Abstract 1217P
Background
Two randomized screening trials of low-dose CT in heavy smokers reported decreased mortality, yet lung cancer can occur in never or light-smokers (defined as ≤10 pack-years). Without screening, most are diagnosed in advanced stages. Electronic noses (eNoses) are promising devices designed to detect patterns of volatile organic compounds (VOC) in breath using environmental sensors, generating point-of-care results. This pilot study aims to build a classifier capable of distinguishing lung cancer patients from healthy controls who are all never or former light smokers (NoFLS).
Methods
This single-centre pilot study recruited NoFLS with or without a lung cancer diagnosis. Breath profiles were obtained in duplicate by a metal oxide semiconductor eNose. Data analysis was performed after advanced signal-processing, and results were compared using Wilcoxon rank sum tests and Principal Component (PC) Analysis.
Results
Of 48 NoFLS lung cancer cases, all had stage IV disease, 69% were female, 70% Asian, and were either EGFR+ (n= 36) or ALK+ (n=10); 38 were on or had previously received active treatment (tyrosine kinase inhibitors n=28, chemotherapy n=5, trial treatment n=5) . Age, sex and ethnicity were similar between cases and the 16 controls. Ten of 13 electronic nose sensor signals differed significantly between cases and controls. The top two PCs were significantly different (P<0.05) between cases and controls. When sensitivity was set at 100% (95%CI 0.93-1.00), specificity was 88% (95%CI 0.62-0.98). Signals of three sensors were significantly different (P<0.05) between treatment-naive cases and patients receiving systemic treatment at timepoint of measurement, suggesting a potential influence of treatment on VOC signatures.
Conclusions
Non-invasive breath analysis with eNose technology could distinguish healthy controls from advanced lung cancer cases in a NoFLS population. Studies of earlier stage cases including post-operative cases where tumor burden is low, are required to determine whether this technology might serve as a tool for the selection of never-smokers for low dose CT screening.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Geoffrey Liu.
Funding
Grant #706247; Grant Type: CCSRI Innovation Grants – 2019; Grant Name: Breathomics as a non-invasive, inexpensive, point-of-care predictive test for immune checkpoint inhibitor efficacy.
Disclosure
S. Schmid: Honoraria (institution): Boehringer Ingelheim; Honoraria (institution): MSD; Research grant/Funding (institution): AstraZeneca; Research grant/Funding (institution): BMS; Travel/Accommodation/Expenses: Boehringer Ingelheim; Travel/Accommodation/Expenses: MSD; Travel/Accommodation/Expenses: Takeda. E. Stewart: Full/Part-time employment, Pentavere. All other authors have declared no conflicts of interest.