Screen-Round–Based Risk Strategies for Population-Based Mammography Screening

Background: The widespread use of organized screening strategies of mammography screening at population level stood a chance of reducing the threat to women's life from breast cancer. However, such a population-wide strategy is often faced with the questions like “How many rounds of mammography screening are require before detecting cancer in question?” and “Can the attendee be classified the low risk after several negative screening rounds?” In addition to the concerns on resource allocation for the purpose of planning an efficient population-based screening program, the recent emergence in precision medicine also makes it attractive for considering such a risk-based decision on breast cancer prevention under the context of mass screening. Aim: To quantify number of screen rounds required for detecting an asymptomatic cancer in question and dispensing with further invitation to screen. Methods: By applying a series of Bayesian negative–binomial family–based stochastic process models taking sensitivity and specificity into account, we elucidated the aforementioned issues based on the empirical data on population-based breast cancer screening program in Finland with international collaboration. The Finnish nationwide biennial mammographic screening program was implemented and targeted to women aged 50-59 years since 1988. The panel data on the regular invitation of eligible population by Pirkanmaa screening center excluding the women who had been diagnosed by breast cancer before their first invitation were enrolled in this study. Results: Based on the estimated results, we are able to determine the rounds of screens required before detecting an asymptomatic breast cancer according to the risk profile determined by age and the performance of screening tool. Based on the empirical data, an average of 2.77 (95% CI, 2.61-2.91) screen rounds will be required to detect an asymptomatic breast cancer cases. A woman may not be invited after a series of negative findings of 8 rounds of screen. Considering the sensitivity of 83% (95% CI, 61%–95%), the required screen rounds become 2.81 (95% CI, 2.65-2.94). The screening rounds required for the young (<55 years) and the old (≥55 years) age group, the corresponding figures was 2.81 (95% CI, 2.55-3.11) and 2.76 (95% CI, 2.43-3.05). Conclusion: We quantified the screen round, 2.77 on average, required to detect an asymptomatic breast cancer and 8 rounds of screen to dispense with further invitation based on the risk and the performance of screening tool. The findings may aid in risk-based interscreening interval determination but also provide information on resources required by different screening policies with target population with different risk levels.