Artificial Intelligence Can Detect Brain Blockages and Reduce Time to Treatment

Radiologists can now use a sophisticated type of artificial intelligence to quickly identify artery blockages, also known as large vessel occlusions (LVO). With this new deep learning model with CT angiography, it can help rapidly detect these clots, which can potentially accelerate treatment.

Large vessel occlusions are clots that block the supply of oxygenated blood to the brain and are usually detected in a significant proportion of ischemic strokes, the most common types of strokes.

A person who experiences LVOs loses an average of 1.9 million neurons per minute. This is the reason why prompt diagnosis is critical in this time-sensitive situation. The recanalization or opening of the blocked artery through a treatment known as endovascular therapy can begin. The higher the chances of extending the patient's disability-free life by a week. According to Matthew T. Stib, M.D., a radiology resident at the Warren Alpert Medical School at Brown University, who leads the research team in the article published in Radiology.

LOOK INTO: Professional Photographer Shares Story of Cancer-Stricken Daughter During Pandemic

Both Johanna M. Ospel M.D., radiology resident from the University of Basel in Switzerland, and Mayank Goyal, M.D., Ph.D., director of imaging and endovascular treatment in the Calgary Stroke Program in Canada, agreed that having a reliable artificial intelligence tool to detect these LVOs is critical in initiating treatment and facilitating and streamlining image interpretation.

CT angiography (CTA) is the current gold standard for detecting these blockages while providing a detailed view of the blood vessels. This a three-minute exam where radiologists can accurately identify and interpret these large vessel occlusions. But because radiologists are not always available, depending on the time and location or any backlogs at the hospital, it can delay care needed by the patient, resulting in dangerous patient outcomes.

To avoid these types of delays in patient care, Dr. Stib and his team worked closely with Brown's computer science department to create a deep learning model from scratch. The goal was to rapidly and accurately detect LVOs on CTA and reduce the time to treatment.

(Photo: pixabay)

How Artificial Intelligence Works

During this multi-center retrospective study, they used the CTA scans from 540 adults with suspected ischemic strokes to train the algorithm in recognizing these large vessel occlusions and differentiate it from other conditions.

The researchers used a newer approach called multiphase CTA, which can provide more comprehensive information than the single-phase technique.

The test set includes 62 patients, with an average age of nearly 70 years old. When the researchers applied the deep learning model on multiphase CTA examinations,  the model could detect all 31 large vessel occlusions for a sensitivity of 100 percent. Dr. Stib found the results promising, and according to him, the model's sensitivity must be optimized to make sure all cases were picked up as missing even a single case will have dire consequences.

This is the first study to make use of multiphase CTA to look at occlusions in both anterior and posterior arteries. The researchers clarified that this algorithm is not meant to replace radiologists but rather speed up the time to diagnosis.

MORE ON: New Research Addresses Women's Lost Interest in Sex as They Age 

Check out more news and information on Neurology on MD News Daily.