Deep Learning for EVAR surveillance

  • Research type

    Research Study

  • Full title

    Deep learning applied to plain abdominal radiographic surveillance after Endovascular Aneurysm Repair (EVAR) of Abdominal Aortic Aneurysm (AAA).

  • IRAS ID

    259034

  • Contact name

    SR Vallabhaneni

  • Contact email

    fempop@liverpool.ac.uk

  • Sponsor organisation

    Royal Liverpool and Broadgreen University Hospitals NHS Trust

  • Clinicaltrials.gov Identifier

    5851, RD&I "Intent to Sponsor" at Royal Liverpool University Hospital

  • Duration of Study in the UK

    1 years, 11 months, 31 days

  • Research summary

    Abdominal aortic aneurysm (AAA) is a condition in which the largest blood vessel in the body becomes weak and forms a bulge. If it becomes large enough, the AAA can burst often leading to death. We therefore repair AAAs before they burst. Endovascular aneurysm repair (EVAR) is a standard treatment in the majority of patients. It is a keyhole technique that reinforces the aorta with a synthetic tube called a “stent-graft”.

    EVAR is a safer option in the short-term compared to traditional open surgery. However in the long term, 1 in 5 patients require further surgery to correct problems developing with the stent-graft such as loss of position and integrity of the stent-graft. Therefore, patients are followed up for life after EVAR with scans performed, usually on an annual basis, to look for signs of a failing stent-graft.

    Stent-grafts are visible on x-rays of the belly and by comparing series of images taken over time, it is possible to diagnose a myriad of stent-graft problems including loss of positioning, disintegration of the stent-graft material as well as stent-graft distortion. But these changes can be subtle and difficult to spot, even to the trained human eye. As a result, patients undergo more detailed scans that unfortunately carry a risk of kidney damage and radiation-induced cancer.

    Our study will explore the use of artificial intelligence in interpreting series of anonymised x-rays to identify features of a failing stent-graft. A deep-learning algorithm will be applied to post-EVAR x-rays that have been performed as part of standard care at our institution over the last 13 years.

    We wish to examine if deep learning program can be as good as or even better than human interpretation.

  • REC name

    North West - Liverpool Central Research Ethics Committee

  • REC reference

    19/NW/0311

  • Date of REC Opinion

    5 Jul 2019

  • REC opinion

    Further Information Favourable Opinion