Over a million developers have joined DZone.
{{announcement.body}}
{{announcement.title}}

Supercomputers and AI Improve MRI Scanning

DZone's Guide to

Supercomputers and AI Improve MRI Scanning

The ability to capture biomedical data in near real-time can open up a whole new world of actionable insights and prompt a new era of science as a service.

· Big Data Zone
Free Resource

Access NoSQL and Big Data through SQL using standard drivers (ODBC, JDBC, ADO.NET). Free Download 

MRI scans are a common tool in the armory of the modern doctor. Such scans are usually very expensive and can take days to generate and interpret. If further scans are required, the patient needs to come back in and go through the procedure all over again.

A new real-time analysis system that is powered by supercomputing promises to change things.  The system, which is developed by a consortium led by researchers at Texas Advanced Computing Center (TACC), is an automated platform that can perform in-depth MRI analysis within minutes, thus enabling further scans to be performed whilst the patient is still in the scanner.

The system combines the imaging of a Philips MRI scanner with the computational grunt of the Stampede supercomputer, with an API developed by TACC supporting communication and data transfer between the two.

“The Agave Platform brings the power of high-performance computing into the clinic,” the team say. “This gives radiologists and other clinical staff the means to provide real-time quality control, precision medicine, and overall better care to the patient.”

Put to the Test

The group put the technology through its paces by scanning a patient with cartilage disorder to assess the disease. Data from the MRI was sent via a proxy server to Stampede, where it was parsed through the GRAPE (GRAphical Pipelines Environment) analysis tool. This allowed the scanned tissue to be analyzed, with the clinician then instructed to explore certain areas in more depth.

A T1 mapping process was used, whereby raw data was converted into something useful in terms of imagery. It’s the kind of intensive computing that is a good marker for the kind of work that can be performed on a regular basis.

The whole process, from the original scan to meaningful insights returned, took around five minutes, with no extra input required from staff. It’s currently setup to alert the operator if a scan was corrupted (such as if the patient moves) or to prompt for additional scans if required.

“We are very excited by this fruitful collaboration with TACC,” the team say. “By integrating the computational power of TACC, we plan to build a completely adaptive scan environment to study multiple sclerosis and other diseases.”

Further:

“Another potential of this technology is the extraction of quantitative, information-based texture analysis of MRI,” they continue. “There are a few thousand textures that can be quantified on MRI. These textures can be combined using appropriate mathematical models for radiomics. Combining radiomics with genetic profiles, referred to as radiogenomics, has the potential to predict outcomes in a number diseases, including cancer, and is a cornerstone of precision medicine.”

The ability to capture biomedical data in near real-time can open up a whole new world of actionable insights and prompt a new era of science as a service.

“Here, we demonstrated this is possible for MRI. But this same idea could be extended to virtually any medical device that gathers patient data,” they conclude. “In a world of big health data and an almost limitless capacity to compute, there is little reason not to leverage high-performance computing resources in the clinic.”

The fastest databases need the fastest drivers - learn how you can leverage CData Drivers for high performance NoSQL & Big Data Access.

Topics:
big data ,ai ,machine learning ,healthcare

Published at DZone with permission of Adi Gaskell, DZone MVB. See the original article here.

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}