Non-Invasive Detection of Parkinson’s Disease: VOC Profiling Using Chromatography and Mass Spectrometry

A review article (1) written by researchers at Imperial College London (United Kingdom) explored the potential of volatile organic compounds (VOCs) as early, non-invasive biomarkers for Parkinson’s disease (PD), a neurodegenerative condition that currently lacks reliable diagnostic tools for early-stage detection. While traditional imaging methods like [123I]FP-CIT SPECT assist in diagnosis, they are insufficient for disease differentiation or early intervention.

VOCs—metabolic byproducts detectable in breath, skin, stool, blood, and urine—offer a promising diagnostic avenue due to their sensitivity to physiological changes. Although VOC-based diagnostics have been more thoroughly investigated in oncology and respiratory diseases, their application in neurology, particularly PD, remains limited but is expanding. This review evaluates existing research on VOCs in PD, the technologies used for their detection, and their potential biological relevance, with a focus on standardization challenges and clinical applicability. LCGC International spoke to Ilaria Belluomo of Department of Surgery and Cancer, Imperial College London and lead author of the review, about how chromatography and mass spectrometry played a part in the team’s research.

Why is GC-MS considered the gold standard for VOC analysis in Parkinson’s disease (PD) research, and what specific strengths make it useful in this context?

GC-MS is considered the gold standard not only for VOC analysis in PD, but for VOC analysis in general. This technique has exceptional sensitivity and allows for confident identification of a wide range of compounds. This level of detail is essential in PD research, where differences in VOC levels may be subtle, since they are likely linked to complex biological processes, such as changes in microbiome or oxidative stress. The robustness and reproducibility of GC-MS makes it particularly well suited for identifying disease-specific VOC signatures, that could support non-invasive diagnosis or monitoring.

What are the main technical or biological limitations of GC-MS when applied to complex samples like breath or stool, and how does this impact biomarker discovery?

Sample preparation can be technically challenging. Due to the high volatility of these compounds, sample collection and storage needs to be carefully planned to maintain sample integrity. Stool contains a complex mixture of chemicals with many interfering substances that can affect the overall VOC detection. These limitations can reduce sensitivity and reproducibility. Biological confounding factors need to be considered as well. VOC profiles vary greatly between individuals due to factors like diet, microbiome, medications, and environment, making it hard to identify specific disease-related VOC patterns. All these factors together can ultimately impact the reliability of potential biomarker identification and slowing down their translation in clinical use. Large-scale validation studies need to be performed to confirm the specificity of the biomarkers and their biological link to the disease.

In what ways can LC-MS complement GC-MS in detecting less volatile or more polar metabolites relevant to PD pathology?

LC-MS can be a complementary technique to GC-MS for some applications. The use of this technique may broaden the range of detectable compounds, allowing the analysis of less volatile and more polar metabolites. GC-MS is an ideal technique for small, volatile molecules, while LC-MS can be the technique of choice to analyze lipids and small peptides, that may be relevant in Parkinson’s research. By combining both these techniques, we could obtain a more comprehensive metabolic profile, improving the chances of identifying biomarkers involved in different aspects of PD.

What role does thermal desorption (TD) play in sample collection and standardization for large-scale PD studies using GC-MS?

Thermal desorption (TD) plays a critical role in standardizing breath sample collection for large-scale PD studies using GC-MS. While direct sampling techniques are valuable for real-time analysis and reduce the risk of sample loss, they are less practical for large-scale multi-center studies where storage and transport of samples are essential. TD allows for reliable collection and storage of breath samples on sorbent tubes, enabling centralized, high-throughput analysis with minimal degradation or loss of VOCs. TD tubes are very easy to transport, robust and of practical use. This is crucial for large validation studies, which are needed to identify robust biomarkers across diverse populations. Since PD is a common disease that can affect anyone, having a scalable and standardized method like TD ensures that we can capture a representative picture of the disease and validate findings across different cohorts and sites, allowing large national and international studies.

How do solid phase microextraction (SPME) and TD compare in terms of sensitivity, reproducibility, and ease of integration in clinical protocols?

SPME and TD are both powerful techniques for pre-concentrating volatile organic compounds (VOCs) prior to GC-MS analysis, but they differ in several key aspects that impact their practical use in clinical research. SPME uses a coated fiber to adsorb analytes directly from the sample matrix, offering solvent-free extraction and high sensitivity, especially for specific target compounds. However, it requires breath to be collected in bags, syringes, or other enclosed vessels before the fiber can be inserted in the GC-MS. This adds complexity and introduces potential variability. Additionally, the fiber itself is delicate and has a limited lifespan, making it less suited for high-throughput studies. TD, on the other hand, involves collecting breath directly onto sorbent-packed tubes, which can be sealed, stored, and shipped easily. This method is far more practical for large-scale, multi-center studies, especially in diseases like Parkinson’s, where large, diverse populations are needed to validate findings. TD systems are more robust and better suited for processing large numbers of samples consistently, with fewer logistical challenges. In terms of sensitivity, both SPME and TD can achieve high analytical performance when properly optimized, but TD tends to offer better reproducibility, sensitivity and stability over time. Overall, while SPME still remains a valuable tool, TD is generally the preferred choice in clinical protocols due to its robustness, practicality, and scalability.

What are the challenges of aligning VOC profiles identified by sensor arrays with those detected by GC-MS? What accounts for these discrepancies?

Sensor arrays offer several practical advantages for detecting VOCs: they can be miniaturized, are highly sensitive to chemical changes, and provide rapid response times, making them ideal for real-time breath analysis. However, they lack specificity due to potential cross-reactivity with other compounds in breath and environmental interferences, which can lead to false positives or inaccurate readings. Unlike GC-MS, sensors have difficulties to identify or quantify individual VOCs but rather detect global patterns that help distinguish patients from controls. This pattern-based approach is useful for classification but cannot reveal the exact biochemical differences and the possible biological link with the studied disease. As a result, sensor data often needs confirmation by GC-MS, which provides detailed chemical identification. Discrepancies between the two methods can arise from different sampling approaches (for example, direct sampling for sensors vs. TD for GC-MS), analytical techniques, sensitivity and dynamic range of the detector. Sometimes, GC-MS may not replicate the same disease-control separation observed by sensors due to these differences. Overall, sensor arrays are a highly promising tool for future non-invasive diagnostics. However, further technical improvements are needed to enhance their sensitivity, specificity, and integration with validated analytical platforms like GC-MS. In general, MS techniques are currently preferable for discovery studies, and once defined the range and identity of the compounds, sensors will be ideal for translation.

Can you discuss the value of untargeted versus targeted GC-MS approaches in PD biomarker discovery?

In PD biomarker discovery, both untargeted and targeted GC-MS approaches play important and complementary roles. Untargeted GC-MS is typically used in the early stages of research, when the specific compounds linked to the disease are still unknown. This approach aims to detect as many VOCs as possible without prior assumptions, providing a broad chemical fingerprint. Tentative identifications are made by matching spectra to libraries. However, this method can be imprecise and sometimes lead to inaccurate identifications, especially when VOCs share similar structures or mass spectra.

Targeted GC-MS, on the other hand, is applied when specific compounds of interest have already been identified, usually with authentic standards. It uses pre-set acquisition parameters and quantification performed with calibration curves built with specific chemical standards, allowing for accurate, sensitive, and high-throughput analysis. Since no validated breath biomarkers currently exist for PD, untargeted approaches are essential for initial discovery. Once reliable biomarkers are identified, targeted methods will become key for routine clinical application and large-scale screening.

How reliable is compound identification through spectral library matching in untargeted GC-MS, and what strategies are used to improve confidence in annotations?

Compound identification through spectral library matching is a central step in untargeted GC-MS workflows, but it comes with limitations in reliability. Annotation is crucial for biomarker analysis, as identifying the correct compound is key to linking chemical changes to underlying biological processes. Online spectral libraries (such as NIST or Wiley) contain thousands of reference spectra and offer a valuable resource for matching unknown compounds detected in samples. However, matches are based on similarity scores, and misidentification can occur, especially when structurally similar VOCs produce overlapping or nearly identical spectra.

To improve confidence, researchers must ensure the mass spectra is free from interference from coeluting compounds, and this is one of the biggest challenges in GC-MS. Utilizing techniques such as multidimensional GC (GC×GC) and signal processing algorithms can aid in producing robust and clean spectra to be compared to commercial libraries. In addition, to further improve reliability of results, researchers must go beyond in silico matches and confirm identities using analytical standards. These standards need to be run under the same experimental conditions to validate retention times and fragmentation patterns, ensuring a higher accuracy in compound identification. This chemical validation also enables downstream biological interpretation, allowing assessment of whether identified compounds are linked to specific biochemical pathways disrupted, thereby strengthening their potential as true disease-related biomarkers.

Nuclear magnetic resonance (NMR) is sometimes used despite lower sensitivity. What does NMR offer that MS-based methods might miss in the context of PD biomarker discovery?

While NMR has a lower sensitivity, it can offer some advantages compared to mass spectrometry techniques that can make its use valuable in PD. NMR provides detailed structural information and be preferable for the detection of some classes of metabolites that may be relevant for PD. The reliable identification of metabolites enabling confident assignment. This level of confidence is particularly useful in hypothesis-free untargeted metabolomics, like for PD, where accurate metabolite annotation is essential. However, the already mentioned low sensitivity of NMR remains a significant limitation, especially when working with complex biological samples like biofluids, where many relevant metabolites may be present at trace concentrations. This means that NMR may miss subtle but potentially important biomarkers that MS-based methods can detect due to their much higher sensitivity and broader dynamic range.

How close are we to integrating validated GC-MS-based breath or skin tests into clinical workflows for PD screening or diagnosis?

We are still some distance from fully integrating validated GC-MS-based breath or skin tests into clinical workflows for Parkinson’s disease screening or diagnosis, but the path forward is becoming clearer. Such non-invasive tests would represent a major step toward earlier, easier, and more accessible diagnosis, which is crucial since motor symptoms appear only after substantial neurodegeneration has already occurred.

However, several critical steps remain. Large-scale, high-quality studies are needed to validate findings across diverse and representative populations, with appropriate control groups and standardised protocols that account for variables like diet, environment, and medication. Many studies to date have small sample sizes and inconsistent control matching, limiting the use of the results. Moreover, deep profiling using high-resolution GC-MS is essential to confidently identify VOCs related to PD pathology. Once key biomarkers are selected, more practical targeted approaches can be developed for population-level screening, such as simplified GC-MS setups or even sensor-based platforms.

There is an urgent need for simpler, more accessible methodologies, especially in Parkinson’s disease and other conditions where reduced mobility can make traditional diagnostic approaches challenging. Sensors are expected to play a central role in the future of non-invasive diagnostics due to their portability and ease of use. However, as mentioned earlier, technical refinement is still required to improve their specificity, sensitivity, and reproducibility, and to better align their outputs with those of gold-standard GC-MS methods. While we are not there yet, with continued investment in robust, large-scale, multi-center research and methodological standardization, breath or skin testing could soon become a valuable part of PD clinical care.

What standardization steps—across sample collection, instrument calibration, and data analysis—are most urgently needed to make GC-MS-based VOC analysis clinically viable for PD?

Standardization across all stages of the analytical process is urgently needed to make GC-MS-based VOC analysis clinically viable for Parkinson’s disease. Each of these steps plays a critical role in ensuring reproducibility, reliability, and comparability of results across studies, different laboratories, and populations. In sample collection, methods must be harmonized, especially for breath analysis where variations in collection devices (e.g., bags, tubes, direct interface), timing, and environmental conditions can significantly influence VOC profiles. Instrument calibration must also be tightly controlled. Without this, even small technical differences across labs or time points can skew data and make comparison difficult. Finally, data analysis needs to be carefully built as well, since inconsistency in processing workflows leads to divergent results. Without process standardization it is impossible to determine whether observed VOC signatures reflect true biological variation or methodological artifacts. For VOC biomarkers to move toward clinical use in PD, standardized, reproducible methods are a top priority.

Reference

1. Belluomo I, Tarazi M, Lao-Kaim NP, Tai YF, Spanel P, Hanna GB. Detection of Volatile Organic Compounds as an emerging strategy for Parkinson's disease diagnosis and monitoring. NPJ Parkinsons Dis. 2025 Jun 12;11(1):161. DOI: 10.1038/s41531-025-00993-2