Background and Purpose: Di-isononyl phthalate (DINP) is a high molecular weight phthalate widely used as a plasticizer in flexible PVC products. Chronic bioassays have indicated that DINP induces liver tumors in mice and rats. Rodent liver tumors are hypothesized to occur via the rodent-specific peroxisome proliferator-activated receptor alpha (PPARα)-activator mode of action (MOA). PPARα activation is the first and only key event shared across species; the downstream biological responses leading to tumor formation are widely considered to be unique to rodents and not relevant to humans. To further inform whether rodent liver tumors observed following DINP exposure are occurring through the rodent-specific PPARα MOA, a comprehensive assessment of the evidence base for DINP was conducted. Methods: The animal and human evidence base for DINP was evaluated in the context of the key events (KEs) and associative events outlined in a published MOA framework for PPARα activator-induced rodent hepatocarcinogenesis. The PPARα MOA for rodent liver tumors includes four KEs: (KE1) PPARα Activation, (KE2) Alteration of Cell Growth Pathways, (KE3) Perturbation of Cell Growth and Survival, and (KE4) Clonal Expansion. First, the weight of the evidence (WoE) for each KE and liver tumors (the adverse outcome [AO]) was evaluated for DINP and its metabolites. Next, the human relevance of this MOA and the potential for alternative MOAs for DINP-induced liver tumors were critically examined. The analysis incorporated aspects of the modified Bradford Hill Criteria and characterization of data gaps and uncertainties. Studies from the peer-reviewed literature and unpublished laboratory reports related to DINP and liver carcinogenesis were first identified from existing authoritative assessments of DINP, followed by literature searches in PubMed to identify literature pertaining to DINP, its metabolites and each KE and potential alternative MOAs. Results: WoE for KEs 1 and 3 and the AO strongly support the PPARα MOA for rodent liver tumors. Although limited data were available to inform KEs 2 and 4, these events can be reasonably inferred given the strong evidence for the other KEs and AO. Regarding alternative MOAs for liver tumors, evidence for genotoxicity and the aryl hydrocarbon receptor were not supported. DINP activation of PPAR gamma and constitutive androstane receptor/pregnane X receptor were weakly supported based on limited in vitro studies. Evidence possibly consistent with cytotoxicity (e.g., elevated serum liver enzymes and focal necrosis) occur at later timepoints and at higher DINP doses than those that elicit PPARα activation—indicating they are likely secondary to the PPARα MOA. Rodent liver tumors resulting from a PPARα MOA are generally regarded by the scientific community to lack human relevance. Comparative data from human and non-rodent mammalian models indicate decreased sensitivity and responsiveness to DINP and its metabolite MINP compared to rodents. Limited data for humans (in vitro) and monkeys (in vivo) suggest perturbation of cell growth and survival is not elicited by DINP or its metabolite. Further, no evidence was identified regarding DINP exposure and liver cancer in humans. Collectively, DINP-specific evidence supports a lack of human relevance for this MOA. Conclusions: Overall, the WoE for DINP strongly supports that the liver tumors observed in rodents following DINP exposure are occurring via the rodent-specific PPARα MOA. Moreover, based on quantitative and qualitative species differences in KEs of this MOA, the WoE indicates liver tumors observed in rodents following DINP exposure are likely not relevant to humans.