Radium applicators and pure beta emitters have been widely used in the past to treat skin haemangioma in early childhood. A well defined relationship between the low doses received from these applicators and radiation-induced cancers requires accurate dosimetry. A human-based CT scan phantom has been used to simulate every patient and treatment condition and then to calculate the source target distance when radium and pure beta applicators were used. The effective transmission factor psi(r) for the gamma spectrum emitted by the radium sources applied on the skin surface was modelled using Monte Carlo simulations. The well-known quantization approach was used to calculate gamma doses delivered from radium applicators to various anatomical points. For 32P, 90Sr/90Y applicators and 90Y needles we have used the apparent exponential attenuation equation. The dose calculation algorithm was integrated into the ICTA software (standing for a model that constructs an Individualized phantom based on CT slices and Auxological data), which has been developed for epidemiological studies of cohorts of patients who received radium and beta-treatments for skin haemangioma. The psi(r) values obtained for radium skin applicators are in good agreement with the available values in the first 10 cm but higher at greater distances. Gamma doses can be calculated with this algorithm at 165 anatomical points throughout the body of patients treated with radium applicators. Lung heterogeneity and air crossed by the gamma rays are considered. Comparison of absorbed doses in water from a 10 mg equivalent radium source simulated by ICTA with those measured at the Radiumhemmet, Karolinska Hospital (RAH) showed good agreement, but ICTA estimation of organ doses did not always correspond those estimated at the RAH. Beta doses from 32P, 90Sr/90Y applicators and 90Y needles are calculated up to the maximum beta range (11 mm).