Study on statistical theory of carsinogenesis model and its application to data in radiation epidemiology

这项研究的目的是建立一种统计方法，以阐明辐射诱发癌素的发生机制。我们利用广岛大学辐射生物学和医学研究所对原子弹幸存者的死亡率进行跟踪调查，并从辐射流行病学和辐射生物学的角度对结果进行了调查。2000年，作为遗传不稳定性模型的一种，我们在Moolgavkar博士提出的双突变克隆扩展模型的基础上，提出了辐射暴露的持久效应模型。在延长模型中，考虑了暴露后一定时间内干细胞的变化。我们拜访了Moolgavkar博士(FHCRC，美国)，他回顾了我们的研究。肺癌、胃癌和结肠癌的研究结果在国际癌症发生数学模型研讨会(京都)上发表。2001年，我们重新考虑了致癌模型的零星效应的数学规范，并再次与Moolgavkar博士讨论了这个问题。2002年，我们分析了辐射的剂量响应曲线及其可能的阈值。我们发现男性肺癌呈曲棍球杆型曲线，女性肺癌在低剂量范围内风险相对较高。当我们结合这两种不同的曲线时，肺癌出现了线性的剂量效应。我们在第16届IEA世界流行病学大会(加拿大蒙特利尔)上发表了流行病学发现，并在该杂志上发表了一篇论文。

The purpose of this study is to develop a statistical methodology for clarifying the mechanism of carsinogenesis induced by radiation. We used the mortality among atomic bomb survivors followed-up by Research Institute for Radiation Biology and Medicine, Hiroshima University and investigated the results from the radiation-epidemiological and radiation'biological point of view. In 2000, as one of the genetic instability model we proposed a protract effect model for radiation exposure based on the two-mutation clonal expansion model developed by Dr. Moolgavkar. In the protract model, changes of stem cell during a certain period after the exposure was considered. We visited Dr. Moolgavkar (FHCRC, USA) and he reviewed our study. Results for cancers of lung, stomach and colon were presented in the International Workshop on Mathematical Modeling in Carcinogenesis (Kyoto). In 2001, we reconsidered mathematical specification for the sporadic effect of the carcinogenesis model and discussed this problem with Dr. Moolgavkar again. In 2002, we analyzed the dose response curve for radiation and its possible threshold. We found a hockey-stick type curve in male lung cancer and a relatively high risk in low dose range in female lung cancer. When we combined these two heterogeneous curves, a linear dose response were appeared in lung cancer. We presented epidemiological findings in The XVI IEA World Congress ofEpidemiology (Montreal, Canada) and published a paper in the journal.