I know there's more than one measure of accuracy in medical studies.
I forget the jargon, but there's always a measure of false-negativeness and a measure of false-positiveness. While they do have to admit a false negative, "some afflicted lacked the indicator", they may be able to say, "100% of healthy patients lacked the indicator". Ergo if you ever found it, you just made an inarguable diagnosis.
So maybe a reporter didn't grok the stat.
The false negatives are curious, though - some small percentage of the population has diseased brain tissue but clean fluids, as if their braincar had a great new oil filter.