In the \textsc{2-Dimensional Knapsack} problem (2DK) we are given a square knapsack and a collection of $n$ rectangular items with integer sizes and profits. Our goal is to find the most profitable subset of items that can be packed non-overlappingly into the knapsack. The currently best known polynomial-time approximation factor for 2DK is $17/9+\varepsilon<1.89$ and there is a $(3/2+\varepsilon)$-approximation algorithm if we are allowed to rotate items by 90 degrees~{[}Gálvez et al., FOCS 2017{]}. In this paper, we give $(4/3+\varepsilon)$-approximation algorithms in polynomial time for both cases, assuming that all input data are {integers polynomially bounded in $n$}. Gálvez et al.'s algorithm for 2DK partitions the knapsack into a constant number of rectangular regions plus \emph{one} L-shaped region and packs items into those {in a structured way}. We generalize this approach by allowing up to a \emph{constant} number of {\emph{more general}} regions that can have the shape of an L, a U, a Z, a spiral, and more, and therefore obtain an improved approximation ratio. {In particular, we present an algorithm that computes the essentially optimal structured packing into these regions. }