Page 161 - 20dynamics of cancer
P. 161

146                                                 CHAPTER 8

                              final number of cells by around 60 months of age, causing cell division
                              to slow and reducing the risk per cell with age. Change in overall risk
                              with age depends on the opposing effects of the rise in cell number and
                              the decline in the rate of cell division.
                                Hethcote and Knudson (1978) developed a mathematical theory based
                              on cellular processes of retinal development, and fit their model to an
                              extended set of data on inherited and sporadic retinoblastoma. The
                              basic pattern in the data remains the same as in Figure 8.1, but the later
                              model fits parameters for the somatic mutation rate and for aspects of
                              cell population size and cell division rate.
                                At first glance, the realistic model based on cell populations and cell
                              division may seem attractive. However, many factors affect the inci-
                              dence of human cancers, including environment, cell-cell interactions,
                              tissue structure, and somatic mutations during different phases of tu-
                              mor development. No model can account for all of those factors, and
                              so incidence data can never provide accurate estimates for isolated pro-
                              cesses such as somatic mutation rate or cell division rate.
                                Knudson’s main insight was simply that age-specific incidence of in-
                              herited and sporadic retinoblastoma should differ in a characteristic way
                              if cancer arises by two hits to the same cell. He obtained the data and
                              showed that very simple differences in incidence do occur. The next step
                              is to understand why the observed differences follow the particular pat-
                              terns that they do. Detailed mathematical theory based on cell division
                              and mutation rate provides insight about the factors involved, but with
                              regard to data analysis, that theory depends too much on the difficult
                              task of estimating parameters of mutation and cell division from highly
                              variable incidence data.



                              RATIO OF SPORADIC TO INHERITED INCIDENCE
                                I advocate theory more closely matched to Knudson’s original insight
                              and to what one can realistically infer given the nature of the data (Frank
                              2005). According to Knudson’s theory, bilateral tumors arise from single
                              hits to somatic cells with an inherited mutation. The rate at which a
                              hit occurs in the developing retina at a particular age depends on many
                              factors, including the number of target cells and the rate of cell division.
                              But we cannot get good estimates for those factors, so let us use the
                              observations for bilateral cases at different ages to estimate the rate
   156   157   158   159   160   161   162   163   164   165   166