Abstract Introduction Carbon-11-labelled positron emission tomography (PET) tracers commonly used in biomedical research expose subjects to ionising radiation. Dosimetry is the measurement of radiation dose, but also commonly refers to the estimation of health risk associated with ionising radiation. This review describes radiation dosimetry of carbon-11-labelled molecules in the context of current PET research and the most widely used regulatory guidelines. Methods A MEDLINE literature search returned 42 articles; 32 of these were based on human PET data dealing with radiation dosimetry of carbon-11 molecules. Radiation burden expressed as effective dose and maximum absorbed organ dose was compared between tracers. Results All but one of the carbon-11-labelled PET tracers have an effective dose under 9 μSv/MBq, with a mean of 5.9 μSv/MBq. Data show that serial PET scans in a single subject are feasible for the majority of radiotracers. Conclusion Although differing in approach, the two most widely used regulatory frameworks (those in the USA and the EU) do not differ substantially with regard to the maximum allowable injected activity per PET study. The predictive validity of animal dosimetry models is critically discussed in relation to human dosimetry. Finally, empirical PET data are related to human dose estimates based on homogenous distribution, generic models and maximum cumulated activities. Despite the contribution of these models to general risk estimation, human dosimetry studies are recommended where continued use of a new PET tracer is foreseen.