I've shot NGC 7822 before as a two panel mosaic, but I was cutting a lot out to get it to fit.  It was clear that to get the whole thing, I would need at least four panels.  At that point, I figured why not go broke with another four panels and pick up SH2-170, the dot on the question mark.

As I had data from an earlier version, I experimented with using subs from that project to supplement the new subs.  As they didn't line up, it was somewhat a pain as PixInsight's WBPP has the habit of choosing an old sub as the reference frame for alignment, which completely messes up the mosaic.  I had to manually select a reference frame.  This is a pain as I typically drop all the frames for the mosaic into WBPP and group them using a PANEL keyword, but you can't manually select reference frames on a panel by panel basis, so I needed the process those panels independently.

I tried this early on when I didn't have a lot of data, but it was fairly obvious where the old data overlapped since it was so much cleaner.  When I got enough new data so it wasn't obvious, the benefit of the extra subs was pretty marginal.   In the end I conclude it wasn't worth the effort.  Besides, when you get to a large mosaic, those fine details will never be visible except those of us with the pixel peeping affliction.  As it is I should really downscale these big mosaics, but I really like zooming it.

Where I think makes more sense to fold in additional data is when you're covering a huge area, perhaps even using a camera lens, and want to enhance the smaller detailed object in the large sea of nebulosity.  I've seen that used in some peoples huge mosaics covering multiple constellations.  The bulk of the mosaic is pretty low res when you zoom in, but the cool little object are sharp.

I have a lot of appreciation for those people who work on large mosaics as part of collaboration of multiple people with data all over the map in terms of image scale, not to mention different color balances due to different cameras and filters.