Google-like data bank of kids' brain scans could aid docs

The goal of the cloud-computing project is to let physicians and radiologists "google" a brain scan to find patients with similar abnormalities, review their anonymous medical records, and improve both diagnosis and treatment.

Say a doctor orders an MRI scan of a child's brain to try to determine what might be at the root of a list of troubling symptoms.

She eyeballs the results to look for abnormalities that might indicate certain diseases or disorders, but nothing seems terribly amiss. So she submits the scan anonymously to a database that includes thousands of other scans of children with healthy and abnormal brains to find matches. She then gets the medical records -- anonymously, of course -- of kids with similar scans and voila, she makes a diagnosis that involves a lot less guesswork than if she'd used her eyes and knowledge alone.

Michael I. Miller, a biomedical engineer and director of the school's Center for Imaging Science, is a lead investigator on the project. Peter Howard/Johns Hopkins University

Such is the goal of a cloud-computing project being developed by engineers and radiologists at Johns Hopkins University.

By collecting and categorizing thousands of MRI scans from kids with normal and abnormal brains, they say the resulting database will give physicians a sophisticated, "Google-like" search system to help find not only similar pediatric scans but the medical records of the kids with those scans as well. Such a system could help not only enhance the diagnosis of brain disorders, but the treatment as well -- perhaps before clinical symptoms are even obvious to the naked eye.

"If doctors aren't sure which disease is causing a child's condition, they could search the data bank for images that closely match their patient's most recent scan," Michael I. Miller, a lead investigator on the project who also heads up the university's Center for Imaging Science, said in a news release. "If a diagnosis is already attached to an image from the data bank, that could steer the physician in the right direction. Also, the scans in our library may help a physician identify a change in the shape of a brain structure that occurs very early in the course of a disease, even before clinical symptoms appear. That could allow the physician to get an early start on the treatment."

Susumu Mori, a radiology professor at the Johns Hopkins School of Medicine and co-lead investigator on what he calls the "biobank," says that a collection of brain scans of this size will also help neuroradiologists and physicians identify specific malformations far faster than is currently possible. It's sort of like the difference between using a library's card catalog, where for starters you had to know how to spell what you were looking for, and typing a few words into Google to instantly review a long list of results -- often despite a misspelling.

Mori has spent the past four-plus years working on a clinical database of more than 5,000 whole brain MRI scans of children who've come through Johns Hopkins -- a project that involved indexing anatomical data on 1,000 structural measurements in 250 brain regions that were ultimately sorted into 22 brain disease categories, including infections, psychiatric disorders, epilepsy, and chromosomal abnormalities. The new data bank will not only help recognize established pediatric disorders, but possibly even help identify and classify new diseases, too. (Researchers are also working on a similar project collecting scans of elderly patients to focus on age-related diseases.)

While the project involving pediatric scans -- made possible by a three-year $600,000 grant from the National Institutes of Health -- is at least in its pilot stage available only to physicians and patients within the Johns Hopkins medical system, the researchers say it could open up and expand to other networks in the coming years. Such an expansion would presumably benefit not only other physicians and patients, but the database itself.

Miller goes into his research further in the video below.

Autoplay: ON Autoplay: OFF