OFR-7 Breath testing for colorectal polyps and cancer- the colorectal breath analysis1 study (COBRA1)

Introduction Colorectal cancer (CRC) is the 2nd commonest cause of death in UK, making up 11% new cases. Despite this, 5-year survival rates are 58%, amongst lowest Europe. There need for improvement because CRC can exceed 90% if detected early. This study aimed to investigate diagnostic accuracy breath testing CRC. If successful, a test could provide highly acceptable non-invasive tool improve and streamline existing referral pathways. lead earlier more targeted diagnosis, ultimately saving both lives resources. Method COBRA1 was multicentre cohort investigating volatile organic compounds (VOCs) 1432 patients attending one seven London hospitals colonoscopy or colorectal adenocarcinoma resection. Patient collected using ReCIVA sampler thermal desorption tubes. Patients were nil by mouth at least 6 hours sampled immediately prior their procedure, separate area. All pathologies diagnosed standard reference (colonoscopy), with histological confirmation CRC, polyps, inflammatory bowel disease. Breath analysed gas chromatography mass spectrometry (GC-MS), resulting VOC data processed GC-MS platform MSHub. A machine learning algorithm used identify components best discriminating ability develop multivariate discriminant analysis model. VOCs identified NIST spectral library. Results The presence (n=162) be predicted from positive negative controls (n=1270) sensitivity 77%, specificity 87% an area under ROC (AUROC) curve 0.87. 15 most predictive organosulphur, alcohol, alkane, ester phenol chemical groups. abundances also appeared related T stage tumour. Polyps (n=592) (n=619): 66%, 54%, AUROC 0.65. Subgroup symptomatic (n=855), demonstrated 82%, 88% detection, 0.90. Conclusions Analysis exhaled discriminate those without, gold-standard analytical method. Results compared favourably stool-based tests, particularly polyps patients. These exciting findings pave way larger studies validate results explore implementation into clinical practice.