This work develops a mechanistic understanding and a simplified model to understand changes in drug release profiles of an active pharmaceutical ingredient in a long-acting parenteral polymer formulation from changes in its microstructure determined using X-ray computed tomography (XRCT) imaging. The system studied is an implant composed of a solid dispersion of crystalline drug microdomains embedded in a polymer matrix and fabricated by hot melt extrusion. We conduct material characterization of these devises using microCT and introduce, for the first time, imaging of such pharmaceutical devices by means of nanoCT, showing finer structures that cannot be captured by ordinary CT methods. We propose a simplified theoretical mechanistic model that estimates the drug release by extracting relevant microscopic features that modulate the release rate from microCT images to construct an idealized geometry. Using this analysis, we found that although the primary mechanism of resistance was originating from submicrometer features, which are below the resolution of the microCT, the main features that determine changes in diffusivity as a result of different process conditions were microscale pore volume and pore connectivity distributions. The parameters were extracted from the microCT images and in conjunction with the theory were able to explain changes in experimental drug release under different process conditions. These results demonstrate that the modeling strategy can explain the dominant release mechanism and estimate drug release of formulations, even when the equations are simplified due to many assumptions and a lack of subscale information. The proposed approach, due to its simplicity and reliance on the widely employed microCT technique, holds promise for guiding formulation development by providing a rapid initial assessment of batch consistency and process parameter selection before embarking upon lengthy and costly clinical trials.