Digital breast tomosynthesis (DBT) is an x-ray imaging technique that can provide three-dimensional reconstructions of the breast. Despite improvements over digital mammography (DM), current DBT systems continue to experience high false positive rates and additionally are subject to image reconstruction artifacts absent from DM. A next-generation tomosynthesis (NGT) system had been designed and developed by the X-ray Physics Laboratory (XPL) to overcome these limitations. The NGT system incorporates multi-directional acquisition geometries, which achieve improved spatial resolution and artifact reduction. This work investigates the relationship between the acquisition geometry and the spatial dependent image quality in DBT. This research highlights the inadequacies of traditional position-based geometric calibration methods when applied to NGT geometries. To address this, a novel line-based iterative calibration method was developed, which calculates system geometry using virtual line segments created by pairs of fiducials in a calibration phantom. The method is robust to all NGT system motions and is superior to traditional methods through simulation and experimental validation. Fourier analysis was employed to assess the sampling comprehensiveness of different acquisition geometries. The study found that geometries with broader angular coverage (i.e., NGT geometries) provide more comprehensive sampling and improved spatial isotropy. Experimental validation confirmed that NGT geometries offer superior image quality when compared to the traditional linear geometry, significantly reducing reconstruction artifacts and enhancing overall lesion detectability. This thesis also presents a comparative evaluation of ray-tracing and Monte Carlo virtual clinical trials pipelines for lesion detection in DBT. The results indicate that ray-tracing provides a quicker and more efficient alternative with comparable outcomes to Monte Carlo simulations, suggesting its potential for practical clinical applications. In conclusion, the advancements in acquisition geometries, calibration methods, and virtual clinical trials presented in this thesis represent significant steps toward improving DBT. These innovations hold promise for enhancing the precision of breast cancer detection and diagnosis, ultimately contributing to better patient outcomes and advancing the field of medical imaging.