[yaakov34]

That's not really true for pancreatic cancer. It is absolutely notorious for escaping even when it's detected early, and even after a (seemingly) complete resection. This was in fact Randy Pausch's case: he had a resection with negative margin (meaning no cancer detectable on the surfaces from which the tumor was cut) and negative lymph nodes. For most cancers, his chances of survival would have been excellent, but not for pancreatic cancer. I remember at the time he wrote that it was a 50:50 proposition, but it's actually a lot worse. We don't know why this happens - it probably has to do with early metastasis by very small clumps of cells, or even individual cells.

This is one reason nobody recommends widespread screening for pancreatic cancer - apart from the inevitable false positives and so on. We currently don't have anything very wonderful to offer even to people in whom it was detected early, although it does improve the chances somewhat.