And of other things, like hemorrhoids, diverticulitis or an abscess. Colon cancer is typically so slow-growing that this is unlikely to save more lives than currently prescribed check-ups and self-observation. The increasing flood of data remains constant as it has since language was developed; it'll be interesting to see how we (or our software agents) adjust sensitivity this time to avoid mass panic, or more pseudo-science trends. See: increased frequency of mammograms leading to more false positives, more stress, more unnecessary invasive procedures.
Yeah when I say above checks against millions of records ... Those records would be over many years and possibly many many years of data that can point to data points showing this X group of people died of this disease and their specimen(urine too) all had this same trait.