This study aims to determine whether a breath test could be used for early detection of hepatic fibrosis, cirrhosis or hepatocellular carcinoma.

Patients who are attending for a planned liver outpatient services or investigations will be approached to provide a breath sample.

Multi platform mass spectrometry analysis will be performed to establish volatile biomarkers that can discriminate between fibrosis, cirrhosis and hepatocellular carcinoma.

Volatile organic compound (VOCs) emitted from the human body have been of interest to researchers for several decades. In 1971, Pauling et al reported that breath and urine contained approximately 250 and 280 VOCs respectively in normal human subjects. VOCs are traditionally measured in food industry and counter-terrorism. Routine clinical applications include the alcohol breathalyser, 13-C urea breath test for H. pylori, nitric oxide in asthma, and hydrogen/methane test for bacteria overgrowth.

Preliminary work has suggested that there is alteration in the VOC profile of patients with liver disease. It is suggested that as fibrosis increases, liver function is impaired, resulting in increased levels of certain compounds. These compounds are then present at higher levels within the blood stream and when they reach the lungs, they are exhaled. There has also been work demonstrating a change in the profile of VOCs within cancers, demonstrating specific "breath signatures" in patients with oesophageal cancer. At Imperial College London, Marker et al have developed a breath test as screening test for oesophageal cancer with comparable sensitivity and specificity to other screening methods (e.g. faecal occult blood testing). The investigators would aim to achieve similar with hepatobiliary disease via a prospective cross-sectional study aimed at determining the diagnostic accuracy of exhaled breath volatile organic compounds (VOC) for the detection of liver disease.

Patients will be recruited from outpatient clinics and hospital wards based across Imperial College Healthcare Trust. They may also be recruited from the radiology department when attending for liver biopsies or hepatocellular carcinoma surveillance ultrasound scans. When attending for their appointment, a member of the research team will discuss the research project with the patient, explain that is requires the patient to provide a breath +/- urine sample. Patients will be provided with a patient information leaflet. If the patient is to agree to this, they will be asked to sign a written consent form. If a recent PHES test has not been performed, one will be conducted otherwise no additional investigations, invasive or otherwise will be required of the patients.

For patients under regular follow up within the trust, they may be asked to consent to further breath samples being taken at future appointments to facilitate longitudinal follow up. Patients will be entitled to withdraw at any point.

Patients attending for endoscopy, fibroscan or liver biopsy will have been nil by mouth for 4-6 hours prior to the procedure as per local guidelines. For those attending outpatient appointments, if identified prior to their appointment by their clinical team, the investigators will aim to contact the patients and request that they fast. Should they not be fasted, they may be asked to perform basic oral hygiene techniques (e.g. rinsing mouth with water), prior to the breath sample being taken. A food diary of the preceding 24 hours will be taken as part of the pre-sampling questionnaire.

As part of the consent process, patients will also be asked permission to access their hospital records for the purpose of the research study only, including their blood tests, radiology and pathology results. All data will be anonymised and all hospital records shall be handled with strict confidentiality in accordance with the Data Protection Act 1998 and the General Data Protection Regulation (GDPR).

Patients will be asked to give a number of exhalations into a breath sampling device until a defined volume of breath has been sampled. This may be done via one of several techniques. The investigators preferred option is that they breathe into a custom designed nalophan bag which is attached to thermal desorption tubes and a breath sampling device. Alternatively, patient may be asked to perform normal tidal breathing whilst wearing a face mask, which takes around 2 minutes on average. Patients located at St Mary's Hospital may also be asked to breathe directly into a mass spectrometer machine (SYFT Voice200Ultra). This mass spectrometer instrument is a self-contained mobile device which can be moved between clinical areas within St Mary's Hospital. Sampling by this method will involve a single deep nasal inhalation followed by complete exhalation via their mouth into a plastic cylinder connected to a tube leading to the mass spectrometer itself. This is repeated three times within a 60 second window. The investigators expect the breath sampling process to take no longer than 5-10 minutes in total. Patients will be requested to pass urine into a standardised 60ml urine specimen vial, which will be immediately sealed. 20ml of urine will be aliquoted into a standard 60ml specimen vial.

The VOCs from the breath will be trapped onto adsorbent tubes inserted into the breath taking device. After sampling the adsorbent tube is removed from the device, capped and transported back to the laboratory at St Mary's Hospital London for analysis. Samples will be transported by courier or by a member of the research team. The sample(s) shall undergo analysis via mass spectrometry instruments, at St Mary's Hospital London.

Breath VOCs sampled on adsorbent tubes will be analysed within 48 hours of collection. The sample constituents are destroyed within the mass spectrometer for detection meaning that no clinical material will remain after analysis by the mass spectrometry instruments. Urine samples will be stored within a registered tissue bank in accordance with the Human Tissue Act and disposed of following analysis.

As part of the study, the investigators will aim to follow patients up longitudinally to ascertain if a participant's VOC profile changes with progression or regression of liver disease. At each appointment, before sample is taken, the participant's consent will be reviewed and if the participant is still willing to continue in the study, further breath and urine samples will then be collected. This would be for a maximum of 24 months.

Inclusion Criteria:

• Patients seen in secondary care with suspected or confirmed liver / pancreatic disease or hepatocellular carcinoma / pancreatic cancer
• Patients able to understand and retain the information provided, thereby being able to give informed consent for inclusion in this study

Exclusion Criteria:

• Patients who lack capacity or unable to provide informed consent.
• Any patient outside the established age range (18-90years).
• Patients unable to fast.

Patients will be identified and recruited from Imperial College Healthcare NHS Trust inpatient admissions, routine outpatient hepatobiliary clinics, the radiology departments (e.g. for patients undergoing liver biopsy or imaging of their liver) and endoscopy units (e.g. liver disease patients undergoing endoscopic surveillance for varices). Patients with hepatocellular carcinoma may also be recruited from oncology outpatient clinics.