There are a lot of ways to slice those numbers; I like deaths / km as it describes the risks of traveling from point A to point B by various methods. On the other hand, it's likely skewed somehow by the fact that you rarely have more than one fatal accident per trip.
Another way of looking at it is deaths / trip; a quick search shows 4,800,000,000 bicycle trips in 2019 for a rate of 48 deaths / billion trips and 229 deaths versus 3,600,000,000 car trips and 610 deaths or 169 deaths per billion trips.
The average trip length in the US is apparently something like 10-15m which might make the number of automotive deaths per billion trips something like 130-200. Maybe. (I can't find an actual estimate of the number of trips.)
Another way is deaths per population, see one of my other comments.