I've seen plenty of work on accelerator-drive heavy isotope reactors but nothing for light isotope reactors like lithium. Accelerator driven heavy isotope reactors still deal with many of the problems of conventional fission reactors - they're greatly improved in many regards, but still problematic (you still have some plutonium, you still have some fuel availability/cost limitations, you still have some long-lived waste, you still have some harder to shield radiation, you still have a wide range of daughter products making corrosion control more challenging, etc - just not to the degree of a regular fission reactor). A light isotope reactor using lithium would virtually eliminate all of these problems. And it has a higher burnup ratio, which is of course critical for space uses.
And while everything I've seen about past improvements in accelerator efficiencies and spallation process improvements, and what's being worked on now, suggests no limit any time soon on neutron production efficiencies - at least that's how it looks from the papers I've read. Plus, even if efficiencies couldn't be improved any further (there's not that much further one needs to go), one could hybridize a heavy isotope and light isotope reactor, using a heavy isotope target as a neutron multiplier to bombard the lithium. You'd require significantly reduced quantities of heavy isotopes relative to a pure heavy isotope reactor, and most of the energy from the lithium side could be as mentioned captured without Carnot losses, which is a big bonus. Any non-thermalized neutrons of sufficient energy would produce tritium as a byproduct, which of course would be a value-added product - in fact, given that the tritium-breeding reaction with 7Li and a high energy neutron yields a lower-energy neutron, the thermalization could potentially be done via tritium breeding in the first place. And tritium is a valuable product whether one has interest in D-T fusion or not.
I just think it's weird that I've not come across any work on a lithium-based accelerator-driven spallation reactor, and was just wondering if there's a reason for that. It certainly looks appealing to my non-expert eyes. I mean, it looks even cleaner and more fuel-available than D-T fusion, and looks closer to being viable on a full-system perspective. Versus accelerator-driven heavy isotope fission you get less power per neutron (about a quarter as much), of course, even accounting for Carnot losses in the former - but that's not what matters. Cost is what matters, and if you're eliminating the use of super-expensive fuel, not producing any costly-to-manage waste, have no incident radiation, no proliferation concerns, etc, you're completely changing the cost picture - without even considering the possible joint production of saleable tritium.