Forget biopsies, ultrasounds, mammograms, pap smears, rectal exams, and other unpleasant cancer screenings—the race is now on for simple, affordable blood tests that can detect all sorts of cancers extremely early.
Source: Big names gamble big bucks on blood tests for early cancer detection
This kind of screening is of a very different type to PSA or mammography, where many things can cause elevated PSA or high radio opacity in breast tissue. You are looking either for circulating tumor cells, or as in the Grail case, nucleic acid fragments from said cells. You sequence the DNA you have, looking for known oncogenic mutations. If you find them, the patient has cells with active oncogenes in them. As some oncogenes are known to be particularly deadly, mutant KRAS for instance, they would presumably be prioritized. I would worry more about false negatives than false positives with this technology, as it would presumably not be able to pick up lesser known mutations, and might be completely blind to tumors caused for example by gene amplification.
Why screen early? There are some very rapid cancers, and there are some which appear to be very rapid because they are diagnosed very late. Many of these cancer, pancreatic (and KRAS) being a poster boy ,have been around and undetected for many years, yet the average patient only survives about 6 months after diagnosis. For these cancers, this kind of screening could be a real lifesaver. The average ovarian tumor is 400 grams when detected, and survival is then usually a brutal land war.
We might start screening and then find out tumors can be detected in everyone. This would give us useful data on how frequently cancers arise, and how frequently they are overcome by the immune system. In such a case we would learn not to treat everyone, but maybe wait until certain thresholds were exceeded. Or, since we know what the mutations are the tumor carries, we could use highly targeted immunotherapies, with low toxicity. Everything is impossible to implement perfectly on day 1, but take a look at the landscape, try some things, and learn, and a decade later you have made real progress.