Pretty small, but maybe that's doable. Why does it have to be 10 meters from the scope? Why not 100 meters? It would be difficult, but it still seems easier than building a fleet of telescopes linked together.
Actually to counteract the shadow of the occluding disk it might need to be some kilometers away. The issue there is keeping a sub-arcsecond alignment between these two objects. Although the disk can get larger as it grows more distant, it then becomes more subject to atmospheric drag and solar wind--making it even more difficult to keep in alignment. And of course the telescope size needed to resolve this milli-arcsecond planet would be enormous, requiring an ever larger disk at even greater distance. It seems that when you try to control one problem area, another blows up in your face. The more I think about it, the more I think the interferometer wins. Even that faces massive difficulties though.