Hi
S11 and S22 or Sxx represent the same thing, the ratio of the reflected voltage/power to input voltage/power at a given port.
Usually when you only have one port you use S11. Using S22 usually indicates a having at least two ports. But you can donate a port whatever you want. The antenna in your figure seems to be a 1-port device, but they have donated the reflection coefficient S22, which is not the usual way. I would named it S11 otherwise you could be trying to find the 1-port.
But to your original question, S22 (S11) is the reflection coefficient at your port.
A simulation program like CST or any other EM-simulator do not calculate the gain of the antenna as S21.
You can use S21 or S12 if you have at least two ports in the antenna structure. But this only show the transmission between the ports and are not linked to the gain of the antenna.
You need to setup the simulator correctly with boundaries. The gain is then calculated by the program can give you the gain or other antenna properties directly.
Hope this helps.