Deflection correction:
Microwave beams, even the most tightly focused beam from the large 70 meter dishes of the Deep Space Network are still physically wide when they get to their destination, so relativistic deflection effects would not be big enough to require corrections, although I'm assume these effects are included in the pointing algorithms anyway.
As @DavidHammen notes in this comment and @DuffBeerBaron noted in this answer there is increasing use of Very Long Baseline Interferometry (VLBI) and Delta-DOR to try to obtain angular measurements of spacecraft, and those might be subject to some relativistic deflection.
Light time correction (non relativistic):
However the light time correction is huge, though not considered relativistic in nature. You have to point in the direction that the spacecraft will be in the future or the past. If for example the spacecraft is out near Jupiter, an incoming signal will come from the direction in space where Jupiter was about 45 minutes ago and if you want to send a signal, you point where it will be about 45 minutes later. However correcting for the finite speed of light does not really count as a relativistic correction.
Situations where relativistic effects are/will be important::
1) Radio and in the future, optical links to spacecraft usually include navigation signals like delay-doppler for precise tracking as well as data and telemetry channels, and relativistic effects will be much more significant for these. See the questions and their answers:
2) In the future, a navigation satellite constellation may be deployed at Mars, equivalent to GPS/GNSS for Earth. Relativistic effects are absolutely critical when using those signals to get a location fix. See for example the questions and answers at:
3) Missions close to the Sun will experience much larger relativistic effects as well. See the currently nicely but incompletely answered question: