With a DC controlled layout, the track is electrically broken into sections. If a train derails on a point frog, then the controller limits the maximum current at around 1A or so. If the back edge of a wheel contacts the "wrong" rail while going through a point, the the controller may remove power momentarily, or may not even notice. In either event, power loss is to one zone only and the remainder of the layout will continue uninterrupted.
With DCC, the current limit is usually set at around 5A: sufficient to power several trains. If DCC didn't allow this, it wouldn't achieve much! However, that does mean that any fault has a higher current available to flow through it: this in turn provides a higher capacity to cause heating & arcing. Consequently DCC systems usually have a much quicker shutdown. Often, a DCC booster will then wait around 0.1s before trying to reapply power; this can cause noticeable "hesitation" if a wheel causes a temporary short.
The most effective way to prevent this being a problem is to stop it happening in the first place. This is achieved by changing the electrical connections within the pointwork to eliminate "back to back" shorts etc.
The definitive guide to this subject is in Allan Gartner's web page. That sets out the reasons, and what can be done. My track is Peco code 55. This is covered in that site and doesn't require a lot of work.