As far as microlensing is concerned:
Not likely
This is a bit of a dodge on the question, but it is hard to come up with a scenario where microlensing becomes the dominant PSF term.
Microlensing occurs during conjunction with a stellar or planetary object, and results in a momentary (seconds-to-days timescale) increase in the brightness of the observed star. It is an extremely rare event, with rates around 100 in 17 million stars in the most recent surveys.
Stellar densities ($\rho$) around our solar system are on the order of 0.14 stars per cubic parsec. With such a low number density, microlensing does not pose any practical limitation on the resolution of radio telescopes. Lets do some scratch-pad estimates to see:
If we assume a synthetic aperture ($D$) of 1AU ($4.84\times10^{-6}$parsecs) , a wavelength ($\lambda$) of 1mm (similar to the view of BetelgueseBetelgeuse by the ALMA observatory) , and a desired spatial resolution ($s$) of 10,000 km at a distance $L$:
$$ \begin{align} \frac{s}{L} &= 1.22 \frac{\lambda }{D} \\ L &= \frac{sD}{1.22 \lambda } = 40\;\text{kilo-parsec} \end{align} $$
This is roughly the diameter of the milky way galaxy. So an interferometer operating in the 1mm wavelength with a 1AU baseline has a 10,000 km resolution at a distance of 40 kpc.
How many stars are between the 1AU-wide observatory and the target star? Using a cone of base diameter $D$ and height $L$ as a volume and the above stellar density, we get the average number of stars ($N$) between our aperture and the observed star: $$ N = \rho \frac{\pi}{12}D^2 L = 3 \times 10^{-8} \; \text{stars}$$
You would have to ramp $D$ up by a factor of $10^4$ before $N \approx 1$ and 'microlensing aberrations' were more than just an extremely rare event.