Many games do something like this, but realities of networked games undermines how well it works as a anti-cheat feature.
The games I’m most familiar with are the Quake series, but the techniques were not unusual. The QuakeWorld and later servers used the same visibility data in the maps that was used to avoid drawing hidden parts of the map, to limit data send to the client, but the checks were deliberately more relaxed. For rendering it would try to only draw a surface that had at least some part visible, but for network updates you were sent anything in that visible area, and the visible areas directly adjacent to it. The intent of this extra data is that a player that has moved around a corner should not need to wait for their client to send that movement to the server, and for the server to respond with what they can now newly see, before they can show who is around the corner.
This was never really done for anti-cheat reasons, but to reduce network traffic, if your sending every client updates for every player, bullet, rocket, etc. then there is a chance that an update for something critical to that player will get delayed behind updates for things on the opposite end of the map.
A modern game with a big open map and not a lot of hallways and tunnels this becomes more about distance then specific position, further reducing the value as an anti-cheat tool.
