relevant experimental and theoretical methods to this problem, which is generic to medical imaging as a whole. Digital radiography of the chest and the breast has been employed as the clinical imaging domain vehicles for the study. The project addressed the problem from the following directions: role and relevance of pathology, human observer studies including receiver operating characteristics, image quality criteria analysis, structural noise analysis, physical measurements on clinical images, physical measurements on imaging system, modelling of imaging system, modelling of visual processes, modelling of doses delivered and IT-based scientific support strategies. This paper presents an overview of the main outcomes from this project and highlights how the research outcomes actually apply to the real world. In particular, attention will be focused on new and original findings and methods and techniques that have been developed within the framework of the project. The relevance of the project's outcomes to future European research will also be presented.

Simulation on the molecular radiosensitization effect of gold nanoparticles in cells irradiated by x-rays.

Abundant studies have focused on the radiosensitization effect of gold nanoparticles (GNPs) in the cellular environment with x-ray irradiation. To better understand the physical foundation and to initially study the molecular radiosensitization effect within the nucleus, a simple cell model with detailed DNA structure in the central nucleus was set up and complemented with different distributions of single and multiple GNPs in this work. With the biophysical Monte Carlo simulation code PARTRAC, the radiosensitization effects on both physical quantities and primary biological responses (DNA strand breaks) were simulated. The ratios of results under situations with GNPs compared to those without GNPs were defined as the enhancement factors (EFs). The simulation results show that the presence of GNP can cause a notable enhancement effect on the energy deposition within a few micrometers from the border of GNP. The greatest upshot appears around the border and is mostly dominated by Auger electrons. The enhancement effect on the DNA strand breakage becomes smaller because of the DNA distribution inside the nucleus, and the corresponding EFs are between 1 and 1.5. In the present simulation, multiple GNPs on the nucleus surface, the 60 kVp x-ray spectrum and the diameter of 100 nm are relatively more effective conditions for both physical and biological radiosensitization effects. These results preliminarily indicate that GNP can be a good radiosensitizer in x-ray radiotherapy. Nevertheless, further biological responses (repair process, cell survival, etc) need to be studied to give more accurate evaluation and practical proposal on GNP's application in clinical treatment.

[First clinical experience with a full-size, flat-panel detector for imaging the peripheral skeleton - Part II: Post-processing with a newly developed adaptive autowindow algorithm]. / Erste klinische Erfahrungen mit einem grossformatigen Flächendetektorsystem bei Aufnahmen des peripheren Skelettsystems - Teil 2 : Nachverarbeitung mit einem neu entwickelten adaptiven Autofensteralgorithmus -

PURPOSE: of the second part of the investigation was the evaluation of a newly developed adaptive autowindow algorithm in comparison to the system processing radiographs of the wrist and ankle to further optimize the image quality with softcopy reading. MATERIAL AND METHODS: All 120 radiographs of the wrist and all 100 radiographs of the ankle used in the 1st part of this paper were processed with the adaptive autowindow algorithm. The evaluation was again performed by 5 radiologists with softcopy reading. For the data analysis a variation of the Visual Grading Analysis (VGA) was used. RESULTS: Up to 19 % of the wrist radiographs and 2 % of the ankle radiographs processed with the system software had to be processed manually afterwards to get acceptable results. By the application of the adaptive autowindow algorithm a manual post-processing was no longer necessary. Highly significant (p less-than-or-equal 0.001) differences for all criteria to be evaluated were found for the wrist radiographs and in the case of the ankle radiographs for the bone contrast, the contrast in soft-tissue regions, the fine details in the bone and the artifacts, the adaptive autowindow algorithm performed always better than the system software. CONCLUSION: Using half of the exposition dose on a flat-panel detector, an optimized post-processing leads to comparable or better results compared to the conventional film-screen-system concerning the image quality.

[Chest radiography: ROC phantom study of four different digital systems and one conventional radiographic system]. / Thoraxübersichtsaufnahme: ROC-Phantomstudie an vier verschiedenen digitalen und einemkonventionellen Röntgenaufnahmesystem.

PURPOSE: To compare the diagnostic quality of five different radiographic systems used in chest radiography for visualization of differently configured, clinically relevant pathologic pulmonary structures. MATERIALS AND METHODS: Four digital detector systems using as detection unit a CsI/aSi-based flat panel detector, a transparent imaging plate, a selenium detector and a conventional storage phosphor plate were analyzed for this study, as well as an asymmetrical film-screen system. The analyzed imaging material consisted of radiographs of an anthropomorphic chest-phantom with superimposed simulated pulmonary structures. The images were evaluated for different pathologic structures by a newly developed multiple structure ROC (ms-ROC). RESULTS: The performance of each system was found to have a strong structure-related variability. The flat panel detector system had the best overall performance. The theoretical advantage of the 4k-matrix of the transparent imaging plate over the 3k-matrix of the flat panel detector was only confirmed for reticular structures. CONCLUSIONS: In addition to comparing the image quality of the different systems, this study shows that the performance of a radiographic system depends on the structure to be analyzed. The modified ROC (ms-ROC) provides valid results with less effort.

[First clinical experience with a full-size, flat-panel detector for imaging the peripheral skeletal system]. / Erste klinische Erfahrungen mit einem grossformatigen Flächendetektorsystem bei Aufnahmen des peripheren Skelettsystems - Teil 1: halbierte Patientendosis -

PURPOSE: This investigation was intended to show that exposures of the peripheral skeleton system can be done with half of the dose used for conventional screen-film systems with a full-size CsI/a-Si flat panel detector. MATERIAL AN METHODS: 120 exposures of the wrist and 100 exposures of the ankle have been made on a full-size flat panel detector system (43 x 43 cm). The patient dose has been reduced by a factor of two compared to conventional images. Five radiologists evaluated every image as a softcopy and a hardcopy image. For the evaluation, a variation of the Visual Grading Analysis (VGA) without reference images was used. For the determination of the patient entrance dose, measurement of a phantom were performed. RESULTS: A dose reduction of about 50 % is possible with the same or even better image quality in routine diagnostics. Only 3 % of the ankle and approx. 21 % of the wrist exposures required a postprocessing. Exposures with implants did not show any artifacts and some of the those achieved better evaluation results compared with exposures without implants. CONCLUSION: A halving of the patient dose is possible with acceptable results for the image quality. The effect of an improved image processing remains to be evaluated. The patient entrance dose is suitable for an evaluation of a radiographic detector and especially for a dose-referred comparison of digital X-ray units.

[Optimized image processing with modified preprocessing of image data sets of a transparent imaging plate by way of the lateral view of the cervical spine]. / Optimierte Bildverarbeitung durch ein modifiziertes Pre-Processing von Bilddatensätzen eines transparenten Speicherfoliensystems am Beispiel der seitlichen HWS-Aufnahme.

PURPOSE: To improve the diagnostic quality of lateral radiographs of the cervical spine by pre-processing the image data sets produced by a transparent imaging plate with both-side reading and to evaluate any possible impact on minimizing the number of additional radiographs and supplementary investigations. MATERIAL AND METHODS: One hundred lateral digital radiographs of the cervical spine were processed with two different methods: processing of each data set using the system-imminent parameters and using the manual mode. The difference between the two types of processing is the level of the latitude value. Hard copies of the processed images were judged by five radiologists and three neurosurgeons. The evaluation applied the image criteria score (ICS) without conventional reference images. RESULTS: In 99 % of the lateral radiographs of the cervical spine, all vertebral bodies could be completed delineated using the manual mode, but only 76 % oft the images processed by the system-imminent parameters showed all vertebral bodies. Thus, the manual mode enabled the evaluation of up to two additional more caudal vertebral bodies. The manual mode processing was significantly better concerning object size and processing artifacts. This optimized image processing and the resultant minimization of supplementary investigations was calculated to correspond to a theoretical dose reduction of about 50 %. CONCLUSION: The introduction of optimized organ programs for the upper and lower cervical spine based on the 12-bit data of the images should improve the evaluation of the lateral radiograph of the cervical spine without reducing the latitude value.

Effects of diet-matrix on volatile organic compounds in breath in diet-induced obese mice.

Breath gas analysis in humans proved successful in identifying disease states and assessing metabolic functions in a non-invasive way. While many studies report diagnostic capability using volatile organic compounds (VOC) in breath, the inter-individual variability even in healthy human cohorts is rather large and not completely understood in its biochemical origin. Laboratory mice are the predominant animal model system for human disorders and are analysed under highly standardized and controlled conditions. We established a novel setup to monitor VOCs as biomarkers for disease in the breath gas of non-anesthetized, non-restrained mice using a proton transfer reaction mass spectrometer with time of flight detection. In this study, we implemented breath gas analysis in a dietary intervention study in C57BL/6J mice with the aim to assess the variability in VOC signatures due to a change in the diet matrix. Mice were fed a standard laboratory chow and then exposed to four semi-purified low- or high-fat diets for four weeks. Random forest (RF++) was used to identify VOCs that specifically respond to the diet matrix change. Interestingly, we found that the change from a chow diet to semi-purified diets resulted in a considerable drop of several VOC levels. Our results suggest that the diet matrix impacts VOC signatures and the underlying metabolic functions and may be one source of variability in exhaled volatiles.

[Usability of smartphones for dose alerts]. / Smartphones jetzt noch smarter? - Möglichkeit des Einsatzes als "Dosiswarner"

PURPOSE: Smartphone apps for measuring ionizing radiation use the capability of (CMOS) camera chips to detect not only perceivable light but also electromagnetic wave radiation. The present study evaluates the accuracy of hardware and software and defines possible applications for the detection of X-ray radiation fields. MATERIALS AND METHODS: 2 apps and 2 different devices were tested in comparison with a calibrated ionization chamber and a personal electronic dosimeter. A calibration curve was determined for dose rates between 12 700 µSv/h and 5.7 µSv/h generated by a C-arm system. RESULTS: The measured scattered radiation produced by an Alderson-Rando phantom ranged from 117 µSv/h (at a distance of 2 m) to 5910 µSv/h (at a distance of 0.3 m) and was 1.4 times less than the values of the ionization chamber. The exposure rate for the operator's thyroid was within 4200 - 4400 µSv/h. We found a strong dependence of the measurements on the angulation of the Smartphone, especially for short distances from the phantom (at a distance of 0.3 m, a 45° rotation downwards in a vertical direction caused a decrease from 3000 µSv/h to 972 µSv/h, while an upwards rotation resulted in an increase to 5000 µSv/h). For a distance of 1 m, this effect was remarkably smaller. CONCLUSION: Smartphones can be used to detect ionizing radiation but showed limited accuracy and are heavily dependent on the angulation of the device. Qualitative measurements and utilization for dose alerts are possible.

Dose conversion coefficients for paediatric CT examinations with automatic tube current modulation.

A common dose-saving technique used in modern CT devices is automatic tube current modulation (TCM), which was originally designed to also reduce the dose in paediatric CT patients. In order to be able to deduce detailed organ doses of paediatric models, dose conversion coefficients normalized to CTDI(vol) for an eight-week-old baby and seven- and eight-year-old children have been computed accounting for TCM. The relative difference in organ dose conversion coefficients with and without TCM is for many organs and examinations less than 10%, but can in some cases amount up to 30%, e.g., for the thyroid in the chest CT of the seven-year-old child. Overall, the impact of TCM on the conversion coefficients increases with increasing age. Besides TCM, also the effect of collimation and tube voltage on organ dose conversion coefficients has been investigated. It could be shown that the normalization to CTDI(vol) leads to conversion coefficients that can in most cases be considered to be independent of collimation and tube voltage.

Influences of mixed expiratory sampling parameters on exhaled volatile organic compound concentrations.

Breath gas analysis is a promising technology for medical applications. By identifying disease-specific biomarkers in the breath of patients, a non-invasive and easy method for early diagnosis or therapy monitoring can be developed. In order to achieve this goal, one essential prerequisite is the reproducibility of the method applied, i.e. the quantification of exhaled volatile organic compounds (VOCs). The variability of breath gas VOC measurements can be affected by many factors. In this respect, sampling-specific parameters like flow rate and volume of exhalation, exhalation with or without breath holding, exhalation in single or multiple breathing and volume of air inhaled before breath gas exhalation can play a vital role. These factors affecting the measurements must be controlled by optimizing the sampling procedure. For such an optimization, it is important to know how exactly the different parameters affect the exhaled VOC concentrations. Therefore, a study has been undertaken in order to identify some effects of different breath sampling-specific parameters on the exhaled VOC profile using the mixed expired breath sampling technique. It was found that parameters such as filling the sampling bag with high or low flow rate of exhalation, with multiple or single exhalations, in different volumes of exhalation, with breath holding and under different surrounding air conditions significantly affect the concentrations of the exhaled VOCs. Therefore, the specific results of this work should be taken into account before planning new breath gas studies or developing new breath gas collection systems in order to minimize the number of artefacts affecting the concentration of exhaled VOCs.