Oncologists who treat prostate cancer have long been frustrated by a quagmire: They know they treat many men whose disease won't harm them, but at the same time they fail to catch aggressive cases that kill. There's no good way to separate the two right now, but in tomorrow's issue of Nature, a team at the University of Michigan, Ann Arbor, suggests a new strategy: Look for specific chemicals in urine that could distinguish among patients with no sign of disease, disease that isn't critical to treat, and disease that's most dangerous. Researchers have poured tremendous energy into early cancer...