With the caveat that it is easy to make an arithmetic blunder, I calculate that there is around a 1 in 25 chance per 6-month interval that planet 9 might cause a significant microlensing amplification of a star brighter than 21st magnitude. The problem is that nobody is really monitoring 21st-magnitude stars over a wide area on the sky, with a cadence sufficiently fast to catch the few-minute duration of such an event.
On the other hand, a measurable gravitational deflection is about 100 times more likely. i.e. Planet 9 is likely altering the positions of a few stars in its vicinity by a sufficient amount to be detectable by Gaia. By sifting through epoch-by-epoch positional measurements taken by the Gaia satellite we might be able to see its track across the sky in the form of a wave of perturbed stellar positions.
Details
(1) Microlensing Amplification
The size of any microlensing is characterised by the Einstein angle, given by
$$\theta_E = \sqrt{\frac{4GM}{c^2}\left(\frac{d_s - d_l}{d_s d_l}\right)}\ , $$
where $M$ is the mass of the lensing object, $d_s$ is the distance to the background star and $d_l$ is the distance to the lensing object.
In this case $d_s - d_l \simeq d_s$ and so
$$\theta_E = \sqrt{\frac{4GM}{c^2 d_l}}\ . $$
If we assume $M \sim 6 M_{\rm Earth}$ and $d_l \sim 500$ au, then $\theta_E \sim 0.008$ arcseconds. In order to get significant amplification then the background star and the planet need to be separated by an angle of this order or smaller.
There are about a billion catalogued stars down to around 21st magnitude, spread all over the 41,253 square degrees of sky, with positions measured to this kind of accuracy by Gaia, and with parallaxes and proper motions so that their sky positions can be predicted with this sort of precision into the (near) future. Thus the density of such stars is about 0.0019 per square arcsecond. At any one time, the chance of planet 9 being within its Einstein angle of a Gaia star is roughly $\pi (0.008)^2 \times 0.0019 \simeq 4 \times 10^{-7}$ (less than 1 in a million).
Of course, planet 9 will move. If we take a first-order approximation that it moves and the background stars are fixed, then it gets to roll the dice again every time it moves by 0.008 arcseconds on the sky. How fast does it move? At 500 au it will move at about 1 km/s, so to first order, its tangential motion is entirely due to the 30 km/s orbital motion of the Earth. In 6 months it will execute a parallax motion of about 800 arcseconds, so you have about $10^5$ chances for the Einstein angle to intercept a star and thus about a 1 chance in 25, over 6 months, that it will cause a microlensing event of some sort.
However, let me emphasize - this will be basically the chance that it is close to a star that is probably in the magnitude range 19-21, where the vast majority of these Gaia stars are. Nobody at the moment is monitoring all these stars to see if there is a microlensing event and since there is no well-determined likely location for planet 9, then a chance detection of this microlensing seems remote. The amplification event would only last a few minutes and so the cadence of any monitoring operation would have to be very high. It isn't going to be seen at the cadence of individual Gaia epochs (roughly 1 observation per month).
(2) Gravitational Deflection
Maybe a better bet is to look for systematic distortions of stellar positions for stars located close to planet 9 in the sky. These distortions are similar to those of stars being gravitationally lensed by the Sun. The deflection angle is approximately given by
$$\theta_D \simeq \frac{4GM}{c^2 r}\ , $$
where $r$ is the distance of the closest approach of the light ray to the limb of planet 9.
Gaia is capable of measuring stellar positions to something like 1 milliarcsecond in a single visit for most stars it observes, and at least 10 times better for brighter stars. To get a deflection of this size, that could be compared with the average, and much more precisely determined, position of the star taken over the whole Gaia mission, would require $r \leq 2 \times 10^7$ m, corresponding to an angle on the sky at 500 au of 0.06 arcseconds and close to the limb of a Neptune-sized planet.
It is therefore about 100 times ($0.06^2/0.008^2$) more likely that a star will get close enough to planet 9 to cause gravitational deflection of measurable size than to cause significant microlensing amplification. Thus planet 9 would likely cause measurable deflections for several stars over the course of 6 months.
I would guess that it might be possible to sift through the epoch-by-epoch data from Gaia, looking for significant, transient, gravitational deflections and I would expect that several groups are probably gearing up to do that kind of analysis once the mission and its data processing nears completion.
(3) Occultation
Finally, there is the possibility that planet 9 will "eclipse" a background star, causing it to dim considerably. If its radius is of order $2 \times 10^7$ m (Neptune-sized). This is exactly the same size/angle as for the gravitational deflection calculation above, so the probability of it occuring is similar - i.e. maybe a few faint stars every 6 months. Again, the problem is, that these eclipses would last of order tens of minutes to 1 hour and nobody is monitoring the brightness of such faint stars over a wide area with that kind of cadence.
