## Results: 5

1.

2.

Schematic representation of different possible extrapolations of measured radiation risks down to very low doses, all of which could, in principle, be consistent with higher-dose epidemiological data. Curve a, linear extrapolation; curve b, downwardly curving (decreasing slope); curve c, upwardly curving (increasing slope); curve d, threshold; curve e, hormetic.

3.

Schematic representation of the potential effect of a small (0.25%) population of women, who are extremely sensitive for radiation-induced breast cancer, compared with the general (normal) population. Schematized is the number of radiation-induced breast cancers as a percentage of the overall population. The dose–risk relations for both the normal and the sensitive populations are assumed to be linear. Because the number of radiation-induced breast cancers in the sensitive population would saturate as the dose increases (because all the exposed women would have developed breast cancer), the dose–response for the whole population would be downwardly curving.

4.

Estimated risks (relative to an unexposed individual) of solid cancer in atomic bomb survivors exposed to low radiation doses (12). Data points are placed at the mean of each dose category. The solid curve represents a weighted moving average of the points shown (dotted curves: ±1 SE), and the dashed straight line is a linear risk estimate computed from all the data in the dose range from 0 to 2,000 mSv. Age-specific cancer rates from 1958 to 1994 are used, averaged over follow-up and gender.

5.

Estimated excess relative risk (±1 SE) of mortality (1950–1997) from solid cancers among groups of survivors in the LSS cohort of atomic bomb survivors, who were exposed to low doses (<500 mSv) of radiation (2). The groups correspond to progressively larger maximum doses, with the mean doses in each group indicated above each data point. The first two data points (in blue) are not statistically significant (

*P*= 0.15 and 0.3, respectively) compared with the comparison population who were exposed to <5 mSv, whereas the remaining four higher-dose points (in red) are statistically significant (*P*< 0.05). The dashed straight line represents the results of a linear fit (2) to all the data from 5 to 4,000 mSv (higher dose points are not shown).