I didn't simulate your structures until now. However, my first impression is: your port (in this case coaxial) is wrongly setup in both cases. Moreover, in CST you draw the airbox which is not needed (by defining open boundary it is already defined).
I'll change the structure to see the results (I'm using HFSS 11.1 and CST 2009).
In the paper, it is written 50 ohm microstrip line. However, if you use air as substrate, 16 mm line will not give you 50 ohm at 1.8GHz frequency. Would you please upload the full paper, so that we know the substrate.