This is the easiest way I've found to combine multiple camera projections into one shader.
1. Find the error that needs touchup.
2. Create a new camera to project the correction. Make sure to lock the camera's translate attributes when you have it where you like. Also ensure that the film aspect ratio is correct. (I rendered a 1k image so the aspect ratio is 1)
Render from the new camera and save it as a targa/tiff.
3. Paint corrections. Save out your rgb and alpha for projection in maya.
4. Set up a new lambert shader, projecting these images with the new camera, just like you have done before. (make sure ambient color is all the way to the right and diffuse is set to '0')
5. Create a Layered Shader node. With the layered shader selected open the attribute editor. Middle-mouse-drag the touch up lambert in first, then the initial shader. Delete the extra color by pressing the 'x' under it.
6. Apply this layered shader to your object.
If you had to apply another touch up to the same object, you'd repeat this process and plug it into the layered shader to the left of the previous projections.
Hope this helps the process seem less convoluted!