No, you should not extend your power plane pours. In fact, you probably need to remove some of of your copper (all of it, pours or traces) from the top edge of the PCB.
This isn't about impedance matching, and unless you are actually doing manual impedance matching (estimating with calculations or via simulation, or both), then nothing is going to 'help with impedance matching'. The only thing that will help with that is carefully determining the values for the matching network, which requires taking into account not just the PCB and antenna, but the enclosure.
Does that antenna have a datasheet? It should give a keep out area. Besides the feed line and the antenna itself, no other conductive materials should be within the keep out area. For the best antenna radiation and reception, this keep out area should extend all the way along the 'long' axis, or orthogonal to the feed line. That is what determines if you should put any pours on the top tab.
So don't do that.
If there isn't a datasheet, then you'll have to estimate the keep out area yourself. The rule of thumb is to make the keep out area (which is a minimum distance from any part of the antenna to any other conductor) is twice the reactive near-field distance.
The reactive near-field distance can be approximated like this:
$$ r_{1} = 0.62 \times \sqrt{\frac{D^{3}}{\lambda}} $$
Where \$ D \$ is the largest dimension of the antenna (which would probably be the length of the helix's axis in your case), and \$ \lambda \$ being wavelength of course.
For 868-915MHz, rounding up, twice that amount gives you about a 6mm of minimum keep out distance. So make sure you are meeting that reqiurement, the antenna looks awfully close to those ground planes at the board edge.
But also keep the plane a straight edge along the length of the antenna, but spaced with that 6mm gap. This is also serving as the ground plane for the antenna.
Also, are you using a plastic enclosure? You will want to ensure you maintain that 6mm gap (or more if possible) between the antenna and any plastic enclosure. Plastics of all sorts have much higher dielectric constants than air, and if any get too close to your antenna, it will effectively increase the electrical length of the antenna and detune it, pushing the resonant frequency lower (and thus, your radiation efficiency as well).
Also, a helical antenna mounted like that without a proper ground plane normal to the long axis is hardly any better than a PCB helical antenna. This is because this antenna mounting, or a PCB antenna, are both limited mainly by the size of the ground plane. In both cases, the ground plane is just the edge of a copper pour parallel to the long dimension of the antenna.
So there is probably little advantage to using an actual coil of wire like that. If it were at a right angle to your PCB, that is a different story, but mounted like this, I think you would probably get better performance by following app note from TI to make a PCB helix antenna. That way you have better control over dimensions and spacing, have the matching network already done for you, you can be sure you have things at their optimal position, etc.
Oh, and the best part about PCB antennas is that they're free. Unless you pay for boards by the square inch, then they're still probably cheaper than buying an actual physical antenna part. Also the antenna in that app note is teeny tiny but still quite good (64% EB efficiency typical).
