The Role of Data in Modern Healthcare

Breathomics at a Crossroads: How BreathBase® Data is Advancing the Field

Breath analysis has long been considered a promising frontier in non-invasive diagnostics. The ability to detect disease-specific volatile organic compounds (VOCs) patterns in exhaled breath offers an attractive alternative to traditional blood tests, imaging, and even invasive biopsies. Yet, despite its potential, breathomics has struggled to establish itself as a routine clinical tool.

The challenge is not a lack of interest. Research groups and companies worldwide have developed electronic nose (eNose) technologies and mass spectrometry approaches to analyze breath profiles. Studies have demonstrated that breath analysis can detect conditions ranging from lung cancer and asthma to infections and metabolic disorders. However, the field has been hindered by inconsistent results, small study populations, and a lack of standardized methodologies. Without large-scale, high-quality datasets, translating composite breath biomarkers into clinically validated diagnostics remains a major hurdle.

This is where BreathBase® Data comes in. As the largest and most comprehensive breathomics reference database, it provides the infrastructure necessary to move from exploratory research to clinically validated applications. By systematically collecting breath profiles alongside clinical data, this initiative is helping to bridge the gap between breathomics research and real-world healthcare.

Why Has Breath Analysis Struggled to Reach Clinical Practice?

The breathomics landscape is diverse, with different technological approaches competing for clinical adoption. Some groups focus on mass spectrometry and gas chromatography to analyze individual VOCs, while others use eNose technology to detect specific disease patterns in exhaled breath. Each approach has strengths, but all share a common challenge: the breath signal is highly complex and influenced by numerous factors including environmental exposures.

A common misconception is that successful diagnostic tests must rely on highly stable biomarkers. However, even in well-established blood-based diagnostics, many biomarkers fluctuate over time. For example, inflammatory markers like C-reactive protein (CRP) vary depending on infection or injury, glucose levels change dynamically throughout the day, and tumor markers often require serial measurements to track trends rather than relying on a single value. Similarly, in breathomics, disease-related VOC patterns may shift due to metabolic changes, disease progression, or external influences, making pattern recognition and machine learning essential for accurate interpretation.

Unlike traditional biomarker discovery, which often focuses on isolated molecules, breathomics benefits from biomarker patterns—a collection of VOCs that together create a disease signature. While this complexity has made clinical validation somewhat challenging, recent advances in data science, machine learning, and cloud-based calibration are helping to extract meaningful signals from breath data.

One of the key obstacles in breathomics has been data standardization. In contrast to imaging and genomics, which benefit from large, structured datasets, breath analysis has historically lacked comprehensive, multi-center reference databases. Many studies have relied on small cohorts or single-center data, making it difficult to confirm findings across different populations. Without standardization, promising results from one research group often fail to translate into broader clinical use.

BreathBase® Data was developed to address this gap by providing a structured, large-scale dataset with standardized methodologies, enabling more reliable biomarker discovery and validation across diverse patient populations.

The Role of BreathBase® Data in Overcoming These Challenges

BreathBase® Data was developed in 2018 to address these challenges by providing a structured, large-scale dataset that enables reliable biomarker discovery and validation. The database currently includes breath profiles and clinical data from over 164,000 individuals across 15 countries, 49 partner institutions and 162 professional users covering a broad spectrum of diseases, including cancer, chronic respiratory conditions, cardiovascular disease, infections, and post-transplant complications.

A key differentiator of BreathBase® Data is its focus on methodological consistency. Unlike many other breathomics studies, which have faced issues with comparability between different devices and study protocols, BreathBase® Data integrates advanced cloud-based calibration and standardized signal processing. This ensures that breath signals collected at different sites and on different devices remain comparable, a crucial factor for developing breath-based diagnostics that can be applied in diverse clinical settings.

Machine learning plays a central role in analyzing breath data, enabling the identification of disease-specific breath profiles rather than relying on individual VOCs. This data-driven approach has proven more effective than traditional biomarker discovery methods. Completed studies using BreathBase® Data have already resulted in validated disease signatures, including:

1. Early lung cancer detection in COPD patients (AUC of 0.90).
2. Differentiation of idiopathic pulmonary fibrosis (IPF) from COPD and lung cancer (AUC: 0.93).
3. Identification of sarcoidosis among other ILDs (AUC: 0.91).
4. Detection of chronic lung allograft dysfunction (CLAD) in lung transplant recipients (AUC: 0.82).
5. Detection of bacterial infections in CF patients, including Staphylococcus aureus (AUC: 0.80) and Pseudomonas aeruginosa (AUC: 0.88).

These results demonstrate that, with the right methodological framework, breath analysis can achieve diagnostic accuracies comparable or even better than established clinical tests.

What Needs to Happen Next?

Despite significant advancements, breathomics is still in the process of establishing itself as a widely accepted diagnostic tool in clinical medicine. The next critical steps involve expanding large-scale clinical validation, securing regulatory approvals, and defining standardized guidelines to ensure breath-based diagnostics can be reliably integrated into healthcare.

Regulatory approval remains one of the key hurdles. Health authorities require extensive validation studies demonstrating that breath-based diagnostics are reproducible, accurate, and clinically relevant across different populations. Initiatives like BreathBase® Data play a crucial role in accelerating this process by providing a standardized, multi-center dataset that allows for rigorous biomarker validation.

In addition to regulatory requirements, collaborative efforts between researchers, clinicians, and industry partners are essential for advancing breathomics. While various companies continue to develop proprietary breath analysis technologies, the field will benefit from open scientific dialogue, shared datasets, and cross-validation between independent studies to build a stronger foundation for clinical implementation.

As breath-based diagnostics continue to develop, they may complement, refine, or in some cases replace existing diagnostic methods, particularly in areas where traditional tests are invasive, costly, or impractical. The challenge lies in demonstrating clear clinical utility—whether by enabling earlier disease detection, improving diagnostic accuracy, or offering a more accessible and patient-friendly alternative to current approaches. Achieving this will require continued investment in research, standardization, and real-world validation studies to solidify the role of breathomics in modern medicine.

Conclusion: Closing the Gap in Breathomics

Despite its potential, breathomics has struggled with validation, standardization, and clinical adoption. BreathBase® Data is addressing these challenges by providing a structured, large-scale reference database that enables reliable disease signature identification across diverse conditions.

With over 164,000 breath profiles, validated diagnostic signatures, and a growing network of researchers and clinicians, BreathBase® Data fosters collaboration across institutions, enabling breathomics to move beyond isolated studies toward broad clinical validation. By bridging the gap between research and application, it lays the foundation for breath-based diagnostics to become a practical tool in precision medicine.

References

1. R. de Vries, N. Farzan, Y.W.F. Dagelet, E. Dijkers, T. Fabius, F.H.C. de Jongh, P.M.C. Jak, E.G. Haarman, S. Kester, M. Bekkers, M.M. van den Heuvel, P. Baas, J.C.C.M. in t Veen, A.H. Maitland-van der Zee, and P.J. Sterk. Prospective Early Detection of Lung Cancer in COPD Patients by Electronic Nose Analysis of Exhaled Breath.  CHEST. 2023 Nov;164(5):1315-1324.

2. I.G. van der Sar, M.S. Wijsenbeek, G.J. Braunstahl, J.O. Loekabino, A.M.C. Dingemans, J.C.C.M. In ‘t Veen and C.C. Moor.  Differentiating interstitial lung diseases from other respiratory diseases using electronic nose technology.  Respiratory Research. 2023 Nov 24; 271

3. I.G. van der Sar, N. van Jaarsveld, I.A. Spiekerman, F.J. Toxopeus, Q.L. Langens, M.S. Wijsenbeek, J. Dauwels and C.C. Moor. Evaluation of different classification methods using electronic nose data to diagnose sarcoidosis.  J Breath Res. 2023 Aug 29;17(4).

4. N. Wijbenga, R.A.S. Hoek, B.J. Mathot, L. Seghers, C.C. Moor, J.G.J.V. Aerts, D. Bos, O.C. Manintveld, M.E. Hellemons. Diagnostic performance of electronic nose technology in chronic lung allograft dysfunction. J Heart Lung Transplant 2022 Sep 20:S1053-2498(22)02133-7.

5. E. Seidl, J.C. Licht, R. de Vries, F. Ratjen and H. Grasemann. Exhaled Breath Analysis Detects the Clearance of Staphylococcus aureus from the Airways of Children with Cystic Fibrosis. Biomedicines. 2024 Feb 12(2), 431