Files can be infected again if adequate measures have not been taken to eradicate the threat.
When files get modified or "written to" by malware, a scan does not occur until _after_ the write has occurred. This is why you'll hear us tell you DO NOT rely on WRITE-only scanning configurations, using bold CAPS text, and it may even look rude - but it's a point we cannot stress enough for persons interested in avoiding infections, and reinfections. Make sure READ scanning is turned on. In fact, if you must turn one of them off you can turn off the WRITE scanning. NEVER turn off READ scanning.
In your case though, you still have the threat in your environment and it's still got access to this machine (at the time that detection occurred).
Getting detections from us doesn't necessarily equate to "this machine is infected"; it could mean, "somebody who has access to my files is infected" and they tried to get you but we cleaned up their mess.
Detections should be followed up on to make sure the node is safe/secure/clean. You can inspect if the infector came from a remote location by checking if any IP addresses have been placed on our blocked list .
If you found nothing, perhaps your User launched some code remotely that allowed the threat to become memory resident again - that happens a lot when servers fall victim to malware.
A lot is touched on in this post. It may warrant further break-down/discussion.