Introduction: Accurate diagnosis and personalized treatments involving site-targeted cancer localization, drug delivery, therapeutic strategy, and disease pathways identification, rely on a precise understanding of biomarker kinetics, drug pharmacokinetics, and mechanistic behaviour of functionalized tracers through in vitro and in vivo studies. X-ray fluorescence (XRF) computed tomography (XFCT) offers a potential alternative to current 3D imaging techniques for spatiotemporal localization of nanoparticle-tracers with high spatial resolution and sensitivity. In this work, the applicability of a benchtop cone-beam system with a polychromatic X-ray source was examined with regard to physical constraints of engineered tissue models.Methods: A tissue engineering approach based on a decellularized scaffold was used to establish a 3D breast cancer model with MDA-MB-231 cells in co-culture with primary human fibroblasts. The 3D breast cancer system, in combination with small-animal-sized phantoms, was used to demonstrate the novel integrated pre-clinical imaging approach to perform in vitro surrogate investigations and non-destructive analysis on biophantoms. These models are adopted to evaluate the functionality and optimize the setup for high-spatial-resolution, fast, and fully-3D quantitative imaging. Polychromatic X-rays from a microfocus source are used for XRF stimulation from conventional Gadolinium (Gd) and nanoparticle-based Molybdenum (MoNPs) contrast agents.Results and Discussion: The intestinal scaffold allowed the invasion of the breast cancer cells over this barrier and therefore provides a valuable tool to study metastasis formation of tumor cells from epithelial origin. The breast cancer model was well suited for the development and validation of the proposed XRF imaging, with spatial resolution under <2 mm and contrast dose in the order of a few 100 μg/mL (∼0.3 mg/mL for Gd and ∼0.5 mg/mL for MoNPs), radiation dose in the order of a few 100 cGy (280 cGy for Gd and 94 cGy for MoNPs, with a possible reduction of an order of magnitude for Gd and 67% for MoNPs), and imaging time in the order of 10 min for Gd (33 min total) and 100 min (2.8 h total) for MoNPs, approaching in vivo conform conditions for pre-clinical studies. High-resolution XFCT for tissue-engineered cancer models would be of significant interest in biomedical research and diagnostic imaging, e.g., for an increased mechanistic understanding of molecular processes in tumor formation or early cancer detection.