Tracer Kinetic (TK) modelling stands at the forefront of model-based analysis for dynamic contrast-enhanced images, offering a robust framework to link observed contrast agent kinetics to targeted hemodynamic variables such as blood flow rate or vascular permeability. While being conceptually derived from pharmacokinetics (PK), which addresses extended temporal domains (hours, days) and macroscopic anatomical levels (an organ), TK focuses on shorter temporal scales (seconds, minutes) and more localized spatial scales (a small image region). In particular, one of the founding assumptions in PK modelling is the well-mixed hypothesis [1], which assumes a slow-enough evolution of the concentration bolus in the region of interest. Such scale disparities thus raise concerns regarding the physical validity of TK modeling, challenging the reliability and clinical meaning of the estimated hemodynamic parameters. In our work [2], we mathematically formulated three TK models, based on mass-conservation principles, clearly exhibiting the assumptions underlying their derivation. Through numerical simulations of the most comprehensive model, we evaluated the validity of these assumptions across diverse parameter values, assessing the conditions required to justify the model’s simplified alternatives. Our findings reveal that the robustness of the TK models’ assumptions diminishes with increasing dimensions of the region of interest or vascular permeability. For typical Gadolinium-based MRI contrast agents, the upper validity threshold for the region of interest approximates 4 mm in low permeability settings, dropping below 250 μm in highly permeable vessels. These validity metrics are not intended for direct clinical application; rather, they highlight potential limitations in model interpretability within established DCE-MRI protocols. By delineating the theoretical boundaries of current model-based approaches, this study underscores the risks of misinterpretation when hemodynamic parameters extend beyond the defined ranges. Ultimately, we propose enhancing model interpretability by accounting for apparent diffusion of the contrast agent, breaking ground for more reliable and clinically meaningful insights.