So treatment decisions are clearer if you're talking about high-risk or low-risk patients, but harder in-between?

Everybody agrees when all the characteristics of a case are low recurrence risk, and by the same token, if many of the characteristics of the tumor and the patient's characteristics are poor, then everybody will agree this is a pretty high-risk case. But the majority of cases are a mixture of good and bad. The challenge is taking this mixture, and figuring out how to interpret it, and where to put the emphasis on the good characteristics or the poor characteristics, and making the decision about how to treat the patient. Oncotype DX™ does this based on a weighting of characteristics, generates a score, and that score seems to hold up in all of the testing that's been done so far.