Gadolinium is a heavy metal, it is toxic in high doses, but in patients with normal kidney function 90 percent of it will be eliminated via urine within 24 hours. That means, "detoxify" salespeople will argue, that 10 percent remains and it can be harmful and they mumble stuff about chelation, and Cutter law firm will happily file a suit they hope to settle quickly. There were a few instances of reactions in the past, in people with severe kidney disease, but both the U.S. Food and Drug Administration and European Medicines Association found no issue. Still, Europe being Europe they banned a few of them anyway, and after a public hearing FDA included a warning and listed concerns for people at risk (which Gena Norris would not have been - at least not from anything a hyperbaric chamber and a Chinese folk medicine practitioner could cure.
The benefits of MRI are many, the risks are minute, but there's no downside to optimization
The last thing our public health system needs is people culturally inclined to distrust medicine (take a compass and draw a circle around a Whole Foods store) to not get MRIs until it's too late. Right or wrong, government can force changes and since "the dose makes the poison" the dose of gadolinium is an ideal use of artificial intelligence, which Enhao Gong, Ph.D., of Stanford University and colleagues have done in work they showed at the meeting of the Radiological Society of North America.
Using a convolutional neural network, they fed in three sets of images from 200 patients who had received MRI exams with gadolinium: zero-dose scans (no gadolinium) low-dose (10 percent of the standard dose) and scans with the recommended gadolinium, and found that after optimization by the "artificial intelligence" there was little difference between the full dose and low dose.
Optimization will not quiet down the trace chemicals "endocrine disruptor" or homepathy fringes who insist gadolinium at any level can be harmful, but at least some science is still happening despite the chaos.