You are I believe 1 of 1 who in over a decade has mentioned this concern, at least it is rarely brought up. It has always been odd to me the preoccupation with all the gussets, bits, welding etc added to the K member area, but all the torque/twist of whatever TB is used gets 99.9% transmitted back to the crossmember in question, and no one bats an eye to that area. A normal cage does little IMO to stiffen this crossmember where it counts. The best I can tell is 60's era nascar mopars stiffened this area with basically another crossmember upside down, and directly over the OEM crossmember, among other things. If you watch closely the Ring cuda with the 1.24-6" TB's video here, you will see my comment on if with these new larger TB's, were we seeing a lot of deflection in this area at race speed. That was ruled out by the OP.

First suggestion, unless a TB is securely pinned at one end, its spring rate is reduced, stating the obvious, and anything that increases the depth of the crossmember in this area, increases stiffness.