$X,Y$ are two r.v. $(\Omega,\mathcal{A},\mathbb{P}) \rightarrow (\mathbb{R},\mathcal{B}(\mathbb{R}))$ and have joint density wrt to $\lambda^2$, the two dimensional lebesgue measure.
So $f_X(x) = \int_{\mathbb{R}} f(x,y)\lambda(dy) > 0$ for $\mathbb{P}^X$-a.s. $x \in \mathbb{R}$. And $f_X^{-1}(x)$ ist the density for $\lambda$ absolutely continuous wrt $P^X$.
is then $\int_A \mathbb{P}[X \in dx] = \int_A f_X(x)\lambda(dx)$ ?