Abstract Breast screening consists on systematic and organized exposure of asymptomatic women in mammography for the early detection of any disease in the first stages. However, breast is exposed to ionizing radiation from the mammographic unit, delivering a risk to the population under screening which must be estimated and controlled.Monte Carlo methods have been used in radiation transport for estimating dosimetric quantities, such as the absorbed dose, which is related with the radiation risk. The mamographic unit has been modelled with the Monte Carlo code MCNP5 for estimating mean glandular dose to the breast through physical measurements of the entrance surface air kerma (ESAK) through quality control testing. Several dosimetric scores or tallies have been used, such as the F2, F4 and F5, applying variance reduction techniques. The radiological risk in the Valencian Breast Screening Program has been estimated from glandular breast doses through a multiplicative model derived from Markov processes, considering several case-control cohorts: Life Span Study, fluoroscopy studies in Canada and Massachusetts and the treatment of breast benign diseases in Sweden. The study has been applied to several digital mammography units (CR and DR) acquired by the Valencian Breast Screening Program, for the quality evaluation of the technologies, regarding to mean glandular breast dose. The average radiological detriment for breast cancer incidence has been lower than 9 10-5 women-year and lower than 6 10-5 women-year for fatal cancers, whereas for digital mammography the incidence has been lesser than 1.3 10-4 and the mortality lesser than 8 10-5. Radiological risks derived from population samples from digital equipments are higher because exposure conditions (compression breast thickness and age) are differents. However, mean glandular breast doses derived from the exposure of the mammographic phantom are similar in analogic, CR and DR mammography units.