Yaqub finds the “magic” of AI in healthcare

Wanting to help people is the reason Dr. Mohammad Yaqub journeyed into artificial intelligence and healthcare, and his passion for helping doctors with advanced diagnostic tools is inspiring the next generation of AI students. 

I felt like healthcare was the right place for me as I wanted to do something in smart imaging and wanted to make a difference.

An Assistant Professor in Computer Vision at MBZUAI, Yaqub completed his Ph.D. in biomedical engineering at the University of Oxford. In more than a decade of research and development, he has certainly found the “magic” of AI in healthcare. Yaqub leads the university’s Bio Medical Image Analysis (BioMedIA) lab, which focuses on investigating real-world healthcare problems using AI.

AI infancy and healthcare

“I thought: studying cybersecurity is good, but it’s not great for me,” he said. “I felt like healthcare was the right place for me as I wanted to do something in smart imaging and wanted to make a difference.”

“During my undergraduate in computer science, artificial intelligence wasn’t a big thing. In the early-2000s, AI was a magic word that we used to think about, but we couldn’t see the effect because the algorithms were not mature enough, and we didn’t have computational power at the time.”

A textbook by American computer scientist and founder of the department of machine learning at Carnegie Mellon University, Tom Mitchell, titled Machine Learning (1997) catapulted Yaqub into the path he is now on.

“It’s actually the book of machine learning models and if you want to go back and read about the very classical machine learning algorithms, you go to that,” he said. “He started to think there’s intelligence. That there are smarter tasks that you can make a computer do.”

Fetal and women’s health

Yaqub stayed with Oxford University for his fellowship to work alongside Professor Alison Noble, who he was supervised by during his Ph.D. During this time, the research team developed an AI solution which is now helping millions of women around the world called ScanNav.

During pregnancy, a very important scan happens at 20 weeks called the anomaly scan, which is key to assessing fetal growth and checking for fetal abnormalities such as heart conditions and spina bifida. The aim of this technology is to help support sonographers during the acquisition of ultrasound images and the assessment of fetal organs, which is likely to improve the detection of anomalies and thereby reduce them through early intervention.

“We did a small clinical audit on Oxford University Hospital – one of the best hospitals – yet we found some issues which we thought we could help tackle,” Yaqub said.

“Getting the right images during an anomaly scan is a very challenging task; you have a moving fetus who is about 10 to 20 centimeters long and you are required to acquire many images and perform measurements. This all needs to be done in a limited amount of time. In some hospitals, even in developed countries, they’re only given 20 to 30 minutes to complete the task,” he continued.

ScanNav doesn’t replace this process, however, it automates the image analysis and ensures adherence of the acquired images to clinical guidelines. ScanNav is now exclusively used in GE Healthcare Voluson SWIFT ultrasound machines. As the regulated first-AI based fetal anomaly system, Yaqub estimates it can help up to 10 million women each year.