What type of cancers are you talking about? What subtype? What population are we talking about? What subpopulation? What is the difference between an early diagnostic and prevention? Are CTs and MRIs the right tool for the job? Who is most at risk? How often do you screen your target population? How do you make sure they actually come and do the exams? How do you measure success? Are you ready to deal with the outcomes of the diagnosis?
These are just basic questions really. Some people devote their entire life to these hard topics. The radio host sounded like an "expert" with "great ideas" to an equally clueless audience when really he just has no idea.
At some point, the scan has a higher likelihood of finding a false positive than a true positive. Interventions on false positives can themselves cause harm, both psychological and physical.
If we ignore the fact that they obviously are not preventative (they can't find a cancer that doesn't exist yet...), and as such it's rather silly to present it as an alternative to treatment, there are still a multitude of issues:
- Early detection, for the sake of argument assuming that the proposed screening program would in fact be effective or could be adjusted to be effective, does not always improve survival rates. There is a diminishing rate of return from pushing the detection earlier and earlier.
- Any screening needs to deal with false positives. Telling someone they may have cancer leads to suffering, and may in some cases trigger more serious effects. E.g. consider someone suicidal who is told they may have cancer, and decides to end it. Or "just" decides to engage in more high risk behaviours, increasing risk of harm. And for something that is rare, even a very small rate of false positives during large scale screening might cause disproportionate suffering, and might end up causing significant harm.
- False positives require further investigation, and if not ruled out with reasonably low impact methods like biopsies may result in over-treatment. This is particularly serious for cancer, where many of the treatment alternatives have a substantial risk of causing death.
- The screening also needs to deal with false negatives. Someone who is told they don't display any symptoms of cancer may pay less attention to signs later on, and may end up discovering their cancer later than they otherwise would, increasing risk of death.
All of these affect overall outcomes, and you need to tally up the effects of all of them before you can say if a screening program will be overall positive or overall negative.
One outcome of actually analysing this, for example, has been a push towards ending large scale mammography screening, because newer surveys that takes the above effect into account shows that while it does in fact reduce mortality somewhat for those who actually have breast cancers that would kill, it causes sufficient overdiagnosis to kill many otherwise healthy patients. Here is one article that covers this problem [1].
The problem is that people often assume the worst case outcome of a screening program is that it won't be effective, but the worst case outcome is that you kill more people than you save.
