Markes and SepSolve host seminar on analytical strategies to improve data quality in biomarker discovery

In June, Markes and SepSolve held a one-day seminar entitled “Analytical strategies to improve data quality in biomarker discovery”. The seminar brought together leading experts in volatile biomarker analysis from across the UK, and was hosted by Professor George Hanna and his research team at Imperial College London.

For me, the highlight of the day was a panel discussion on:

Standardisation in breath analysis: Where do we start?

The panel was chaired Dr Dave Worton (NPL), with panellists Dr Waqar Ahmed (University of Manchester), Aaron Parker (Imperial College London) and Prof. Paul Thomas (Bioxhale), whose combined experiences cover the whole spectrum of biomarker discovery and clinical application. The following questions were posed to our expert panellists to work towards a practical solution to standardisation in breath analysis:

What are the most important measurement challenges in breath analysis research?

Aaron highlighted that the complexity of breath sample collection and the sample matrix itself is the most difficult element to control. Collecting a representative sample, choosing end-tidal or whole breath and how to meter the volume collected for quantitative studies is an area his team have focused on. The panel agreed that consistency and ease of use is more important than developing the perfect sampling protocol. Paul highlighted that the wide concentration range of potentially important volatile organic compounds (VOCs) within a single breath sample spans orders of magnitude, meaning high patient numbers in discovery studies are critical to account for variability and to produce statistically significant data for further validation studies.

Aaron also explained the challenge of understanding and controlling for individual errors and their cumulative effect in breath analysis workflow, and how the clinical breath analysis workflow is a multi-step process in which samples pass through numerous pairs of hands. He described the steps taken by the Hanna Group to introduce routine competency checks, automated audit trails, regular additional audits and quality control checks at as many stages as possible in their efforts to move towards accreditation.

Waqar then raised the issue of analytical quality control over long studies as another key challenge facing researchers. Clinical studies take place over extended time periods, but patient recruitment cannot always be spread evenly throughout a study period and samples collected on sorbent tubes can only be stored for several months, not years. This combination of factors means it’s necessary to analyse samples in multiple batches across the study duration, adding in the obstacle of maintaining analytical instrument performance and correcting for drifts due to ageing consumables or routine maintenance across the study. Thankfully, Waqar also presented some exciting new quality control data from his own efforts to tackle this challenge, including the use of pooled breath samples to correct for instrument drift.

Does there need to be a standard sampling approach for breath analysis? Would defining a standard suite of compounds (a universal breath standard) be useful to enable the validation of sampling methods and/or analytical tools?

Paul was keen to emphasise the importance of a standardised approach for comparable and repeatable results, as well as for ease of adoption in a clinic – basically, it must be slick and quick with minimal training burden. He also highlighted that the sample must be as free from contamination as possible – be that environmental or from the patient themselves.

The concept of a universal breath standard was highly attractive to our panel, but as the discussion progressed it became clear that defining what a standard would be used for and how closely it should mimic a real breath sample is much more complex to agree upon. The humidity levels, concentration ranges and stability of stored samples were all mentioned.  Once Dave pressed the panellists for a viable starting point, two streams of work were suggested. The first was to take inspiration from other industries with standardised methods that have similar complexities and challenges, such as environmental air monitoring or food safety methods written by ISO or CEN technical committees. The second was that the complexity, concentration range and different types of matrices are all important factors when validating equipment for standardisation, and how a centralised facility or inter-laboratory study could be used to achieve this. Prof. Hanna even suggested to Dave that this could be developed at NPL for the breath research community to use.

What are the barriers to adoption/translation of breath analysis research into more routine clinical practice?

Waqar approached this question from both the mindset of an analytical scientist and that of a clinician, asking if it would cause more burden in the clinic, and if the timeframe for results was relevant to the situation. He explained that prioritising research into relevant conditions and solving the analytical challenges to streamline and drive down the costs for those tests to deliver fit for purpose tools should be a priority to move towards widespread use of breath testing in clinical practice.

Aaron agreed that the most efficient path to routine clinical practice is to adopt the centralised biochemistry lab model rather than driving for point of care sensors in one giant step. Paul also challenged us all to continue to innovate in parallel, consider the health economics of an aging population and continue to invest into detecting disease states via routine screening before people know they are unwell.

Standardisation is often mentioned but what is stopping progress? What is needed to be able to move forward? What could instrument manufacturers and metrology institutes do to help?

Prof. Hanna suggested the need for a dedicated institution to be able to take models from other industries and to adapt them for the unique challenges in breath analysis. Dave added that the potential for positive impact on national health aligns with the NPL mission, and that they are keen to work with the breath research community to advance the measurement science.

Prof. Hanna was also keen to point out the responsibility of research institutes to recruit staff from beyond academia to bring experience in standardisation, accreditation and good laboratory practice alongside the innovation and research. Aaron mentioned that in his lab, the team have adopted this model and split their efforts into parallel focused streams of untargeted discovery and longer term clinical trials. With this model, they are ensuring good practice and analytical excellence without stifling innovation.

Finally, Waqar reminded us that NHS clinical laboratories, which only operate under accreditation and where quality control and standardisation are paramount to success, are resources the community should consult as they look to develop standard methods and achieve accreditation.

The take home message for me is that there is a real desire in the research community to strive for excellence, and to build in as much standard practice as possible to ensure highest quality data for reliable testing. However, the next 10 years must not aim for perfection – instead, we must pull together to share experiences and innovations to identify and validate fit-for-purpose solutions, remove the barriers to large scale clinical trials and validation studies, and to reach the goal of adding practical, sensitive and specific breath tests to the clinician’s diagnostic tool kit to improve patient outcomes.

My colleague, Dr Laura McGregor of SepSolve Analytical, has also written a blog post summarising the talks given on the day. In this blog you can also find links to the videos of the presentations.