Page 83 - 20dynamics of cancer

68 CHAPTER 4

between two groups to infer underlying processes of progression. In
this case, the comparison pointed to a genetic mutation as a key rate-
limiting step.
The second contribution arose from the hypothesis that two muta-
tions provide the only rate-limiting barriers to tumor progression in ret-
inoblastoma. Knudson’s conclusion that two genetic hits lead to cancer
contributed an important step in the history of the subject. In partic-
ular, Knudson’s study presented the first data in support of the idea
that cancer is primarily a genetic disease driven by mutation and that
progression can be explained by known rates of mutation. Later, when
it was discovered that the genetic basis of retinoblastoma depended on
mutational knockout of both alleles at a single locus—named the retino-
blastoma or Rb locus—Knudson’s hypothesis provided the link between
the rate of cancer progression and the molecular nature of tumor sup-
pressor genes, in which abrogation by mutation of both alleles knocks
out the function of a tumor suppressor protein and releases a constraint
on tumorigenesis.
Knudson (1971) has been cited 2,926 times as of August, 2005. Fig-
ure 4.1 shows the citation history by year. The sharp increase in citations
in the early 1990s follows the rise of molecular studies that confirmed
the key role of tumor suppressor genes in limiting cancer progression
and the contribution of mutations to tumor suppressor genes in tumori-
genesis (Knudson 2003).
A dissonance exists between Knudson’s quantitative method of analy-
sis, which formed the entire basis for his paper, and the molecular anal-
yses of the 1990s that elevated Knudson’s work to classic status. The
later molecular work cited Knudson because he foreshadowed the con-
clusions of the molecular analyses: cancer progression requires knock-
out of both alleles of a tumor suppressor locus. But that molecular work
has ignored the major intellectual contribution of the Ashley-Knudson
papers: the quantitative analysis of progression dynamics by compari-
son of age-specific incidence curves between different genotypes.
I have emphasized several times that a gene has a causal effect on
cancer to the extent that it has a quantitative effect on progression dy-
namics: a genetic change has a causal effect to the extent that the ge-
netic change shifts the age-specific incidence curve. Ultimately, research
must return to this quantitative problem. I develop this issue in the next
chapter.