6.2 Dose-Response & Radiation Effects

Dose-Response Relationships

A dose-response relationship plots a biologic effect against absorbed dose. Two questions define every curve: does it pass through the origin (a non-threshold curve, where any dose carries some risk) or does it start at a threshold dose below which no effect occurs, and is the relationship a straight linear line or a curved linear-quadratic/sigmoid shape. Radiation biology sorts nearly all diagnostic-imaging effects into two great categories built on these curves: stochastic effects and deterministic (tissue-reaction) effects. Distinguishing them is one of the highest-yield concepts in the Safety domain.

Stochastic Versus Deterministic Effects

Stochastic effects are all-or-nothing, probabilistic outcomes. As dose increases, the probability (chance) of the effect increases, but the severity does not depend on dose, and there is no threshold: even the smallest dose is assumed to carry some risk. The two stochastic effects of concern are cancer (radiation-induced carcinogenesis) and heritable/genetic effects (mutations in germ cells). A leukemia caused by 50 mGy is no milder than one caused by 500 mGy; the higher dose simply makes it more likely.

Deterministic effects, renamed tissue reactions by the ICRP, behave oppositely. They have a threshold dose that must be exceeded before the effect appears, and once past that threshold the severity increases with dose. They arise from killing large numbers of cells. Classic examples and their approximate thresholds are shown below.

These thresholds far exceed any properly performed diagnostic exam, so in radiography the practical concern is the stochastic risk of cancer, which is why we optimize every exposure.

The Linear-Non-Threshold (LNT) Model

Because stochastic effects are assumed to have no safe threshold, radiation protection is built on the linear-non-threshold (LNT) model. LNT extrapolates the well-documented cancer risk seen at high doses (atomic-bomb survivors, radiotherapy cohorts) down through the low-dose region as a straight line to zero dose. It deliberately assumes that risk is directly proportional to dose all the way down and that no dose is completely safe. LNT is conservative and probably overestimates low-dose risk, but its caution is exactly why it underpins ALARA (As Low As Reasonably Achievable) and every regulatory dose limit. When an exam item asks which model "assumes any dose carries some risk with no safe threshold," the answer is the linear-non-threshold model.

Somatic Versus Genetic Effects

Effects are also grouped by who is harmed. Somatic effects occur in the exposed individual's own body (soma = body cells); they include radiation-induced cancer, cataracts, erythema, and life-span shortening, and are subdivided into early (appearing within hours to weeks, such as ARS and erythema) and late (appearing years later, such as cancer and cataracts). Genetic (heritable) effects occur not in the exposed person but in that person's descendants, because the damage struck a germ cell (ovum or sperm) whose mutated DNA is passed to offspring. Genetic effects are stochastic and have never been conclusively demonstrated in irradiated human populations, but they are assumed to exist, which is one reason the gonads carry a protection emphasis.

Acute Radiation Syndrome (ARS)

Acute radiation syndrome results from a high, acute, whole-body dose (generally above ~1 Gy) delivered in a short time, a scenario seen in accidents rather than diagnostic imaging, but frequently tested. ARS unfolds in four stages: a prodromal stage (nausea, vomiting, fatigue within hours), a latent stage (a symptom-free interval), a manifest illness stage (the full syndrome), and then recovery or death. Three dose-dependent sub-syndromes appear in order of increasing dose: the hematopoietic syndrome (~1-6 Gy, bone-marrow depression), the gastrointestinal syndrome (~6-10 Gy, loss of GI epithelium), and the cerebrovascular/central nervous system syndrome (above ~50 Gy, rapidly fatal). A key benchmark is the LD 50/60: the whole-body dose that kills 50% of an exposed population within 60 days without medical support, roughly 3-4 Gy in humans.

Early Versus Late Somatic Effects and the Evidence Base

Because the exam often blends timing with mechanism, keep two axes straight. Early (acute) somatic effects appear within hours to weeks, require relatively high doses, and are deterministic, including the prodromal nausea of ARS, skin erythema, epilation, and depression of blood counts. Late somatic effects appear years to decades later and include the stochastic outcomes of cancer and heritable-line mutation as well as some deterministic outcomes such as cataracts and fibrosis. The human risk estimates that anchor these models come chiefly from the atomic-bomb survivor Life Span Study, supplemented by medically and occupationally irradiated cohorts; these populations show a measurable rise in leukemia and solid tumors at moderate-to-high doses, from which the LNT line is extrapolated downward.

Worked Example: Reading a Dose-Response Curve

Suppose an item shows a curve that passes through the origin and rises as a straight line. That is a linear, non-threshold curve, the hallmark of a stochastic effect such as cancer, and the shape adopted for radiation protection. A curve that stays flat until a threshold dose and then climbs steeply is a sigmoid, threshold curve, the signature of a deterministic tissue reaction. Matching curve shape to effect category is a reliable way to answer these questions even when the specific effect is unfamiliar.