What about fusion thruster whose containment fields are powered by a fission?

[hkultala]

The problem of fission NTRs are low temperatures of fission reactors meaning that isp is not much better than chemical rockets, and all the isp advantage over chemical rockets comes practically from lighter propellant, which then means really bad impulse density

And the really bad impulse density means enormous and heavy fuel tanks.
With fusion, MUCH higher temperatures could be used, leading to both very good isp and good impulse density,.

But the common problem with fusion is that the magnetic fields containing the fusion consume more energy than the reaction produces.

But what if we don't even try to generate any electric energy from our fusion reaction. If we only use it as a very-high temperature NTR thruster, AND have another power source for our magnetic fields. And use a fission reactor for that.

The relevant comparison target to evaluate whether this makes any sense is probably an ion engine powered by the fission reactor?

[Stan-1967]

With the reference point being and ion engine powered by a fission reactor, you could examine the criticisms of the VASIMR.  It looks great on paper, but requires a magical power source to reach its full potential.  The problem seems to come down fission reactors must ultimately boil water for generating large amounts of electricity, so basically you have a heat engine, and it must reject huge amounts of heat.  The radiators kill the mass budget. 

Schemes skipping the steam generation, i.e. using fission/fusion fragment to generate the electricity, are the main source of hope for those schemes.

[ANTIcarrot]

Depends if you want pulsed fusion, or sustained fusion.

For sustained...
Take a barnsworth fusor, put in under a big electromagnet & shield, feed some aneutronic fuel into it, and power the whole thing off a fission plant. Congratulations. You have produced net thrust! You have a desktop fusion engine - with all the maneuvering capabilities of the desk it's mounted on. And it's rather tricky to move beyond that due to conventional materials.

For Pulsed...
You have any number of other options, mostly based round inertial confinement or laser-confinement. Still experimental, but we at least have back of the envelope blueprints about what such engines look like.