I have always done it by one gram contributing an estimated 1 mL. It may not be exactly right, but any error that might exist is too small to be of importance.
However, where someone has been using a formula that has worked well for him before, then there is no reason to go and change from that formula.
For extremely high concentration formulas such as this one, it might be the case that 350 mg/mL wouldn't have dissolved anyway. I don't know, as I don't try for extreme concentrations.
You might just see how you like this and if you do, then next time do it the same way, and consider yourself an individualist for using 330 mg/mL instead of the units-of-50 that everyone else pretty much insists on using.
I mean really, if 200 mg/mL is okay and so is 250, then why wouldn't for example 220 or 230 be nice concentrations as well? But people don't see it this way. That though is only an aside
Anyway, point being, if the 330 works well for you, why not stick with it rather than try "improving" the formula by probably-more-accurately accounting for volume of steroid, but as a result perhaps getting a concentration that doesn't hold.
Also, from the standpoint of round numbers, 330 gives you almost a round 1000 mg per 3 mL, which might be convenient, unless worrying overmuch about the "missing" 10 mg.