Thinking out loud, again . . .
Woody, your research would want me to be looking at that 2.5% increased port, to be looking at where that flow is happening in the port and bowl with respect to the valve event, and
how it's filling the cylinder past the valves. You've likely already have done that.
What if you're not necessarily looking to maximize
total flow, but instead, are looking to create a less turbulent, and possibly better directed flow?
IO, given the external constraints of the port, you're effectively creating a low pressure point in the airstream where the port/bowl narrows around the valve guide, but that would begin to recover past the widest point. An airfoil design could be engineered to be less turbulent at that point where the airflow is entering the combustion chamber, no?
If you have a port size capable of delivering flow capacity in excess of the application's needs, it seems to me the shape of the obstructions are fair game.