well, not actually in the feedback loop, I don't think...
but you can include losses down the cable, by moving the tap-off point for the amplifier's negative feedback loop to the far end of the speaker cable... that only covers the "hot leg" of the cable, so for true coverage, you'd add a differential amp, sensing the signal arriving at the far end of the cable (complete with resistive/inductive/capacitant losses etc. over both conductors) and correct for it in the feedback circuit. Any overshoot and "motor effect" back EMF would also be compensated for in a handy manner. In this sense it's just like a set of remote sense terminals on a power supply, only the amp is like an AC power supply.
This is actually rather similar to the 'sense' legs on DRV134 and SSM2142 line drivers.
Drawback: the cable will add phase/delay issues due to cumulative inductnce and capacitance. The effect over cable length gets really, really bad! Depending on your amplifiers equivalent "open-loop" gain/bandwidth product, the result might be a 100Watt radio transmitter! -For that reason most anufacturers probably shudder at the thought! -Also, lower bandwidth amps might not exhibit the problem so readily... the better the amp, the worse the nightmare! -Since you can't control what the user is going to do (like run 100 feet of zip-cord to fix a problem "in a pinch"...) you can't guarantee the amplifier will remain "FCC compliant" unless you make it spec like crap bandwidth wise.... and the effects in the audio band start to become really nasty...
-If you're putting amps inside speaker cabinets though, it might be worth considering. -Then you control the whole thing.