Hello from across the pond.
I have had the same problem with my 1918 D45. I soaked the cage with a 50/50 mix of acetone and ATM, tried hitting the valve with a 5lb hammer and wooden pole from below, tried jacking the car up on the pole pushed up through the bore so the entire weight of the car was taken by the valve and cage and at the same time applied heat (having removed the ATM/acetone mix!) and attached a slide hammer to the valve stem - all at the same time and no luck. All this was taking place over a period of months and I was getting desperate. I finally called in a legendary fettler here in the UK (name supplied in personal communications) who had listened to my tale of woe over the months and eventually made a 200 mile round trip to visit me. He took his 5lb lump hammer and a steel drift and told me to look away... He then made several well aimed blows with the drift around the lip of the cage - driving it down the bore, subjecting the rust on the cage to a percussive sheering force . (Boy was I worried about my beautiful cast iron block...). He then reattached my slide hammer and within a few upward slides the cage started moving and came out within minutes. Genius! He fixed a problem in five minutes which had me baffled for 3 months or more...
His theory was that the counter intuitive blows to the rim of the cage transmitted the shock such that the rust which had effectively welded the cage to the block cracked thus releasing the cage.
I needed to remove the cage because it had turned and had reduced the inlet orifice by about 50%. One of the other cages had spun around and completely blocked off the exhaust outlet to another cylinder, but happily that one moved easily. I found that the inlet cages had rusted, whereas the hotter exhaust cages were free.
Terry Wiegand knows this car because he helped the owner free the engine before he sold it to me. It is now running sweetly with new white metal bearings having had a knock in the engine.