We actually had something similar in version 2.5 of Blesta that would run emails through spamassassin on the local server, if it existed. We could certainly check a header in an incoming email and reject based on the header. What spam filtering software are you using? I imagine this header changes from filter to filter and it might be better to be able to specify the header and the value threshold.

regardless if your hosting provider is using spamassassin or rspamd or ....

You should be seeing

X-Spam-Flag: YES

and then add a parser for the score that presents like

X-Spam-Score: 6.274

Now while the scoring may be different from mail solution to mail solution, setting a threshold for above (or smaller than) $value seems easy to implement.

So the my proposal would be to add the type "Spam" to the Status of tickets (open, closed, client-reply, trash,...) and just move the ones that fall into the junk criteria (x-spam= yes AND/OR x-spam-score) into that. This would also help blesta users in tuning and ensures no potential tickets are actually lost, even if they are sorted as spam initially (with the option to re-tune the scoring value required for their individual setup/mail solution after investigating why items got sorted into spam)