After much thought on this topic, I have pretty much concluded that the cost in volume of the airfoil surfaces needed to generate meaningful aerodynamic lift would exceed the volume required to just bump up the maneuver drive to the point that it can take off and soft-land without aerodynamic lift. Want to land on a world with 1.5 G of surface gravity? Install a Maneuver-2 drive (or a Maneuver-1.6 drive if rules allow). Want to refuel at a gas giant with 2.53 G gravity (Jupiter) at an altitude where the atmosphere is dense enough for the intakes? Install a Maneuver-3 (or 2.6 in allowed) drive.
There would be exceptions. An obvious example is an aerospace plane (defined as a vehicle that flies in both atmosphere and space, using significant aerodynamic lift while in atmosphere) built on worlds that don't have gravitic drives. A few real-world examples are under development, and the Space Shuttle was such a vehicle when returning from orbit. Such vehicles would be most efficient if designed for a specific world; gravity, atmospheric density, and atmospheric composition (oxidizing, reducing, non-reactive).
The constraints of aerodynamic lift make it impractical for starships, except possibly in cases where the starship is designed for a specific route between two very similar worlds.
That doesn't mean streamlining and aerodynamic control surfaces are useless. Streamlining allows a faster trip through an atmosphere, and control surfaces make a vehicle more maneuverable in atmosphere, even if the maneuver drive is powerful enough that lift is unnecessary.
Does anyone see any fault in that conclusion?