Lights, camera, heartbeat

For most patients, not moving an inch is a big part of having an MRI done. That's problem for doctors and researchers who want to take a look at something precisely when it is moving: a heart mid-beat, or a knee mid-bend.

Now, a team of researchers from Biomedical NMR Research Inc. and the Max Planck Institute for Biophysical Chemistry in Goettingen, Germany, have found a way to take a magnetic resonance imaging 'video' of moving joints and organs by reducing the time required for recording images to just one fiftieth of a second.

The team explained its findings in a paper published late last week in the journal, NMR in Biomedicine.

"What we can now do is look at the heart in real-time, heartbeat by heartbeat," said Jens Frahm, who led the research, in an interview with Deutsche Welle. Current technology allows doctors to analyze multiple images to create a kind of 'average' video of a heart beating, but in patients with conditions like arrhythmia, doctors may not be able to get a clear idea of irregularities in the heartbeat.

"The main limitation is that MRI has to acquire data several times, different kinds of projections of an image in order to make a full two-dimensional image reconstruction possible," said Frahm, who is also the director of Biomedical NMR Research Inc., an independent, non-profit research group attached to the Max Planck Institute.

"So it's not just a single snapshot of a few milliseconds – instead, we have to repeat our experiments several times with different spatial encodings to make this a full image."

Math holds the key

Frahm's team was able to speed up the process by developing a new mathematical reconstruction technique that could calculate a meaningful and detailed image using only about 10 percent of the data required in a traditional MRI. That means the imaging can be done about 10 times faster, reducing the time required to record images to 20 milliseconds.

"This sounds like a wonder, but this is novel mathematics," he said.

Other scientists expressed cautious optimism about this new technique.

"I think this is an exciting development," said Dr. Mushabbar Syed, a cardiologist and director of cardiovascular imaging at Loyola University in Chicago, in the United States.

"Currently there are real-time MRI [technologies] available," he said. "But the real-time is fairly rudimentary, it doesn't have very good resolution."

Dr. Syed, who was not part of the study, said the research could be especially useful in patients for whom conventional breath-hold MRIs are unsuitable. Frahm added it could also have non-medical applications.

"We can see the blood flow as it flows into the heart. You can see the jet stream into the left ventricle of the heart, and reverse, so you can see directly from the images turbulent flow phenomena," he said.

"This is fantastic in the heart, but you can also use this to study turbulent flow in a phantom, or in some other natural science environment."

Not ready for the market

The computational effort required by the new method is so immense, however, that Frahm estimates that the technology will not be widely available for another two to three years. A one-minute examination of the heart -- which typically involves between 2,000 and 3,000 images -- took the Goettingen researchers half an hour to process.

A Siemens Healthcare spokesperson confirmed that the company is cooperating with the researchers to find ways to bring the technology to market.

Author: Sophie Tarr
Editor: Cyrus Farivar