ADE XL: chosen instances ignored in statistical corners

Hi! I'm trying to create statistical corners for a Monte Carlo simulation where I only consider variations on a couple of instances. The Monte Carlo run goes fine and only includes mismatch from the selected instances, but the Statistical Corners created from it are ignoring this and are including mismatch from ALL the instances in my tesbench!

I can see that the Statistical Corners generated have entries like the following:

monteCarlo::param::Instances (set to: "my_test_name%/my/instance/one, /my/instance/two%Schematic%Schematic")

but the actual netlists produced don't include any "dut=" entries for the montecarlo analysis. This is in contrast to the original Monte Carlo run, where all points have netlist entries like the following:

mc1 montecarlo numruns=1 seed=123 variations=all sampling=lds \

donominal=no scalarfile="../monteCarlo/mcdata" paramfile="../monteCarlo/mcparam" saveprocessparams=no savemistmachparams=no \

dut=[my.instance.one my.instance.two] savefamilyplots=yes savedatainseparatedir=yes wfseparation=yes firstrun=96

...whan can be going wrong? This is Spectre 17.1.0.307.isr6 64bit running over ICADV12.3-64b.500.22.

Thanks and regards, Jorge.