Researchers at Aga Khan University and the University of Virginia are collaborating on an innovative project that will harness the power of artificial intelligence to understand a particularly complex disorder of the intestine, environmental enteric dysfunction (EED). EED often referred to as a neglected disease of poverty, is widespread among children in low-income countries where the population is exposed to contaminated water and poor sanitation.
EED hinders the gut’s ability to absorb essential nutrients compromising children’s growth potential and leaving them vulnerable to a range of diseases. Data scientists have already demonstrated how “intelligent” computers can outperform experienced radiologists and pathologists in detecting signs of disease in x-rays and biopsies. Dr Sana Syed, an assistant professor in paediatrics at the University of Virginia and Dr Asad Ali, associate dean for research at Aga Khan University, are presently applying “deep learning”, a type of artificial intelligence, to train a computer programme and analyse microscopic images of tissue located deep inside the small intestine. The initiative, funded through an Engineering in Medicine grant from the University of Virginia (UVa), will be conducted in collaboration with the Data Science Institute at UVa.
The project will see computers break down the size, shape and structure of images of the intestine’s cells into a matrix of numbers. Every number corresponds to a pixel “the smallest unit of an image” and as the programme scans more of these images, it becomes alert to abnormal patterns. Eventually, the computer will learn to compare images of healthy intestines to those affected with EED and to pinpoint the differences at the cellular level that trigger the disorder. The images of intestines affected by EED being studied come from work in SEEM, a USD $13m multi-country grant funded by the Bill and Melinda Gates Foundation.
In the longer-term, Dr Syed and Dr Ali believe that these insights could also transform the way doctors diagnose EED. At present, the only way to conclusively identify the disease is through a biopsy, an invasive procedure that involves extracting tissue samples from a person’s intestine. Researchers aim to use the insights from their work to create a comprehensive set of screening biomarkers “chemical warning signs” that would help future clinicians diagnose EED through a simple blood or urine test.—APP