Page 115 - 20dynamics of cancer
P. 115

100                                                 CHAPTER 6

                                      6.3 Parallel Evolution within Each Individual

                                The model in the previous section assigns each individual in the popu-
                              lation to a particular stage of progression. Sometimes, it may make more
                              sense to consider the stage of particular cells or tissue compartments
                              within a single individual. Different components may be in different
                              stages of progression.
                                I described in Chapter 3 how colorectal cancer initiates in individ-
                              ual crypts, perhaps with mutations that occur to a particular stem cell
                              within a crypt. So we might choose to focus on different stages of pro-
                              gression in different crypts or stages of progression in different stem cell
                                                                  7
                              lineages. The human colon has about 10 crypts, and a slightly higher
                              number of stem cell lineages, so each individual has many parallel, in-
                              dependent lines of progression.

                                                          PR ´ ECIS

                                Suppose each individual has L independent lines of progression. We
                              start by calculating the rate of transition into the final, cancerous stage
                              for each independent line—the incidence per line. The incidence per
                              individual is the rate at which one of the L lines moves into the final,
                              cancerous state. The incidence per individual is simply L multiplied by
                              the incidence per line: the cancer rate rises linearly with the number of
                              independent lines that can fail.
                                If we fix the rate of progression per line, then the number of inde-
                              pendent lines does not affect log-log acceleration. However, if we wish
                              to keep constant the overall probability per individual of developing
                              cancer by a certain age, then as the number of lines increases, the prob-
                              ability of cancer per line must decline. Interestingly, slower per-line
                              transformation keeps acceleration higher through later ages, because
                              slow transformation maintains a high number of stages remaining in
                              progression.

                                                          DETAILS
                                Let the number of parallel lines of evolution within each individual
                              be L. We now have to consider progression hierarchically. Within each
                              individual, cancer arises as soon as one of the L lines progresses to the
                              nth stage. For each independent line, the probability of progressing to
   110   111   112   113   114   115   116   117   118   119   120