Metabolic signatures associated with lung cancer using serum samples. Influence of hemolysis

Sampling procedure greatly affects results and conclusions in metabolomic studies. This is the case of serum samples sustaining hemolysis, which need accurate evaluation in assessment of lung cancer (LC). This study considers metabolites present in serum samples after precipitation of proteins for analysis by direct infusion to a Triple Quadrupole Time-Of-Flight Mass Spectrometer (QqQ-TOF-MS). In addition, Hemolytic Index (HI) was semiquantitatively evaluated by Ultraviolet-visible spectrophotometry. The metabolomic profiles obtained were statistically compared by Partial least squares Discriminant Analysis, taking into account different hemoglobin concentration (Hb%) in the samples. Statistical models were built based on (Hb%) in the samples (>20, >40 y >80 mg/dL of Hb). Fig 1.A shows a clear classification between (LC) and healthy controls (HC) samples, regardless of sample hemolysis (LCH>20/HCH>20 -hemolyzed; LC/HC-not hemolyzed). When (HI) is higher than 40 (Fig 1.B), a proper statistical classification between hemolyzed (LC) and controls (HCH) is not achieved, since all hemolyzed samples are grouped in one single cluster. Fig. 1C represents the far tip of hemolysis influence, in which ((LC), green) and ((HC) reed) are not discriminated. Concluding, hemolysis is a disturbing factor in metabolomic studies related to (LC) and has to be considered in the methodological approaches for blood sampling in order to avoid it.