The principle is: if an IR test on a cable reads high, the cable might be good. If it reads low, the cable is definitely bad.
We are surrounded by insulating air, so in a dry environment there is no easy way to detect places where PVC insulation has been punctured or stripped as the air surrounding the cable will provide equally good insulation. In truth, in a dry environment where there is not much earthed metal around, IR tests will miss many kinds of damage. The problem is much more likely to be detected where something non-insulating is present to make contact with the exposed conductors; damp masonry, soil, dead wildlife etc. E.g. if a squirrel gnaws the PVC off a cable cleated along a wooden joist the loss of insulation itself won't cause a low reading. But if the carbonised dead squirrel ends up lying across the exposed conductors then the IR will show that up.
Does it matter? If a screw just grazes the line conductor in a cable then you have a live screw, but it won't shock anyone unless they touch it. There must be hundreds of thousands of live screws and nails around the world, bashed into dry timber and completely undetectable by normal testing. If the screw happens to be holding up a metal shower rail then it is a serious shock hazard, but it is likely to be in contact with various building materials, not all perfectly dry, so at that point the insulation resistance is likely to read low enough to earth to arouse suspicion. This is why it is so important for IR tests to be done with the CPC connected to the MET, unless specifically troubleshooting.
IR tests are very good at finding substandard insulation materials, water ingress and underground faults, because here the good insulators are replaced with bad ones, rather than with air.