a 1% false positive rate. ouch.

If the blood test becomes widespread you could end up with more false positives than true positives. So probably not useful as a routine test with this error rate.

Worth noting that because of anonymity, they couldn't go back and check if those 7 people actually had cancer and just didn't know it.

Recently the woman who could smell Parkinson's Disease got one test wrong but when they rechecked they found that she was right, it was just undiagnosed in that person.

That specificity implies one of these cancers apriori occurs in 5% of the tested population.

Is that about right taking into account pre-screening? We can identify a population that has a 5% chance of having one of these cancers?

That's a lot of false positives. If 4% of people actually have one of these cancers at any point in time (a number that seems quite high) that means that 20% of positive results would be false in the general population.

How does that compare to the base rate of these cancers in the population? The false positive rate in a control group can be very low, but still overwhelming when applied to the general population.

Ah, missed that, thanks. I assumed it was in the paper itself.

Oops, thanks. I've always had struggle remembering these terms because they are so confusing. "Specific" means that something only relates to one particular element X in a set of more than one thing. Given that meaning, one would expect specificity to mean "how good is the test at correctly reporting the thing it is designed to report X rather than confusing a different thing (an element from the complement of X) for X", but that is sensitivity. Instead, specificity means that with respect to the complement: "how good is the test at correctly reporting all the other elements in the set ¬X as negative rather than confusing an element from ¬X for X".

"Sensitive" has a clearer intuition: For example when you want to detect a weak signal, then if your antenna is more sensitive, you can better pick up the signal, thus you are better at telling whether the signal is really there (you have a better true positive rate).

I cannot think of a better name for TNR, though, other than simply sticking with TNR and TPR.

It's not that simple. If the worst effect was some unnecessary secondary screening, it would be a no-brainer.

But first of all you need to account for the negative effects of a false positive, and they are much worse than that. They include anything from stress and psychological effects, cost and risks of all the additional screening, to the much worse issue that you also need to account for the fact that for some cancers in some instances it will be hard enough to determine if the growth is malignant in the sense that some proportion of cancerous growth will never pose a threat to the patient, but once diagnosed there tends to be a strong pressure to treat. As a result too frequent screening will save some, but will also result in a large amount of unnecessary chemo, radiation therapy and/or surgeries - all of which come with risks in addition to the pain and discomfort.

Then you need to account for frequency of testing. A 1% false positive rate per test looks distinctly worse if the test is repeated every year for a 10-20 year timespan, for example.

Then you'll also want to account for how much it improves outcomes. If the cancers it detects are ones that would generally be detected in time anyway, and/or have low mortality, and where treatment outcomes are good, the incremental improvement would not necessarily be big enough to justify the negatives.

For breast cancer, for example, there has in recent years been a push to reduce large scale routine mammogram screening because it's not clear if it does more good than harm for many groups of patients (in large part because it leads to overtreatment). This article goes into some of the issues related to that in some detail (and presents both proponents and detractors of large scale screening):

http://www.healthbeatblog.com/2009/04/mammography-screening-...

No it doesn't, since of the people known to have cancer who were tested, only 70% tested positive.

With a 1% false positive rate that won't work I guess. False positive rate needs to be much smaller.

True. That is really difficult to obtain though. But assuming that most people are eventually diagnosed with cancer (even if just close to death), then a similar maybe a little lower population must have undiagnosed cancers.

In the end you'll need a properly conducted study with random undiagnosed people to get reliable numbers though.

CAT scans are insufficient to determine if a growth is malignant in many cases. Which would lead you down the road to biopsies, which are also insufficient in many cases. Which would lead to unnecessary radiotherapy, chemo or operations in at least some cases.