Mo1336 HIGHLY SENSITIVE DIAGNOSIS OF IRRITABLE BOWEL SYNDROME VIA NON-INVASIVE METHODS

Background: Irritable bowel syndrome (IBS) is a common and difficult to diagnose condition. Experts advise use of the Rome IV
criteria, which can provide a positive diagnosis, but have poor sensitivity. More commonly, IBS is a diagnosis of exclusion after
patients undergo colonoscopy to exclude organic disease. There is a need for a new approach that provides certainty to the vast
majority of IBS patients. We previously achieved proof-of concept for use of bowel sound features to allow differentiation between
IBS and healthy individuals and between sufferers of IBS and inflammatory bowel disease. Previous research employed basic
piezo based sensors to collect the sounds. Here we describe a larger study with clinic-appropriate hardware based on Micro-
Electro-Mechanical systems (MEMS) microphones for noise cancellation and bowel sound collection.
Methods and Results: Our study had a case-control design. We recruited 205 study participants with an existing clinical diagnosis
of IBS or healthy (asymptomatic) digestive systems. We recorded bowel sounds for two hours post-fasting, and then for 40 minutes
after a standard meal, and asked patients about the frequency of symptoms. Using the first 140 participants we built and optimised
a logistic regression based model to characterise the two groups. The optimised model was based on 13 sound features related to
motility and the migrating motor complex and two symptom features. Internal testing of the model via leave one out cross-validation
gave 93% accuracy for IBS diagnosis. Independent testing (data analyst blind to participant condition) using the next 33 IBS and 32
healthy participants revealed 94% sensitivity, 88% specificity and 91% accuracy for IBS diagnosis.
Conclusions: These results provide great promise that combined use of symptoms and sound analysis may provide a more
sensitive, cost-effective and less-invasive approach to IBS diagnosis. It offers a positive diagnosis with much greater sensitivity than
symptom based diagnostic criteria. The test could be combined with limited laboratory tests to screen for organic disease. However,
our next step is to gather more sound data with the new hardware that allows us to broaden the approach to differentiation between
IBS and organic gastrointestinal diseases. We will then conduct a multi-centre prospective study to confirm our findings. Our belief
is that the test will be especially valuable when used in primary care where it would offer greater confidence to the patient and
clinician, whilst reducing the burden of unnecessary colonoscopies.