LTSpice voltage and current phases in matching networks

I am analyzing the voltage and currents of driving a helical coil with a RF generator through a matching network. The coil is modeled as RLC elements with winding resistance and self inductance and capacitance. The generator is an ideal voltage source with a 50 Ohm resistive source impedance. The matching network is a shunt C then series C viewing from the input side of the matching network.

I have determined the matching network component values and verified it on a online Smith chart tool. The reflection coefficient is <0.01 magnitude. Then I verified the impedance match in LTspice by performing AC analysis and plotting V(t)/I(t) at port 1 of the matching network. Zooming into the operating frequency, the impedance is approx. 50 Ohm magnitude and <1 deg phase, i.e. 50 Ohm real and very small reactive. 

However, when I perform transient analysis, I find the voltage and currents are not in phase at all, on this one particular run current lag the voltage by 158 degrees. Can anyone explain why the voltage and current are not near in phase?

In my investigation, I learnt that LTSpice even the simple RLC elements are polarized or orientation specific. The direction of current is opposite if you flip the component 180 degrees. I have double checked that the component orientations are what I believe is the direction of current flow. 

Thanks, Denny