Temperature.
Conservative thought about steam heat transfer processes at the most limiting point.
Length of piping (number of pipes and pipe penetrations going in and out of the primary systems and through the primary pressure boundaries.
Coal and gas-fired boilers blow extremely hot air and combustion gases past the saturated steam to gain their superheat, right?
You take the saturated steam from the boiler piping at the end of the conventional runs, run it back to the front of the boiler in new rows of pipe being hit directly by the hottest of the boiler gases as they leave the burning area.
So, in a nuke, where are you getting the the "extra" heat? The core, right?
So, you need to run hundreds of pipe (or regroup all of your existing pipes into a single one again) back into the the hottest part of the core (where you are trying to control the nuclear reaction by modulating the water fraction, fuel fractions as they change, and control rod fractions, poison fractions as the fuel burns up, and the "contamination" of the various components as fuel becomes replaced by nuclear fission products. Now into that, you are introducing a varying steam+water mix of pipes and pipe materials - none of which are fuel - and all of which vary in density and nuclear cross-section with temperature.
It makes your nuclear control more difficult, your metalurgy more difficult, your primary containment system more difficult. A few plants - I don't recall which, DID try to even use conventional burners + superheaters to reheat the saturated water post-steam generator, but that didn't last long. Too much cost for too little benefit.
All of these problems could be solved, have been looked at before in various ways, but the simple, solid design of "leave it as is" weighs out. You have to be able to sell your design across the public desks of local and state politicians, to the local public.