Breath Research with SIFT-MS

14 months ago

We all prefer to receive a diagnosis from our doctor without having blood taken. A future where this is the norm is envisaged by those exploring breath analysis, using the VOCs in exhaled breath as indicators of a patient’s physiological condition. Advancements in trace VOC analysis technologies such as SIFT-MS bring this vision within reach. However, there remain several barriers to clinical use.

Forerunners in the field of SIFT-MS, David Smith and Patrik Španěl, recently penned a perspective in which “a plea is made for more effort to be directed towards the positive identification and accurate quantification of individual VOCs in exhaled breath.” They noted that, by employing SIFT-MS, reliable concentration ranges have been established for many common breath compounds. However, because of the considerable inter-individual variation in VOC concentrations, large cohorts of sick and healthy subjects are required to statistically validate biomarkers for diagnostic tests.

Analysts at the Chinese University of Hong Kong endeavored to surmount this obstacle by employing the more powerful least absolute shrinkage and selection operator (LASSO) statistical model to accurately predict blood creatine and urea concentrations using breath samples from hemodialysis patients. This novel approach employs full mass-scan SIFT-MS data, which contain all mass counts for the sample at a given time. LASSO selected masses were shown to be better predictors of urea and creatine concentrations than the more well studied indicator compounds acetone, ammonia and trimethylamine, which were monitored in tandem.

With two different approaches addressing the same challenge, which is the true solution? Whatever the chosen route, developing a valid diagnostic is about combining the right analytical technique with a large number of independent test samples and the most appropriate statistical method to provide reliable prediction.

To read Smith and Španěl’s full perspective titled “On the Importance of Accurate Quantification of Individual Volatile Metabolites in Exhaled Breath” follow the link to the publisher’s website: 

Likewise, for a discussion of the LASSO statistical model as implemented for breath analysis with SIFT-MS data, visit the following link:

DSC 3372 2 copy
Lalit Rane Bsc IT
Marketing Manager