I'm not sure but I think that with the current system it's not possible to change that.
AFAIK, if the system was changed to take FM audio to software (like it does while recording) and only then sent to headphones/line out, it would be possible to draw VU meters as well as use any other thing that analyses the audio. But, this implementation would take more resources (CPU at least), and the audio quality would be worse (at least, this is what I've read on another thread).
Apart from this problem -
and this is just brainstorming - having FM chip send output to software instead of directly to the headphones
could (eventually) open the doors to other fancy features, such as pausing the live emission (the emission gets saved on music buffer until it is full), rewind the emission (again, the emission would have to be constantly saved on the buffer) and fast-forward it (this would imply that the emission is not live, and already stored on the buffer - something that would happen when you pause the emission). To resume, perhaps (just a possibility, need developer feedback on this) it would be possible to implement timeshift functions (like what happens on many DTV tuner and recorder boxes, but this would be for FM radio).
[end of brainstorming]