The problem with 1.1 is, that it has to use MemoryImageSource, which doesn't provide a graphics context. If the text/image is static, use textures instead and blit those. If it isn't, you can draw the rendered image into another one, draw your text/images onto that and draw the whole bunch to screen. That will be slower of course, which is why i would limit it to 1.1. and use another pipeline for higher versions.
You can also access the pixels-array directly, but i doubt that this will be very helpful in this case.