Better way to integrate a gmdistribution.fit object class?

I see that objects of gmdistribution class have a cdf method that can be used indirectly to integrate Gaussian mixture models.

Is there a faster, more efficient way to integrate a gmdistribution object over an n-dimensional rectangular region of space compared to evaluating the cdf at the corners of the region and adding/subtracting those values?