This is nearly certain to be oversold. The etiology and composition of brain tumors is highly diverse; they don’t all share the same biology and the study design here (100 “CNS tumors” vs. “normal” individuals) leaves a lot to be desired. It’s not clear to me if they also are basically using the same testing and training dataset or not.

Even the headline here (100% accuracy) is inaccurate; it had 100% sensitivity). The list of cancer bio markers which have turned out not to work for screening is quite extensive.

That said, the most interesting thing here is the microRNA capture technology. If it can be scaled, it might be interesting.