Secondary electrospray mass spectrometry (SESI-MS) and liquid chromatography mass spectrometry (LC-MS) for comparative volatile/semi-volatile organic compound profiling in human plasma and breath

Volatile and semi-volatile organic compounds (VOC/sVOCs) play crucial roles in metabolic processes and serve as potential disease biomarkers. Despite their clinical significance, direct comparisons of VOC/sVOCs detected in both human breath and plasma remain limited. Breath analysis could offer a non-invasive alternative to traditional blood-based diagnostics, but establishing correlations between these two matrices requires optimized analytical workflows. This study aims to compare VOC/sVOCs in human breath and plasma using secondary electrospray ionization-mass spectrometry (SESI-MS) and liquid chromatography-mass spectrometry (LC-MS), respectively. Sample preparation methods common for plasma metabolomics are compared for breath and plasma data, and data processing workflow parameters are critically evaluated to determine their impact on these data sets. Breath and plasma samples from a human cohort were analyzed to identify shared and unique VOC/sVOCs. Plasma metabolites were extracted using several extraction systems with a standard metabolomics dry-down step to optimize overlap with breath features and analyzed using LC-MS. Real-time SESI-MS was used for breath analysis, with data processed via the new BreathXplorer pipeline, a dedicated high-resolution breath analysis workflow. Previous internal processing methods were compared with BreathXplorer to assess data quality. Compound matching between breath and plasma was performed, generating unique feature lists for each. The 100% methanol extraction with a dry-down step yielded the highest number of overlapping VOC/sVOC features between breath and plasma. BreathXplorer demonstrated a superior number of detected features compared to previous internal workflows. Eleven compounds were detected in both breath and plasma, suggesting the potential diagnostic capability of shared features, while distinct breath-only and plasma-only feature lists highlighted matrix-specific feature distributions. These findings expand upon the limited field of blood-breath feature relationships, advancing the potential of breath-based diagnostics for biomarker discovery. Optimized sample preparation and data processing workflows significantly improve reproducibility and sensitivity in VOC/sVOC analysis. Future research should validate these annotated features in larger cohorts and continue to explore the relationship between features that are common to both breath and plasma. Not applicable.