Colditz Graham A, Rosner Bernard A
Cancer Epidemiology Program, Dana-Farber/Harvard Cancer Center, Boston, MA, USA.
Breast Cancer Res. 2006;8(3):208. doi: 10.1186/bcr1414. Epub 2006 Jun 6.
Models of breast cancer incidence have evolved from the observation by Armitage and Doll in the 1950s that the pattern of incidence by age differs for reproductive cancers from those of other major malignancies. Both two-stage and multistage models have been applied to breast cancer incidence. Consistent across modeling approaches, risk accumulation or the rate of increase in breast cancer incidence is most rapid from menarche to first birth. Models that account for the change in risk after menopause and the temporal sequence of reproductive events summarize risk efficiently and give added insights to potentially important mechanistic features. First pregnancy has an adverse impact on progesterone receptor negative tumors, while increasing parity reduces the risk of estrogen/progesterone receptor positive tumors but not estrogen/progesterone receptor negative tumors. Integrated prediction models that incorporate prediction of carrier status for highly penetrant genes and also account for lifestyle factors, mammographic density, and endogenous hormone levels remain to be efficiently implemented. Models that both inform and reflect the emerging understanding of the molecular and cell biology of carcinogenesis are still a long way off.
乳腺癌发病率模型已从20世纪50年代阿米蒂奇和多尔的观察演变而来，即生殖系统癌症的年龄发病率模式与其他主要恶性肿瘤不同。两阶段模型和多阶段模型都已应用于乳腺癌发病率研究。各种建模方法的结果一致显示，从初潮到首次生育，乳腺癌发病率的风险积累或增长速度最快。考虑绝经后风险变化以及生殖事件时间顺序的模型能够有效总结风险，并为潜在的重要机制特征提供更多见解。首次怀孕对孕激素受体阴性肿瘤有不利影响，而增加生育次数可降低雌激素/孕激素受体阳性肿瘤的风险，但对雌激素/孕激素受体阴性肿瘤无效。整合预测模型，将高穿透性基因携带者状态的预测与生活方式因素、乳房X线密度和内源性激素水平结合起来，仍有待有效实施。能够告知并反映对致癌作用分子和细胞生物学新认识的模型，距离实现还有很长的路要走。
