1560P - A breath test for diagnosing malignant pleural mesothelioma

Date 29 September 2014
Event ESMO 2014
Session Poster Display session
Topics Mesothelioma
Imaging, Diagnosis and Staging
Presenter Kevin Lamote
Citation Annals of Oncology (2014) 25 (suppl_4): iv542-iv545. 10.1093/annonc/mdu357
Authors K. Lamote1, J. van Cleemput2, K. Nackaerts3, O. Thas4, J.P. van Meerbeeck5
  • 1Respiratory Medicine, Ghent University Hospital, 9000 - Ghent/BE
  • 2Occupational Health Service, Eternit NV, Kapelle-Op-Den-Bos/BE
  • 3Respiratory Oncology Unit, University Hospital Leuven, Leuven/BE
  • 4Mathematical Modelling, Statistics And Bioinformatics, Ghent University, Ghent/BE
  • 5Thoracic Oncology/moca, University Hospital Antwerp, Edegem/BE

Abstract

Aim

Malignant Pleural Mesothelioma (MPM) is an asbestos-related disease with a dismal prognosis due to its late detection at an advanced stage (van Meerbeeck JP et al, 2011). As blood biomarkers have not shown to be clinically useful for early non-invasive diagnosis (Hollevoet K et al, 2012), breath is currently explored. Breath is easy to retrieve by physicians in a clinical environment and contains volatile organic compounds (VOCs) that arise from (patho)physiological processes (Lamote K et al, 2014; Buszewski B et al, 2012). Since asbestos causes oxidative stress and cancers are known to up regulate their metabolism, we hypothesize that VOCs and, hence, the exhaled breath of MPM patients will differ from healthy controls.

Methods

We compared the breath of 20 MPM patients, 10 asbestos-exposed and 10 non-exposed healthy individuals using a multicapillary column/ion mobility spectrometer (MCC/IMS, B&S Analytik, Dortmund, Germany). After subjects refrained from eating, drinking and smoking for at least 2 hours, 10 ml alveolar air was sampled via a CO2-controlled ultrasonic sensor. Per subject a background sample was taken. Eighty-nine VOC peaks were visually selected via on-board VisualNow 3.2 software. Their intensity was compared between background and breath samples. After calculating the VOCs' alveolar gradient, we performed a logistic LASSO regression using R (R Foundation for Statistical Computing, Vienna, Austria). We selected the optimal MPM diagnostic logistic LASSO model by 10-fold cross-validation using age, gender and the alveolar gradient of the peaks and subsequently estimated this models' sensitivity, specificity, AUCROC and positive (PPV) and negative predictive value (NPV) with 95% confidence intervals.

Results

We were able to discriminate MPM patients from the asbestos-exposed and non-exposed controls with 85% sensitivity (64%-96%) and 90% specificity (71%-98%). The AUCROC was 0,92 and the PPV and NPV were resp. 90% (69%-98%) and 86% (66%-96%). It was clear that age and the VOCs P5, P3, P83, P1 and P67 were the important discriminators.

Conclusions

Breath analysis in a clinical setting by IMS allows to discriminate MPM patients from asbestos-exposed and non-exposed healthy controls. Identification of the underlying VOCs and a further validation in different patient cohorts is being undertaken.

Disclosure

All authors have declared no conflicts of interest.