I suppose it depends on its stem? If its stem was Atamiss- then one would have imagined that it would go Atamiss, Atamissis. I doubt that ss word-finally would be written in Latin, because it can't really be pronounced differently from s; unless the same phenomenon occurred as with hic(c), hoc(c) in which the second c was pronounced when a vowel followed. Word internally ss was simplified to s (hence causa < caussa) except in cases where the ss appeared after this sound change (at least, I assume that's the explanation for eg. admissus).
So it depends on when the word entered the language, and whether (and when) Atamiss > Atamis (which could then be reanalysed as an i-stem). Depending on these factors, you could end up with Atamiss, Atamissis; Atamis, Atamissis; Atamiss, Atamisis; or Atamis, Atamis.
Altrúnia's takeover happened over 100 years ago in-timeline (at least, c. 2420). Thus, I'm not sure the Skarans would have changed anything because Pandoran's official stance on place names is transcription as is unless there's a change needed for pronunciation (Altrunia > Αλτρύνια (Altrúnia)). The language started as transcription-only because there were two parent languages, Old Latin and Kioné Greek.