“…Despite the high level of maturity of deterministic transport methods, there are some serious drawbacks that prevent its use for the most challenging problems: (i) they can be very expensive in three dimensional geometries requiring fine spatial meshes even for moderate angular expansion orders [8,9] and (ii) the solutions often exhibit a lack of angular resolution, especially when considering charged particle transport with highly forward peaked scattering kernels [10,11] (as encountered in proton or carbon radiotherapy). Even a brute-force increase of the number of directions is no cure for these types of problems.…”