Longer answer: Incoming email is - IMHO - a persistent misnomer if not a misconception as usually a workstation doesn't run a mail server but only a client which fetches (downloads) the mail from "somewhere". But why, you ask, not scan the downloads, aren't browser downloads scanned? Yes - and you could do the same with email.
The difference is that email content isn't immediately executed - well, it depends on your settings, "previews" can be dangerous. Anyway the mail can't be "run" immediately. The client first has to decode the content in order to display or execute it (whether HTML, PDF or something else). An on-access scan will catch the threat at this point. If you decide to delete the mail before viewing it or before opening an attachment or a spam filter deletes it before you get to see it there is no need to scan it. Contrariwise you can't skip the on-access scan every time the mail or an attachment is opened as in the meantime new rules might have been released and just because the email seemed to be clean at download time doesn't guarantee it is.
Apart from that the scanner would have to download and decode the entire (at which point is it safe to pass it on to the client?) message and then the client repeats this. And this won't help if the mail contains HTML and the client subsequently happily downloads remote content using the http protocol.
Be assured, the contents are scanned "in time" even without this feature.
