Microwave ablation is a minimally invasive modality increasingly being used for thermal treatment of cancer in various organs. During ablation procedures, treatment planning is typically restricted to vendor specifications of expected ablation zone volumes based on experiments in unperfused ex vivo tissues, presuming parallel insertion of antennas. However, parallel antenna implants are not always clinically possible due to the restricted control of flexible antennas and presence of intervening organs. This paper aims to quantify the effect of non-parallel antenna implants on the ablation volume. 3D electromagnetic-bioheat transfer models were implemented to analyze ablation zone profiles created by dual antenna arrays. Parallel and non-parallel implants spaced 10-25 mm with antenna tips deviated to create converging or diverging configurations were analyzed. Volumetric Dice Similarity Coefficients (DSC) were calculated to compare ablation zone volumes for parallel and non-parallel configuration. Antenna tip displacements of 3 mm/antenna yielded an average DSC of 0.78. Tip displacements of 5 mm/antenna yielded a DSC of 0.78 and 0.64 for 15 mm and 20 mm antenna spacing, respectively. For ablation with dipole antennas as the frequency of operation decreases from 2.45 GHz to 915 MHz the similarity between the ablation zones for parallel and angled cases increased significantly. In conclusion, ablation volumes with non-parallel antenna implants may differ significantly from the parallel configuration. Patient-specific treatment planning tools may provide more accurate predictions of 3D-ablation volumes based on imaging data of actual implanted antenna configurations. Methods to compare ablation zone volumes incorporating uncertainty in antenna positions and experimental results to validate the numerical modelling are also presented.