Defects induced by the fabrication process, which are dependent on the input process parameters, remain a critical issue in the fatigue design of industrial components processed by Laser Powder Bed Fusion (L-PBF). The present work aims to use numerical simulations to assess the impact of typical L-PBF defects on the High Cycle Fatigue strength of a high-strength material. To do so, real defect geometries have been explicitly integrated in the simulations, using X-ray tomography observations of different defect populations of Ti64 alloy processed by L-PBF. The fatigue strengths of the different defects have then been determined numerically for different types of loadings (tension, torsion, and tension-torsion) by applying the Crossland multiaxial fatigue criterion. A non-local analysis of the calculated stress fields was also included in the calculation of the Fatigue Indicator Parameter in order to account for the severe stress gradients at the vicinity of defects [F. Morel et al., International Journal of Fatigue, 2009; A. Karolczuk et al., Computational Materials Science, 2008; Y. Nadot & T. Billaudeau, Engineering Fracture Mechanics, 2006]. With this approach and by considering different loading directions, the fatigue strength anisotropy was assessed for two defect populations (gas pores and LoF defects) for the considered loading types.