Computer software could help doctors predict key stroke treatment outcome


2014-04-23 23:30

Predicting which patients might suffer serious and even fatal consequences from clot busting stroke treatment could be easier for clinicians thanks to airport technology. A computer program that has been developed by researchers at Imperial College London has the capability to recognize subtle changes in the brain of stroke victims using pattern recognition software similar to that used in airport security and passport control.

The findings, published in the journal Neuroimage Clinical, make it easier for doctors to assess which patients are the best candidates for thrombolytic therapy that dissolves clots that cause ischemic stroke by interfering with blood flow to the brain.

Software detects subtle brain changes

Subtle brain changes can mean increased risk of bleeding for patients given stroke treatment with medications that dissolve blood clots.

Six percent of people with stroke will suffer more harm than good from the treatment that is injected into the blood vessels to thin and break up small clots. The result for those patients means bleeding in the brain, more disability and the potential for dying.

The new software detects narrowing and thinning of the blood vessels that can predict bleeding that are subtle and could go undetected.

Dr. Paul Bentley from the Imperial College of London Department of Medicine and lead author of the study explains clinicians have to weigh the risks and benefits of stroke treatment for every patient.

"Intravenous thrombolysis carries the risk of very severe side effects for a small proportion of patients, so having the best possible information on which to base our decisions is vital. Our new study is a pilot but it suggests that ultimately doctors might be able to use our pattern recognition software, alongside existing methods, in order to make more accurate assessments about who is most at risk and treat them accordingly."

The study

For the pilot study the researchers looked at CT scans from 116 patients who had been given thrombolytic therapy to treat ischemic stroke in Charing Cross Hospital. Included in the sample was 16 patients who developed serious bleeding in the brain that causes hemorrhagic stroke.

Three independent experts reviewed the brain scans without knowing the outcome of treatment. The scans were also assessed by a trained computer program developed to detect blood vessel abnormalities associated with increased risk of bleeding.

The computer was able to predict risk of bleeding with 74 percent accuracy compared to 63 percent using standard human interpretation. The software also identified which patient developed brain bleeding based on ten scans with a 56 percent accuracy rate, compared to thirty-one percent of the time using the standard approach.

The software might also be used to help patients who are not offered the clot busting drugs based on current guidelines. Just twenty percent of patients are considered candidates. Treatment decision depends on time of arrival to the hospital after symptoms occur. Patients with severe stroke are also excluded. The software has the potential to help doctors identify patients who would still benefit from the drugs.

The next step Bentley said is to carry out a much larger study to fully assess the potential of the software. The new development could mean better outcomes for 15 million people worldwide who suffer stroke each year.

Image credit: Wikimedia Commons