Hello Pradeep,
we should see the context of the sentence. It is probably inserted when explaining the OSPF neighbor state machine on a point-to-point link.
However, routing from DRother to DRother in the same LAN segment is possible, DR and BDR are central points for routing exchange ( they allow to reduce the number of adjacencies needed in the LAN segment, so saving on resources like cpu and memory on the routers), but user traffic is not constrained to go via DR or BDR, or DR and BDR would become a bottleneck.
Routing from DRother to DRother is possible because both nodes have complete topological information, even if they are not fully adjacent, they are both adjacent = synchronized to DR and BDR. This means they have the same database.
Actually within a single OSPF area each router has a complete and detailed knowledge of the topology, for example what devices are in a remote LAN segment, who is the DR in that remote LAN segment, what are the OSPF neighbors of a given router and so on. The collection of Router LSAs and Network LSAs provides this capability.
As a result of this, sending user traffic to another DRother router is possible when needed as the local node has received the Router LSA of the other node and can understand that there is a valid next-hop in the LAN segment to reach the destination ( in the case the destination is a network attached to the other DRother node, that is the simplest case).
Nowdays there is a trend to use point-to-point links as much as possible, but in the past there have been LAN segments with tens of OSPF routers in a single LAN segment. They would have been a true bottleneck if all user traffic had to go via DR.
Hope to help
Giuseppe
