Would the sharp edge of a corner be a slightly better antenna thereby promoting coupling to adjacent parallel tracks like those in a multilane bus?
Ran some quick simulations
4/4mil track space 2ns rise times, same layer stackup, 2 tracks on top layer
The drive ends of the tracks start far away in one instance the approach is 45 deg the other 90, run together for about 0.4" and then seperate
The 45 degree approach, with a longer parallel run required 240mV threshold setting to avoid coupling indication
90 degree approach, with a shorter parallel run required 450mV threshold setting to avoid coupling