Researchers are testing a database of more than 10,000 pediatric MRI scans of Johns Hopkins patients to help physicians and researchers better detect brain anomalies in children.
The idea behind the project is to create a Google-like search system that matches a patient’s MRI scan with similar scans in the database. After obtaining the matches, a physician can request the de-identified medical records of those patients.
“Our technologies allow you to group children based on their MRI scans,” says biomedical engineer and Center for Imaging Science Director Michael Miller, who is leading the project along with Susumu Mori, professor of radiology, and Thierry Huisman, director of imaging and imaging science at Johns Hopkins Bayview Medical Center.
Mori says the team faced several challenges in creating the database, including converting the MRI data into meaningful information that can be understood by computers. “We have to teach computers what is important and where the important structures in the brain are,” notes Mori.
For the latter, the team of engineers and radiologists used Johns Hopkins-developed software called Brain GPS to analyze each brain image. The software converts the images into bar code-like data and then organizes the information based on anatomical abnormalities related to 22 major brain disease categories, including chromosomal abnormalities, congenital malformations, vascular diseases, infections, epilepsy and psychiatric disorders.
The database, which can only be accessed by Johns Hopkins physicians and researchers via several institutional supercomputers and from behind the Department of Radiology and Radiological Science’s firewall, is now being tested by radiologists for its accuracy.
It continues to have new pediatric brain MRI scans and related patient information added every two weeks.
The project was supported by a recently completed three-year $600,000 grant from the National Institutes of Health and is now being funded by a grant from the National Institute of Biomedical Imaging and Bioengineering.
