Isn't this still too high of a false positive rate to be useful? IIUC, if you give this to 100,000 people, it will give you 2,000 false positive results. But only 13.5 new pancreatic cancer cases occur per 100,000 adults in the U.S. each year [0]. So if you took this test annually, a positive result still means a less than 1% chance you have cancer?
edit: although that NCI data doesn't make sense to me. If only 13.5 new cases appear per 100K, how can that accumulate to lifetime diagnosis rate of 1.7%? 100000 * 0.017 = 1700; 1700 / 13.5 = ~126 years to accumulate a 1.7% chance of diagnosis.
I wonder what the follow up test is like / cost. If it is highly accurate, but you end up doing expensive follow up tests for 2,000 people ... that has to factor into the overall "cost" of such screening right?
You wouldn't just screen the entire population with this test. You'd give it to the high risk factor subpopulations where the rate is way higher than 13.5 / 100K: older people, tobacco users, etc.
[0]: https://seer.cancer.gov/statfacts/html/pancreas.html
edit: although that NCI data doesn't make sense to me. If only 13.5 new cases appear per 100K, how can that accumulate to lifetime diagnosis rate of 1.7%? 100000 * 0.017 = 1700; 1700 / 13.5 = ~126 years to accumulate a 1.7% chance of diagnosis.