sevam is clearly from an Iranian loan in NW Indo-Aryan tongues. While Early Indo-Aryan probably had a name for apple it was lost.
There is a NW IA language Pashayi which has a word wAlI for apple that cannot be etymologized in any Iranian pathway. Some have propose that it is a descendant of the now lost Indo-Aryan abalikA, the original word for apple. Of course, some I-Ir linguists could get hot as a coil hearing this proposal. Iranic probably had a comparable abarnA based on some East Iranic forms and a loan into Uralic. So by reconstruction from the possible form in Corded Wear IE it would be better to form abalikA rather than sevam.
It is interesting that the varuNavR^ikSha (Crateva religiosa) is called abala in Skt. Wonder if that might be a survival of this etymon