In a cell survival experiment, cells are seeded onto petri dishes and exposed to various doses of radiation. The number of cell colonies that subsequently grow is determined. Each colony is assumed to be derived from a single surviving cell. A cell survival curve is a plot of the fraction of cells that survive (normalized by the fraction of cells that survive with no radiation exposure) versus the radiation dose; each point on a cell survival curve corresponds to a single dose of radiation. Usually, the log of the surviving fraction is plotted on the vertical axis versus dose on the horizontal axis. Mammalian cell survival curves generally show an exponential response to high doses of radiation with a “shoulder” of varying widths in the low-dose range. The effect of radiation on cells can be observed by noting how the cell survival curve changes under different conditions. These changes in condition include using different types of radiation, changing the environment, adding or subtracting oxygen, or delivering the radiation at different times or during different parts of the cell cycle. The study of survival curves revealed many aspects of radiation damage long before the structure of DNA was elucidated.

The single-target/single-hit model has little practical application, but is useful in explaining the multitarget and LQ models. In the single-target/single-hit model, it is assumed that a cell has a single target that when hit causes the cell to die. In this case, the cell has no opportunity to repair the radiation damage. The single-target/single-hit model is inadequate to explain most cell survival data from mammalian cells, because it does not account for the shoulder portion of the curve at low doses.

The multitarget (two-component) and LQ models are both considered multiple-target models. Both models assume that each cell contains two or more targets that must be hit before the cell is killed. In order to be killed, the cell must accumulate enough hits in a short amount of time, such that the enzyme repair mechanisms are not capable of repairing all of the damage in between hits. However, after the first target is hit, the cell may have enough time to repair the damage before the next target is hit. In this case, the first hit is an example of sublethal damage. Sublethal damage occurs at a dose that is not sufficient to cause very much cell death. Thus, there are two dose regimes (low dose causing mostly sublethal damage, and high dose, causing lethal damage) that explain the shoulder (i.e., the change in slope) of the survival curve. Importantly, the multiple-target models provide a way to account for the presence of the shoulder region observed in mammalian cell survival curves. The multiple-target models fit the experimental data well, but still have limitations. Regardless of the underlying mechanisms, the major interest in survival curves data is in predicting radiation effects on humans.