Noninvasive Proteomic Markers for Respiratory Tract Infections in Mechanically Ventilated Patients

IntroductionEarly and accurate diagnosis of respiratory tract infections (RTI) in critical care settings is essential for appropriate antibiotic treatment and lowering mortality. The current diagnostic methods face critical challenges, including the lack of noninvasive specimens from the site of infection and molecular biomarkers that can predict disease progression and treatment effect. In this study, we addressed these critical challenges by developing a noninvasive method based on the characterization of truncated proteoforms contained in exhaled air collected from mechanically ventilated patients. MethodsExhaled air samples were collected from twenty-five intubated patients with RTI and twenty-two intubated patients without RTI, determined by clinical data and microbiological testing. Truncated proteoforms were identified using top-down proteomics. Feature selection algorithms were used to identify significant truncated proteoforms associated with RTI. A score system combining the significant truncated proteoforms was constructed and evaluated using multiple logistic regression to predict RTI. ResultsThe results showed that six truncated proteoforms of lung structure and proteolytic proteins were statistically different between intubated patients with and without RTI. Specifically, the truncated proteoforms of collagen type VI alpha three chain protein, matrix metalloproteinase 9, and putative homeodomain transcription factor 2 were found to be independently associated with RTI. A score system named TrunScore was constructed by combining the three truncated proteoforms, and the diagnostic accuracy was significantly improved compared to individual truncated proteoforms. ConclusionsIn this study, we presented a noninvasive method to address the current challenges in diagnosing RTI in critical care settings, by characterizing truncated proteoforms contained in exhaled air from intubated patients. The method provides an accurate prediction for RTI in mechanically ventilated patients and can help diagnose other respiratory tract diseases.