Donate to Science & Enterprise

S&E on Mastodon

S&E on LinkedIn

S&E on Flipboard

Please share Science & Enterprise

A.I. Image Analysis Computes Stroke Damage

e-Aspects software

Data from e-Aspects software displayed on a smartphone. (Brainomix)

10 Oct. 2019. A review study shows an analysis of CT scans using artificial intelligence matches calculations of damage from stroke made with more advanced imaging. Results of the study testing a technology developed by Brainomix, a company in Oxford, U.K., appear in the 30 September issue of the International Journal of Stroke.

Stroke occurs when blood flow to the brain is interrupted, cutting the oxygen needed by brain cells to function. The vast majority (85%) of strokes are caused by blood clots, while many other strokes are caused by blood vessel leakage in the brain. Recovery, often in rehabilitation clinics, can take months or years of continuous exercises. World Stroke Organization says 1 in 6 people worldwide will have a stroke in their lifetime, with some 15 million people suffering a stroke each year.

A team from institutions in Germany and the U.K., led by neurologist Simon Nagel at University Hospital in Heidelberg, Germany, seeks to confirm earlier findings of a technique using artificial intelligence to measure the extent of damage from stroke from blood clots, called ischemia, from an analysis of computed tomography or CT scans. While non-contrast CT scans can provide an initial assessment of stroke damage, these evaluations normally require advanced CT or MRI imaging to provide more precise and stable damage measurements.

Brainomix, a spin-off enterprise from University of Oxford, develops the e-Aspects system, with Aspects standing for Alberta stroke program early CT score. Aspects is a 10-point standard quantitative measure of early indicators of restricted oxygen supply to the brain, derived from non-contrast CT scans. Results of an Aspects score can provide a fast assessment of stroke damage, in circumstances where therapy decisions often need to be made very quickly. Brainomix says its e-Aspects software uses deep-learning algorithms to analyze CT scan images and return an Aspects score in less than a minute, with results transmitted to doctors’ desktop or laptop computers, as well as smartphones.

Nagel and colleagues reviewed results of three previous studies where results of non-contrast CT scans analyzed using Brainomix software with 388 stroke patients were measured against evaluations of the same patients using more sophisticated MRI or CT perfusion images. The technologies measured the volume of stroke damage, with clinical outcomes calculated with the National Institute of Health Stroke Scale of severity at the initial diagnosis, then after 45 and 120 days.

The findings show measurements of stroke damage using the e-Aspects system correlated moderately with stroke severity assessments using the NIH scale. The e-Aspects results, however, correlate more strongly with a measure of lesions detected with the images, and predicted clinical outcomes.

“The study results show that e-Aspects volume provides a robust assessment of the early brain injury in stroke patients using only non-contrast CT scans, without the need for difficult-to-access advanced imaging,” says George Harston, chief medical officer for Brainomix in a company statement. “This is vitally important, as the vast majority of hospitals around the world do not have advanced imaging in their Emergency Departments.” Harston is also a stroke physician at Oxford University Hospitals.

The e-Aspects software is part of a package of e-Stroke software offered by Brainomix, including modules for large blood vessel blockages, and analysis of MRI and CT perfusion images.

More from Science & Enterprise:

*     *     *

1 comment to A.I. Image Analysis Computes Stroke Damage