[okket]

(a bit tangential, not directly related to the special case of pancreatic cancer, maybe in this case early detection provides better results)

Early detection is double-edged sword: It is very hard to tell in early stages if the cancer develops into harmful variants and a small false positive error can have drastic consequences, if applied on scale [1]. The result is that many undergo unnecessary therapy with 100% harmful consequences (not life threatening, but permanent damage like removed organs/sterility).

See also: "The Case Against Early Cancer Detection"

http://fivethirtyeight.com/features/the-case-against-early-c...

I'd put my money (research) rather on better treatment than earlier detection.

[1] https://www.youtube.com/watch?v=M8xlOm2wPAA

[melling]

It sounds like we need better non-invasive early detection. Tests where you can monitor the progress. Simply saying "we've got some early tests that aren't very helpful" doesn't mean we shouldn't be improving the tests.

Also, i'm not sure why we can't have both: better early detection and better treatments.

[shanacarp]

Pathways genomics is actually in clinical trials for blood tests to detect very very early masses that won't show up on scans.

One of the vps from Google's moonshot division,Jeff Huber, just left to head up a company that is a spinout of illumina, called Grail

No affiliation, but they have a bunch of interesting software jobs open after looking at thier site

http://newton.newtonsoftware.com/career/CareerHome.action?cl...