PDF is a read-only format, for the most part. While some operations can be done on it after it is created -like watermarking, adding or removing pages, signing, etc.- accessing its exact contents and layout is not possible (unless you use something like Acrobat Pro).
If you want to go to a lot of trouble you could probably adapt the
PDFRenderer library for reading the PDF, and Apache POI for creating a Word document. I'd estimate that to take a week's worth of work to get even halfway decent results, though.