then you can treat it as the superposition of two lamda/quarter antennas.Perhaps you can quickly solve this Problem using the antenna module in AMANOGAWA (its free and available on the web..but does not run well on a MAC..and make sure that you have the latest version of JAVA.).You may need to intropduce a virtual groundplane under 45 deg..
It works perfectly well. The radiation pattern is no longer circularly symmetrical, because in the plane of the bend the whole antenna is not in phase, so there is a bit more gain in the direction normal to the plane the wires are in. The nulls at the poles are filled in, too.
It is still pretty much a dipole antenna. Try building one or modelling it and you will see. With two 7.5 mm arms 1/2 mm diameter it is resonant close to 10 GHz as would be expected.