The committee based that judgment on the following considerations:

• Substantial new epidemiologic evidence has accumulated since the 1990 BEIR V report was published. The present committee's phase-1 report cites 39 new epidemiologic studies that fall into this category (see Table 1). Additional studies that have a direct bearing on the subject should become available in the next 3 years, the estimated period required to complete the phase-2 study.

• Some of the new epidemiologic data are on subjects on which information previously had been sparse, such as cancer mortality in those exposed to whole-body irradiation in childhood.

• Studies of carcinogenesis completed since publication of the last BEIR report have focused on mechanisms and the cellular and molecular events that are involved in the neoplastic process. The understanding of molecular events involved in carcinogenesis has increased significantly. Mechanisms that might be involved in radiation carcinogenesis have been identified. Further knowledge of these mechanisms that should become available in the next 3 years might affect estimation of the radiation-response curve at low doses.

• Over the next few years, investigators will be applying two closely linked approaches using animal models of carcinogenesis. These will likely contribute to a better understanding of mechanisms of radiation-induced cancer. In the first of these two approaches, genetically engineered mice with alterations in specific genes will be used to determine the influence of these genes on the susceptibility of the mice to radiation-induced cancer. In the second approach, studies will be conducted of the inherent differences in susceptibility to radiation-reduced cancer among different mouse strain, the objective being to identify the genes involved in controlling susceptibility. Researchers responsible for this new generation of animal studies are taking advantage of the current rapid developments in molecular genetics. Progress on both approaches should be substantial over the next few years. Significant results of relevance to risk estimation are expected to be available for the proposed BEIR VII phase-2 study.

• Evidence regarding specific biologic events that can affect the shape of the dose-response curve at low doses is accumulating. Information on such phenomena as DNA repair, signal transduction, chromosomal instability, ''bystander" effects, and adaptation, although preliminary, might eventually affect risk analyses of low-dose and low-dose-rate exposures.

The Committee Recommends That the Individuals Responsible for the Proposed Phase-2 Study

• Include a comprehensive review of all relevant epidemiologic data related to low-LET (low linear energy transfer), i.e. sparsely ionizing, radiation.