Automatic Measurement of Abdominal Aortic Diameters
Research type
Research Study
Full title
Automatic Measurement of Abdominal Aortic Diameters
IRAS ID
308123
Contact name
Sean Bayly
Contact email
Sponsor organisation
University of Kent
Duration of Study in the UK
2 years, 0 months, 1 days
Research summary
Abdominal Aortic Aneurysm (AAA) is a treatable condition in which the wall of the descending aorta begins to swell. Unfortunately, the condition is often overlooked or not reported during routine procedures. Left untreated, the aorta is likely to rupture resulting in a medical emergency from which only 10% of patients are expected to survive. We aim to work towards an AI solution for automatically detecting the presence and severity of existing aneurysms in this region.
Considering parameters such as disease stage/severity and patient data such as age, gender, ethnicity, medical
history effectively transforms each patient into a big data challenge. Machine learning has been used to tackle this big
data challenge, however, this requires large datasets.Our proposal is to develop a novel data processing pipeline that extracts information from axial image slices of the abdominal aorta. This is subsequently used to detect and quantify AAA.
we will use a few-shot, deep learning, algorithm to segment regions of interest, located in images slices derived from several imaging modalities including CT/CT-Angiogram (CT-A) and MRI. The diameter of the segmented regions will be measured to determine the presence and severity of the aneurysm(s).
Also of interest is the use of Super-Resolution (SR) techniques to increase the resolution of input data while minimizing artifacts. This is to gauge the impact of SR on segmentation accuracy.
REC name
East of England - Cambridge East Research Ethics Committee
REC reference
22/EE/0088
Date of REC Opinion
5 Sep 2022
REC opinion
Further Information Favourable Opinion