I need to calculate or at least estimate the expected value of this term
$$ \mathbb{E} \left[\frac{1}{(L-K \cdot\left \| \mathcal{N}(0,\Sigma) \right \|_2)^2}\right] $$
where $L,K\in \mathbb{R}>0$ are constants and $\left \| \mathcal{N}(0,\Sigma) \right \|_2$ stands for the 2-Norm of a zero mean multivariate normal distribution. Since $\mathbb{E}[1/X] \neq \mathbb{E}[1]/\mathbb{E}[X]$ for almost every case, I'm stuck on how to approach this problem.