Why is it that in a cascode LNA, only the input transistor contribute to the noise?

Why is it that in a cascode LNA, only the input transistor contribute to the noise, while the cascode transistor doesn't? Thank you.

If the parasitic capacitance of the node between cascaded LNA?s is high, the noise contribution of the second LNA could be significant at high frequencies.

Thanks for the reply, vfone.

In CMOS LNA, since noise current is proportional to gm, the cascode transistor's gm is degenerated by input transistor, so I suppose that explains why it contributes less. But how about bipolar LNA? The shot noise is 2qI. How does it work over here? Thanks.

you can find the answer in the Razavi's book.