Ah, but it doesn't matter to the FDA whether there are good data or not. Logically the potential exists. An example: look at the malignancy warnings in every single TNF-alpha blocker label. No good data exist there either.