Colorectal cancer incidence trends in the United States and United kingdom: evidence of right- to left-sided biological gradients with implications for screening

Abstract

Several lines of evidence support the premise that screening colonoscopy reduces colorectal cancer (CRC) incidence, but there may be differential benefits for right- and left-sided tumors. To better understand the biological basis of this differential effect, we derived biomathematical models of CRC incidence trends in U.S. and U.K. populations, representing relatively high- and low-prevalence screening, respectively. Using the Surveillance Epidemiology and End Results (SEER) and the Office for National Statistics (ONS) registries (both 1973-2006), we derived stochastic multistage clonal expansion (MSCE) models for right-sided (proximal colon) and left-sided (distal colon and rectal) tumors. The MSCE concept is based on the initiation-promotion-progression paradigm of carcinogenesis and provides a quantitative description of natural tumor development from the initiation of an adenoma (via biallelic tumor suppressor gene inactivation) to the clinical detection of CRC. From 1,228,036 (SEER: 340,582; ONS: 887,454) cases, parameter estimates for models adjusted for calendar-year and birth-cohort effects showed that adenoma initiation rates were higher for right-sided tumors, whereas, paradoxically, adenoma growth rates were higher for left-sided tumors. The net effect was a higher cancer risk in the right colon only after age 70 years. Consistent with this finding, simulations of adenoma development predicted that the relative prevalence for right- versus left-sided tumors increases with increasing age, a differential effect most striking in women. Using a realistic biomathematical description of CRC development for two nationally representative registries, we show age- and sex-dependent biological gradients for right- and left-sided colorectal tumors. These findings argue for an age-, sex-, and site-directed approach to CRC screening.

Figure 1

A. Three-stage CRC carcinogenesis model. The model agrees with the notion that adenoma initiation requires the biallelic inactivation of the APC gene. The parameters are: the number of susceptible stem cells, X, the mutation rate of the first hit at the APC gene, μ0, the mutation rate of the second copy of the APC gene, μ1, the adenoma cell division rate, α, the adenoma cell death or differentiation rate, β, and the malignant transformation rate, μ2. B. The hazard (or incidence) function of the model (as a function of age t) exhibits three distinct phases that reveal the carcinogenic process in reverse. In particular, for older ages the hazard increases linearly reflecting the incidence of adenomas. The line approximating this (linear) phase has an intersect with the age axis that represents the mean time between adenoma initiation and its conversion into a clinical carcinoma (Ts). For mid-ages, the hazard rises exponentially, with a rate equal to the net growth rate of adenomas (α-β). For younger ages, the hazard increases as a power of age, consistent with the Armitage-Doll theory of multistage carcinogenesis

Figure 2

Left. US and UK CRC incidence standardized to the US 2000 population by sex and subsite. Right. Estimated calendar-year effects.

Figure 3

Three-stage CRC model hazard by country, sex and subsite. The figure shows the combined effects of differential tumor initiation rates and tumor growth rates on the estimated hazard functions that represent the intrinsic age-effects on CRC risk. In all panels, left-sided lesions dominate during early ages and are more prevalent, whereby in older ages, there is a right-sided dominance.

Figure 4

Predicted adenoma (≥1mm) prevalence in asymptomatic individuals by country, sex and subsite. We assume that there are about 500,000 cells in a 1mm adenoma (Pinsky, JTB 2000) and that about 1% of those are stem cells (Boman, JCO 2008). The relative prevalence of right-sided versus left-sided adenomas increases with age in all panels.