Im trying to improve the regex for email matching (Im not running the tracking token).

We have some case registration build (Custome entity). Each case get a unique increasing number. We send emails from that entity and use the unique number in the subject.

The problem is that when users answer the emails they are not always getting tracked. It can be many reasons, ex. addin small changes to the subject or adding additional contacts to the mail.

My idea was to improve the regex to leave anything out besides the unique case number -> maybe it should be "Case # Case number" (To make it more precise). The case number can be either 4 or 5 digits.

A typical subject could be: "Case # 4.2345 // Title from the case and maybe some other reference (33243)".
If that email is answered it can be "RE; Case # 4.2345 // Title from the case and maybe some other reference (33243)". 
Or maybe a user also add something to the subject "RE: Case # 4.2345 // Title from the case and maybe some other reference (33243) users have now added a new number 45 that will typially break the matching".

The regex will then have have to filter anything out besides "Case # 4.2345 //" But i ca not get the regex working. Is it even possible to create such regex (As far as i understand it will have to filter out anything before "Case" and anything after the "number". (maybe the first time "//" appear that than be used to filter out anything after / to the right"). And yes if the user messes with the case number then it doesn't work - but that is acceptable.