Breath data of patients who underwent eNose examination is stored in an extensive, online database, which is linked to an artificial intelligence application called BreathBase. This platform provides a real time answer at point-of-care within a few minutes. In addition, it allows collaboration between physicians and researchers using the database for shared analysis.
BreathBase is predominantly used in lung diseases but other applications are certainly possible. BreathBase allows assessment of the most probable diagnosis, but can also be used for phenotyping individual patients and for predicting the response to (novel) interventions (tailored therapy). Hereby the system meets the growing demand for personalized or precision medicine.
Why using exhaled breath?
Volatile organic compounds (VOCs) present in exhaled breath provide a picture of a person’s metabolism, which can be used to diagnose diseases before other physical symptoms become apparent. Just like tracing molecules in our blood or urine, one can measure the molecules – and they are quite numerous – in our exhaled breath. There can be more than 3,000 different molecules in our breath. Where do they come from? Not just from the lungs, but also from other parts of our body. By means of gas exchange in the lungs, molecules in our bloodstream can also eventually end up in our exhaled breath. Our breath can therefore contain molecules that are being released in case of lung disease or illnesses elsewhere in the body, e.g. inflammation, infection or cancer. These abnormalities change the behaviour of cells in our body and thereby altering the molecules that are being released by these cells. VOC biomarkers are present even at the earliest stage of a disease and directly reflect real-time disease activity. The main advantage of VOC measurement over other types of diagnostic tests is that VOCs can be collected easily an non-invasively from our exhaled breath.
Diagnostic tests are an essential part of modern medicine. The ultimate goal of diagnosis and monitoring is to optimize the outcome or prognosis for the patient by giving the clinician directions for a clinical management strategy. Even though physiological and cell-based procedures, such as spirometry, blood test and induced sputum are often routinely available in respiratory medicine, molecular diagnostic tests are not yet widely applicable in the doctor’s office.
The missing link in breathomics provided by Breathomix B.V. is the combination of the easy to use eNose and an online reference database of exhaled biomarker profiles combined with validated clinical data that can be implemented at any location in the world.
The online reference database is linked to a computer program based on artificial intelligence; BreathBase. The link with validated signal processing and machine learning algorithms allows real-time analysis of the obtained exhaled biomarker profiles.
REPRODUCIBLE AND INTERCHANGEABLE
Because the sensor data of the eNose is reproducible and interchangeable (De Vries ERJ 2018 & De Vries JBR 2015) it will now become possible for researchers and physicians to exchange data all over the world via the internet in order to learn from each others results.
DATA DRIVEN MEDICINE
BreathBase enables point-of-care personalized medicine by providing immediate diagnostic answers for the individual patient. The breath data of each patient will be added to the expanding database to make subsequent diagnoses more accurate.
The patient takes a deep breath and calmly exhales through the eNose. The cross-reactive sensors in the eNose capture the complete mixture of molecules in the patient’s exhaled breath.
The sensors respond to the exhaled breath of the patient. All the sensors react in a different way, because each sensor is sensitive to a partly diffferent group of molecules.
The various sensor readings are subsequently processed and corrected for the environment by BreathBase. This is necessary since the air you breathe in obviously also influences the air you exhale.
Finally, the obtained breath profiles are compared to the online reference database. On the basis of machine learning algorithms a final diagnostic and/or phenotypic answer will be established.
To demonstrate that breath signals are useful as biomarkers for diagnosis, they had to be tested at the time of diagnosis and in the population in which the test should be used clinically. We have embraced this concept and tested thousands of patients (with and without an established clinical diagnosis) at the point-of-care, which is why we now know that BreathBase makes the assessment of the most probable diagnosis possible.
AT RISK POPULATIONS
Prediction of disease development is one of the most important premises of any biomarker. Breath analysis is especially suitable for this because measurements can be repeated frequently, which allows for monitoring the likelihood of disease development. Additionally, breath analysis is one of the few acceptable methods for molecular analysis in preschool aged children because of its non-invasive nature.
With development of very costly new targeted therapeutics, prediction of treatment response is becoming ever more important. Therefore, prospective studies to validate the predictive performance of exhaled breath analysis in patients starting treatment are mandatory. BreathBase cooperates with clinical partners to identify patients that most likely will benefit of immunotherapy (Pembrolizumab and Nivolumab), Mepolizumab, Reslizumab and Benralizumab.
What is next?
It is of great importance to currently test the eNose on a bigger scale in research centres in and outside The Netherlands. This is required for: 1) Even higher accuracy for determining the most probable diagnosis, 2) External validation studies for the prediction of the most effective therapy (tailored therapy) and 3) Cost-benefit analysis. In doing so, the collaborating medical centres are focusing on patients with a broad range of respiratory symptoms. The results provided by the BreathBase system are then being compared with the current clinical diagnosis, based on standard pulmonary function tests, blood tests and imaging. This allows validation of more than 1,500 cases with a differential diagnosis of asthma, COPD, lung cancer and other respiratory conditions.
During this program the online reference database filled with breath profiles will increase in size, whilst the self-learning algorithm based on Artificial Intelligence (BreathBase) is further optimized. This will ensure that the system will offer every new patient a personal advice within one minute, regarding the most probable lung disease, the individual phenotype and the most effective therapy.
As a pulmonologist, we often diagnose lung cancer too late. I would like to use the simple and fast eNose technology in combination with the BreathBase platform to detect disease at an earlier stage. Then I can use specialized equipment for further research in the right patients.
Prof. Dr. Paul Baas
The Netherlands Cancer Institute
The eNose and BreathBase are representing a very modern, molecular assessment of disease. It includes accurate and real-time analysis of the entire metabolomics spectrum from exhaled air. It provides a wealth of data on both health and disease, which a few years ago was just not possible.”
Prof. Dr. Peter Sterk
Amsterdam University Medical Centre
Our technology partners:
Breathomix B.V. consists of a team with over 20 years of combined experience in the field of
breathomics, cloud applications, machine learning and big data analysis. It is our strong conviction that this team can bring eNose technology from a research setting to the patient.
Maurik van den Heuvel
Chief Executive Officer
Maurik van den Heuvel has professional experience in Artificial Intelligence and data science & analytics. He has a degree in economics and is working on a PhD in drug utilization research. His principal areas of expertise are in the pharmaceutical industry, healthcare, FMCG and finance. He has broad international business experience with top tier manufacturing and consulting companies. Maurik brings the experience and the people needed to implement the cutting edge technology that helps make BreathBase scalable, secure and successful.
Rianne de Vries
Chief Operational Officer, Technical Physician
Rianne de Vries studied Technical Medicine with a masters degree in Medical Signaling. She forms the hybrid between classical medicine and advanced technology and is therefore a new professional who has the knowledge, skills and problem-solving mind-set to design and safely apply improved diagnostics and therapeutics. She developed BreathBase; the online platform that is connected to the eNose and automatically processes and analyzes the eNose sensor signals in order to diagnose and phenotype disease or to predict therapy response.
Prof. Peter Sterk, MD, PhD, has a background in clinical respiratory physiology and is focusing on a systems medicine approach to link complex biology to clinical management of disease. He has launched exhaled breath analysis by eNose in various clinical settings during the past decade. His primary interest is phenotyping of chronic airways diseases, in which eNose analysis has been validated for diagnostic purposes and for subclassification of patients. This enables the identification of responders to existing and new treatments (biologicals). In this way he is contributing to the dissemination of Data Driven Medicine at point of care.
Razvan Furca currently serves as the CEO of Tecknoworks, a worldwide leader in custom software service. He is a highly-driven, customer-focused software developer with a passion for breathing life into big ideas. Razvan grew Tecknoworks from a one man operation into a world-wide brand with offices around the globe. Over the course of his career, Razvan has become known not just as a visionary leader in the world of software development, but also as a trusted partner among firms seeking ways to prepare for the future and do better business using cutting-edge technology.
Breathomix is here to provide you with more information, answer any questions you may have and create collaborations