Fig. 2

Schematic of the IVIVE procedure followed. Our approach included initial literature search and data extraction, followed by a six-step procedure to determine the feasibility of extrapolating between in vitro markers and in vivo fibrosis. First we aligned the in vivo and in vitro dosimetry by converting in vivo dosimetry to mass per surface area of tissue (Step 1) and then converted the mass per surface area dosimetry to units that reflect the surface-driven effect (i.e. cm²/cm², Step 2). Next we compared endpoints by first determining if there is a correlation between PMN influx and fibrosis in vivo (Step 3), investigating the relationship between fibrosis and in vivo study protocols (Step 4) and then comparing PMN influx in vivo with cytotoxicity and inflammatory endpoints in vitro (Step 5). Step 5 was further split into two possible approaches, the first (Step 5a) involved dose comparisons at a central point of the curve e.g. by EC50 analysis, when in vivo and in vitro responses were aligned (P < .05). A second approach (Step 5b), which in theory can be applied to any curve shape, used a point-of-departure (PoD) analysis of the curves using BMD analysis. Finally, we studied the applicability of the method for inferring lung fibrosis from promising cytokine/chemokine endpoints in vitro using nano-CeO₂ as a case study material