Non-invasive breath testing to detect colorectal cancer: protocol for a multicentre, case-control development and validation study (COBRA2 study).

Background: Colorectal cancer (CRC) is the fourth most common cancer in the United Kingdom. The five-year survival rate from CRC is only 10% when discovered at a late stage, but can exceed 90% if diagnosed early. Symptoms related to CRC can be non-specific, and therefore the decision to refer for a colonoscopy can be challenging. Breath analysis potentially offers a simple and quick method to detect CRC specific volatile organic compounds (VOCs) in breath. This protocol describes the COBRA2 study which aims to develop and validate the clinical prediction model (CPM) in the detection of CRC based on the breath test. An exploratory comparison between the breath test and faecal immunochemical test (FIT) will also be carried out to assess whether combining both tests improves diagnostic performance.

Methods: The COBRA2 study is a multicentre, case-control development and validation study. Breath samples will be collected from participants attending hospital for a planned colonoscopy (control group) or from participants with histologically confirmed colorectal adenocarcinoma (CRC group). A total of 720 participants (470 controls, 250 CRC) will be recruited. All participants will maintain a clear fluid diet for a minimum of 4-6 h prior to sampling, which will take place at outpatient clinics to avoid bowel preparation. The FIT result will be recorded where available. Breath samples will be analysed using gas chromatography-mass spectrometry to identify the VOCs present. Relationships between VOCs of interest and presence of CRC will be explored, and the CPM will be developed using statistical and machine learning methods. We will also assess whether incorporating FIT into the CPM improves diagnostic performance. The CPM will be subsequently validated in an independent sample of up to 250 participants (125 controls, 125 CRC) using the same case-control design and the potential clinical utility of decision rules for triaging will be assessed. If successful, broad validation in an unselected target population of symptomatic patients is required.

Discussion: The non-invasive breath test may provide direct patient benefit through earlier and accurate detection of CRC, and higher patient acceptability. It can help ensure timely secondary care referral, potentially translating to improved curative treatment and survival for patients.

Trial registration: The study is registered with ClinicalTrials.gov (NCT05844514).