Volatile organic compound profiles of lung fluids for PGD detection and monitoring

Volatile organic compound profiles of lung fluids for PGD detection and monitoring 

Pierre-Hugues Stefanuto1,2, Christiaan A. Rees3, Rosalba Romano4,5, Mavra Nasir3, Louit Thakuria5, Nicole Pagani5, Anna K. Reed5, Nandor Marczin4,5, Jane E. Hill1,3

1 Thayer School of Engineering, Dartmouth College, Hanover NH, USA, 2 Organic and Biological Organic Chemistry group, University of Liège, Liège, BE, 3 Geisel School of Medicine, Dartmouth College, Hanover NH, USA, 4 Department of Surgery & Cancer, Section of Anaesthetics, Imperial College of London, UK,  5 Harefield Hospital, Royal Brompton & Harefield NHS Foundation Trust, Harefield, UK

Primary graft dysfunction (PGD) is a major complication following lung transplantation. PGD reflects the summation of injury inflicted on the donor lung by the transplant process including donor related factors, preservation and reperfusion injury, intraoperative factors, and consequences of intensive care management. There has been significant progress in identification of genetic or molecular biomarkers of PGD. However, continuous method development may provide insight into disease pathogenesis.  The volatile organic compound (VOC) profiles of 58 bronchoalveolar lavage fluid (BAL) and blind bronchial aspirate (BBA) samples from 35 transplanted patients were analyzed using SPME-GCxGC-TOFMS. Data were split evenly into training and testing. The support vector machine algorithm (SVM) was used to identify VOCs that could differentiate patients with severe (grade 3) PGD from low (grade 0-2) PGD.  Using 20 VOCs we achieved an AUROC of 0.899 (95% confidence interval: 0.777 – 1.000) and an accuracy of 0.828 (95% CI: 0.642 – 0.941) on test set samples. In terms of statistical performance indicators, the final model has a sensitivity of 0.636, specificity of 0.944, positive predictive value of 0.875, and negative predictive value of 0.809. The miss-classification are mainly coming from patient with grade 2 PGD indicating the potential risk profile of these patients.  The analysis of lung fluids provided high accuracy, specificity, and sensitivity for PGD detection at time of transplant. Our data suggests that BAL/BBA has potential and should be investigated in a larger scale study. The potential translation to breath volatile analysis should also be investigated.