Within your published "final document", can viewers select and copy the text? Or is it "unselectable" and essentially an image.

If text is "selectable", then an OS/system-wide screen-reader should function.