Two points: commercial vehicle manufacturers such as Peterbilt and Kenworth do tilt-table testing to determine the point of incipient rollover, and therefore the cg height.
Second, the government's so-called "roll-over ratings" are as Greg says, "schoolboy physics" and are worse than meaningless in the real world. They consider ONLY cg height and track width.
For example, a Corvette has the best roll-over rating according to NHTSA, but per an analysis of crash data done by a writer for Car Driver magazine, the Corvette is something like 187 times more likely to be involved in a roll-over crash than the much-maligned Explorer. And for that matter, the two-door Explorer was 50 times more likely to overturn than the four-door, when both have identical NHTSA ratings.
According to his analysis, roll-over correlated to the number of doors better than to any vehicle dynamics factor.
i.e., a two-door vehicle is much more likely to be driven in a fashion that would produce a roll-over than is a four-door.
Sorry, this doesn't have as much to do with commercial vehicles as you might like. Without a study of the trailer, since most commercial vehicles have semi-trailers, I don't think you could make any meaningful conclusions about comm vehicle dynamics.
And load-shifting is probably more of a factor in roll-overs than any vehicle characteristic you can name.