Artificial intelligence may help reduce gadolinium dose in MRI

According to a study presented today in the annual meeting of the Radiological Society of North America (RSNA), researchers are using artificial intelligence to reduce the dose of an opposite agent in the body after the MRI examination.

Gadolinium is a heavy metal that is used in Contrast material that enhances images on MRI. Recent studies have found that trace amounts of metal remain in the body of those who have examined some type of gadolinium.

 The effect of this statement is not known, but the radiologist preserving critical information is working continuously to optimize the patients, which provide a gadolinium-enhanced MRI scan.

Anhão Gong, chief of the study, Anhão Gong, a researcher at Stanford University in Stanford, California, said, "There is concrete evidence that there is gadolinium in the brain and body.

read more: Xiaomi Redmi Note 7 comes with water-sealed ports and buttons: Lei Jun reveals

While its implications are unclear, while potential patients are reducing risks. It is imperative to maximize the clinical value of the MRI exam. "

Dr. Gong and Stanford colleagues are studying deep learning as a way to achieve this goal. Deep learning is a sophisticated artificial intelligence technique that teaches computers through examples. 

Through the use of a model called homogeneous neural network, the computer can not only recognize images but also can find a subtle difference between imaging data that a human supervisor may not be able to be sensible.

In order to train an intensive learning algorithm, researchers used MR images from 200 patients who had received opposite-enhanced MRI examination for various types of signals. 

They collected three sets of images for each patient: Pre-opposed scans, referenced in contrast to the opposite administration and referred to as a zero-dose scan.

 Low-dose scan, acquired after 10 percent of the standard gadolinium dose administration; And complete dose scan, 100 percent dose administered after administration.

The algorithm learned to estimate full-dose scan with zero-dose and low-dose images. After this, neuroradiologists evaluated the images for contrast enhancement and overvalue quality.

Results showed that image quality was not quite different between low-dose, algorithm-enhanced MR images, and full-dose, contrast-enhanced MR images.

 Early results demonstrated the ability to make full-dose, contrast-enhanced MR images equal to the use of any opposing agent.

These findings Dr. According to Gong, instead of sacrificing the clinical quality, dramatically suggest the ability of the method to reduce the dose of gadolinium.

"The images of low-dose gadolinium provide significant unexpected clinically useful information, which is now accessible using intensive education and AI", he said.

Now, when researchers have shown that the method is technically possible, then they want to study it further in clinical settings, where Dr. Gong believes that it will eventually find a house.

In future research, a wide range of MRI scanners will include evaluation of algorithms and various opposite agents.

"We are not trying to replace existing imaging techniques," Dr. Gong said. "We are trying to improve it by looking at it for the safety of our patients and generating more value than current information."