Note that the OP's crank is tricky in regard to having the small ring attach to the large ring, instead of to the spider. This complicates assuming that the spider is straight just because the small ring may be running straight.

Note also that a longer wrench, clamped onto the outer hoop/ring of the big ring, will effect more of a pure twisting force, which would better localize the corrective bending to the outer hoop versus the spider. This is what I would do here if the small ring and spider appear to be running true.
This is the opposite of what I usually do while truing chainring runout on cotterless alloy cranks, where I typically use a 3# hammer and wooden dowel drift to impact directly on the chainring bolts or nuts. Alloy chainrings tend to be both strong and springy, and bending them with a wrench can snap the tabs off of a cast crank's spider.
But there are always a lot of different possibilities why a chainring doesn't run true, hence my deferral to francophile's methods above.