There is also the related idea to 11 .. Nce7 of 10 .. Nce7. This suffices to defend vs Ng5 and can turn independent if black doesn't put his queen on c7. It's certainly tempting to keep the option of Qe8/Qa5 in one move if possible.
(Although there then isn't a bishop on e3 to hit with Nf5!).
Eg:
11 h4 Qa5 12 Bd2 c4 13 Bxg6 fg 14 Qg4 Qa4 ^ 15 h5 Qxc2/ Qe8 if white defends c2. (or maybe Qxa3 - might be too risky).
The sacrifice known from 10 .. Qc7 11 h4 c4 doesn't seem to work here with black defending g6.
Kindermann and Dirr give another line here for black which they also think is fine.
11 Bg5 is both logical and annoying for black since he has to defend the e7 knight with his queen & unpin. So Qc7 or Qe8 I suppose.
11 dxc5 is the main line if anything is. Black then seems to depend on 11 .. b6 working.
12 Be3 Qc7 is back to the 11 .. Nce7 lines (and meant to be OK) so white's got 12 c6 or 12 cxb6.
c6 was what happened in the only GM game I know of which reached this (Lutz - Pelletier, Zurich 2002) which turned very random indeed after 12 .. Qc7 13 o-o f5 14 ef gf 15 c4 etc.
Psahkis suggests 13 .. Qxc6 14 c4 Bb7 which seems a little calmer
After 12 cb Psahkis just says Black has an easy game.
I assume that black goes Ba6 swaps on d3, provokes d4 by hitting e5 and pummels away at a3 & c3.
This certainly looks to be enough to make it hard for white to do anything with his extra pawn. One question to ask is whether black has any realistic winning chances here. He's got to keep some bits defending the kingside. White might also be able to avoid playing d4 when life is harder.
Still this seems interesting to play and I think 11 Bg5 might be most worrying. Unless I've missed something this line does look interesting though.