In familial cases, the sensitivity is the same 78%, the specificity is a bit better 94%, so the math is still quite unforgiving.
If we consider that the prior probability of finding pancreatic cancer at a single point in time around age 50 is 10/100,000, and for someone with a family risk, it’s about 5 times higher, i.e., 50/100,000 = 0.0005 = 0.05%.
By using that test and getting a negative result, the probability is still 0.0012%.
Relatively, the risk decreased a lot, but absolutely, it was low and remains low.
The entire lifetime risk is, of course, higher than at a single point in time. So the test should be repeated annually. In doing so, one very likely exposes oneself to a false positive result