Yes, we solved it with some losses.
We made a database export with cleanup and repair commands.
I cant find the exact commands we used, but you should find them on multiple posts about database repairing.
I recommend to do a backup and test these actions on a cloned instance before running in prod.
We decided that the few artifacts we lose can easily be reuploaded by us. Due to the corruption we could not find any other way to fix this and keep everything. So it was better to make a cut now and prevent even more data going corrupt.
If anyone find a way to fix this easily, I’m still listening for future problems 