I think the conclusion they're eluding to in the article is that: "if MRI says 99% of people have abnormalities, MRI is not trustworthy".

Not "MRI is not trustworthy" but "abnormalities are not harmful". ("Allude", by the way; to "elude" is to escape.)