Titanium dioxide (TiO2) has been characterized as a poorly soluble particulate (PSP) with low toxicity. It is well accepted that low toxicity PSPs such as TiO2 induce lung tumors in rats when deposition overwhelms particle clearance mechanisms. Despite the sensitivity of rats to PSPs and questionable relevance of PSP-induced tumors to humans, TiO2 is listed as a possible human carcinogen by some agencies and regulators. Thus, environmental toxicity criteria for TiO2 are needed for stakeholders to evaluate potential risks from environmental exposure and regulatory compliance. A systematic review of the literature was conducted to characterize the available data and identify candidate datasets upon which toxicity values could be derived. Key to this assessment, a survey of mechanistic data relevant for lung cancer was used to support quantitative inhalation risk assessment approaches. A total of 473 human studies were identified, 7 of which were epidemiological studies that met inclusion criteria to quantitatively characterize carcinogenic endpoints in humans. None of these studies supported derivation of toxicity criteria; therefore, animal data were used to derived safety values for TiO2 using different dose-metrics (regional deposited dose ratios, TiO2 particle surface area lung burden, and volumetric overload of alveolar macrophages), benchmark dose modeling, and different low-dose extrapolation approaches. Based on empirical evidence and mechanistic support for nonlinear mode of action involving particle overload, chronic inflammation and cell proliferation, a no significant risk level (NSRL) of 300 μg/day was derived. By comparison, low-dose linear extrapolation from tumor incidence in the rat lung resulted in an NSRL value of 44 μg/day. These toxicity values should be useful for stakeholders interested in assessing risks from environmental exposure to respirable TiO2.