Abstract ID: 49 Monte Carlo simulation of breast screening programmes. (October 2017)

• Record Type: Journal Article
• Authors: Forastero, C.; Zamora, L.I.; Guirado, D.; Lallena, A.M.
• Is Part Of: Physica medica. Volume 42(2017)Supplement 1
• Page Start: 10
• Publication Date: 2017-10
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2017.09.025
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Nowadays, there are still controversies about the justification of breast screening programs (BSPs). In this work we have developed a Monte Carlo tool that reproduces the outcomes of BSPs (annual detection rate, invasive/ in situ rate, interval cancer rate, etc.) and that, once tuned, permits evaluating the mortality reduction rate and the overdiagnosis for different BSP configurations (age ranges and mammography frequencies). In the simulations, the tumors diagnosed within BSP and those clinically detected (cancers of interval) were taken into account. This permitted estimating the annual reduction in breast cancer mortality due to mammographic screening for any configuration[1, 2] . The comparison of the results obtained with those found for the same population without screening provided the overdiagnosis due to BSPs. The tool was adapted to simulate randomized controlled trials (RCTs) and estimate the corresponding relative risks[3] . By considering the same population characteristics for different RCTs, their external validity was addressed; their internal validity was analyzed by investigating the disagreements observed between the relative risks obtained and those quoted for the RCTs considered. One of the main results was that assuming a quicker growth of the invasive cancers was not necessary to reproduce empirical results. Mortality reductions of 12%–20% (between two and four deaths per year and 100000 women) were obtained for acceptances above 50%. This should be a threshold for the acceptance, which appeared to be a critical parameter. Our tool allowed reproducing the known results of overdiagnosis. This varied between 10 and 20%, depending on the configuration. It was found that, after the end of the BSP, the number of invasive cancers detected was similar in control and screening groups and thus overdiagnosis is almost exclusively associated with in situ tumors. Monte Carlo simulations appear to be a very powerful tool to investigate BSPs and RCTs, helping to establish which of their results may be extrapolated to other populations, to design the trial strategies and, eventually, to adapt them during their development.