Cloning is probably the easiest way. If you clone then you need only, at a minimum, update the pixel data with the subtraction image (watch out for negative pixel values!) However, you may want to also change other fields to perhaps place the subtracted image in a new series of the same study or to change the sequence number and leave it in the same series as the originals. You may also want to change the image type to indicate that this is a derived image, not a primary data image.