Since you mentioned NDL dives, I ran a 100' dive with 32% through the SAUL probabilistic decompression algorithm and it came up with a 0.169% chance of DCS for a 30 minute dive; 0.047% for a 25 minute dive. So your risk is less than a third if you shorten your dive by 5 minutes. The algorithm does include a 3 minute safety stop. A 30 minute dive is the PADI RDP NDL, and those tables are aggressive compared to what a dive computer may come up with.
These numbers, with your undeserved DCS hit, seem to hint towards increasing the conservatism (shorter dive or longer safety stop, or both) during the dive to avoid the hit altogether, rather than trying to mend it out of water.
If I had the O2 available, it would probably have a bigger impact used underwater rather than at the surface (if feasible of course). You'd get the benefit of the higher ambient pressure keeping the bubble sizes under control, while simultaneously keeping the tissue gradient high (0% N2 in your breathing mix) to offgas quickly. Even just a few extra minutes on O2 or Nitrox underwater might have a bigger impact than at the surface for reducing your risk factors.
Another thing to think about is having to deal with the O2 logistics - if your surface O2 session is say 30 minutes (based upon the prior post) you would probably get 2 sessions out of an AL40 cylinder of O2 depending on fill and type of breathing apparatus.
In a pinch, I might just breathe down the remainder of whatever Nitrox I had left in my cylinder at the surface as that would be better than air.