Cancer potency estimates were derived for 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF) using data collected from the recently published National Toxicology Program bioassay in female Sprague-Dawley rats. By using a toxicokinetic model for 4-PeCDF, the dose-response relationship for combined liver tumors (hepatocellular adenomas and cholangiocarcinomas) in rats was assessed in terms of lifetime average liver concentration and lifetime average adipose concentration with data from both the lifetime and the stop-exposure components of the bioassay. Benchmark dose modeling was performed to estimate tissue concentrations at two points of departure (EC(10) and EC(01) and their 95% upper and lower confidence limits). The same toxicokinetic model with human input values was then used to back-extrapolate human equivalent doses that corresponded to the internal tissue concentration measures at the points of departure. Information regarding the cancer mode of action was used to support the development of several toxicity criterion values based on a nonlinear method, e.g., reference dose or tolerable daily intake. Nonlinear estimates of toxicity criteria based on observed noncancer toxic events as possible precursors to tumor formation were also derived and were similar in value to those based on combined liver tumors. For comparison purposes, linear estimates of cancer potency were also derived.