The short answer is what you are asking cannot be done in the direct manner in which you are assuming.
This will only work in certain conditions. The easiest to describe would be if the SMTP server you are using to send the message, is the same server that hosts the domain of the recipient email messages (the server you refer to when setting your .HOST property of smtpClient). So, if you were only sending to recipients on your local SMTP mail server, then this might work pretty reliably. But that depends on the specific SMTP server software being used and potentially also on how it is configured.
To explain why this is, you must realize that only the last SMTP mail server receiving the message that actually hosts the desired email addresses, will be able to authoritatively answer the question, is this a valid email address. If the message has to pass through any other email servers on the way to getting at this final authoritative server, the message has to be handed off sequentially from one server to the next server in the chain until it reaches that final authoritative server. This means that there is not a guaranteed method for authenticating a specific address. Couple this with the fact that some domains are configured to act as a black hole and swallow illegitimately addressed mail, and you can see that there are many reasons why you cannot rely on that methodology.
So, many messages to external domains are going to have to hit at least one separate SMTP server and depending on how that server answers or forwards the mail, it will determine the results for any specific receiving domain. In fact, monitoring the FROM address for bounced messages is not foolproof either as my previous comment about some hosts putting some messages into a black hole if they do not appear to be valid.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…