Aga Khan University’s department of paediatrics and child health in partnership with Integration Xperts (iX) will develop an artificial intelligence (AI) algorithm on T2*CMR images, the gold standard to gauge cardiac iron overload in thalassaemia patients, a leading cause of morbidity and heart failure.
T2*CMR requires a post processing software and highly-trained personnel to interpret images, making it a costly procedure. The study team’s AI algorithm will help in countering this challenge in resource-constrained countries like Pakistan.
An inherited disorder of blood, an estimated 5000 to 8000 children are born with thalassaemia in Pakistan every year. Cardiac iron overload is a major cause of morbidity in thalassaemia caused by numerous blood transfusions and increased gastrointestinal iron absorption. According to an international survey of T2*CMR in thalassaemia major, 43 per cent of patients having T2* values less than 10 milliseconds indicated severe iron overload which is associated with heart failure and death.
“Mortality due to thalassaemia complications in Pakistan is comparable to numbers in Greece in the 1960s,” said Dr Babar Sultan Hasan, an associate professor at AKU’s department of paediatrics and child health. Thalassaemia is one of the most frequent genetic disorders in Greece and governmental programmes aiming at prevention and management have effectively brought down their occurrences over the years. “Through timely diagnosis and appropriate management, mortality due to iron overload can be brought down,” said Dr Hasan.
Beginning this fall, Dr Zahra Hoodbhoy, principal investigator of the study and an assistant professor at the department of paediatrics and child health will be partnering with Integration Xperts (iX), a group of professionals providing digital solutions through AI and Big Data, to automate MRI imaging and reporting.
The conventional method requires breath holding to acquire images and had been difficult to perform on children of ages 10 and less, and those with anaemia and heart failure. The computer-assisted diagnosis through an AI algorithm also examine if T2* values can be directly derived from free breathing images of the heart as opposed to the conventional breath-holding method.
“Our team will put together efforts that will make big differences in medicine in the future”, said Umair Azam, founder and managing partner of iX. “We believe in the democratisation of technology so that it serves the common people of Pakistan.”
The University’s research team aims to review records of estimated 800 thalassaemia patients at AKU who had T2* cardiovascular magnetic resonance imaging in the last five years. The AI models will differentiate between images and estimate iron overload.
“The study will have an impact internationally and can be upscaled to other conditions besides thalassaemia,” said Dr Devyani Chowdhury, co-investigator in the study and a paediatric cardiologist in Pennsylvania, USA.
The research team from AKU includes Dr Babar Sultan Hasan and Dr Zahra Hoodbhoy from the department of paediatrics and child health, data scientists Ilsa Khan Baqai, Yusra Shahid and Zia Saleem from Integration Xperts and Dr Devyani Chowdhury from Cardiology Care for Children in Pennsylvania, USA.