Find replace to fix mixed Greek and Latin characters

[greek]

The idea is to fix words like these where the first character appears to be Latin, but in fact it is Greek (one can tell that when changing it to lowercase):

Αstart
Τgjk
Μdlkj
Νsdjfskl
Εsdjfskl
Ζdjfskl
Ηjfskl
Ιjfskl
Κjfskl
Οjfskl
Υjfskl
Χjfskl
Ρagioi

So my idea was to find any of those Greek characters when followed by an English one and then replace the Greek character with correct English character (i.e. the Greek "Τ" with the Latin "T"). Is this possible with a single replacement operation?

So far the find is the one below but I am not sure about the replace:
([ΑΒΕΖΗΙΚΜΝΟΡΤΥΧ])([A-z])

The English characters would be:
[ABEZHIKMNOPTYX]

The Greek as Hexadecimal code points

Please Login or Register  to view this content.

[macropod]

That would be feasible to do with a macro, but the code to do so would run much slower than simply iterating through all the upper-case Greek characters. The former would require a loop to be built into the replace process to analyse each upper-case Greek character as it's found, then overtype it with the corresponding Latin character. By the time you've processed a few dozen such strings, replacing characters one at a time and with perhaps with thousands more to go, a looping process iterating through the same 22 characters but using ReplaceAll would have finished.