A general probabilistic model of carcinogenesis: analysis of experimental urinary bladder cancer.

A theoretical model of two-stage carcinogenesis has been hypothesized. Variables that are modeled include the populations of normal, initiated, and transformed cells; mitotic rates of these cells; hyperplasia; and the probabilities of cell initiation and transformation during replication. The size of the cell populations can be estimated and mitotic rates determined directly from animal studies. Tumor occurrences at different time intervals following varying periods of N-[4-(5-nitro-2-furyl)-2-thiazolyl]formamide (FANFT) and sodium saccharin administration are known and are used in the indirect estimation of values for unobservable model variables. Model-based analyses suggest FANFT markedly increases the probability of cell initiation in addition to its experimentally verifiable effects on increasing the stem cell population and mitotic rates. Further, experimental results appear inconsistent with the hypothesis that FANFT increases the probability of cell transformation over background levels. Similarly, the effect of sodium saccharin was found to be attributable entirely to increases in stem cell populations and mitotic rates without influencing either the probability of initiation or the probability of transformation.