I think that usually that would not work, but depending on the ester you probably could extract successfully with DMSO, and then recrystallize by adding water to the DMSO.
I did this successfully for the purpose of NMR analysis a long while back (more than 10 years) but don't recall what the compound was.
The compound would have to have a melting point well above room temperature for this to work. So for example, testosterone propionate could be a suitable choice, but nandrolone decanoate would not be.
For whatever reason, it can be surprisingly (to me) hard to get recrystallization in some instances from adding water. One could wind up with a cloudy mixture that stays that way. Patience will be a virtue.
In cases where the oil would remain clear in the refrigerator or freezer, but hard solids precipitate out, then yes those could be isolated as you describe.
In some instances you might aid the process by adding hexane, perhaps at equal volume, before putting into the freezer. This would be reasonably likely where a product is particularly concentrated, meaning that there is a substantial percentage of solubility enhancers. The solubility of the steroid could become much less on addition of hexane, even enough so to more than compensate for the added volume. However that would be something to try not with an entire batch but just as its own side experiment.