Cancer as an Evolutionary Process
Cancer is evolution in miniature. Within the body, cells compete for resources, acquire mutations that enhance their proliferative capacity, and undergo natural selection favoring unchecked growth. The Armitage-Doll multi-hit model, published in 1954, provided the first quantitative framework linking somatic mutation accumulation to the dramatic age-dependence of cancer incidence — one of the most important insights in cancer biology.
The Multi-Hit Model
The model proposes that a normal cell must acquire k specific, rate-limiting mutations to become cancerous. If each mutation occurs independently at rate μ per cell division, and the body performs N cell divisions per year, then the probability of cancer by age t follows:
P(cancer by age t) ≈ 1 − exp(−μ^k · N · t^k / k!)
The key prediction: on a log-log plot of cancer incidence versus age, the data should fall on a straight line with slope k−1. Empirical data for most adult epithelial cancers show slopes of 4-6, suggesting 5-7 rate-limiting steps — remarkably consistent with the number of driver mutations identified by modern cancer genomics.
DNA Repair: The Guardian Against Drift
The raw mutation rate per cell division is approximately 10⁻⁷ per base pair, but DNA repair mechanisms correct the vast majority of errors. The effective mutation rate is μ_eff = μ_base × (1 − repair_efficiency). With 99.9% repair efficiency, the effective rate drops a thousandfold. This is why inherited defects in DNA repair genes (BRCA1, BRCA2, mismatch repair genes in Lynch syndrome) cause such dramatic increases in cancer risk — they shift the effective mutation rate by orders of magnitude.
From Armitage-Doll to Modern Genomics
Bert Vogelstein's landmark work on colorectal cancer identified the specific sequence of mutations required: APC → KRAS → SMAD4 → TP53, each conferring a selective growth advantage. This molecular confirmation of the multi-hit model, combined with large-scale cancer genome sequencing, has validated Armitage and Doll's 1954 insight with extraordinary precision.
Tomasetti and Vogelstein (2015) extended this framework by showing that cancer risk across different tissues correlates strongly with the number of stem cell divisions — the 'bad luck' component of cancer that is a direct consequence of unavoidable replication errors, exactly as the multi-hit model predicts.
Using the Simulator
Adjust the mutation rate and DNA repair efficiency to see how they interact. Notice that the number of hits (k) controls the steepness of the age-incidence curve: k=2 gives a shallow curve (as in retinoblastoma, consistent with Knudson's two-hit hypothesis), while k=6 gives the steep power-law increase typical of adult carcinomas. The animated DNA strand below shows mutations accumulating in real time — a visual reminder that cancer is, fundamentally, a disease of accumulated DNA damage.