临床和政策讨论需要有关数字乳腺断层合成 (DBT) 筛查（伴或不伴辅助乳腺磁共振成像 (MRI)）的长期益处和危害的信息，特别是对于乳腺致密的患者。基于乳腺癌乳房 X 光检查策略（DBT 或数字乳房 X 光检查）的结果，有或没有根据乳腺密度进行补充 MRI。使用美国乳腺癌监测联盟数据提供的 3 个癌症干预和监测建模网络 (CISNET) 乳腺癌模拟模型进行协作建模。模拟的 1980 年出生的具有平均乳腺癌风险的女性也被纳入其中。建模分析于 2020 年 1 月至 2023 年 12 月进行。从 40、45 或 50 岁到 74 岁，每年或每两年进行一次乳房 X 光检查，根据乳腺密度进行或不进行补充 MRI。避免终生乳腺癌死亡、假阳性回忆和每 1000 名从 40 岁到死亡随访的模拟女性的假阳性活检建议总结为跨模型的平均值和范围。从 50 岁开始对所有模拟女性进行两年一次的 DBT 筛查与 40 岁相比，分别避免了 7.4 例和 8.5 例乳腺癌死亡，分别导致 884 例和 1392 例假阳性召回以及 151 例和 221 例假阳性活检建议。与 DBT 筛查相比，每两年一次的数字乳房 X 光检查可以避免类似的死亡，并且假阳性检测结果略多。在对 50 至 74 岁女性进行每两年一次的 DBT 筛查的基础上，将针对乳房极其致密的女性的 MRI 纳入其中，可以减少死亡（7.6 比 7.4）、假阳性召回（919 比 884）和假阳性活检建议（180 比 151）。将补充 MRI 扩展到乳房不均匀或极其致密的女性，进一步增加了避免的死亡（8.0 vs 7.4）、假阳性召回（1088 vs 884）和假阳性活检建议（343 vs 151）。针对 40 至 74 岁女性的相同策略避免了 9.5 例死亡，但导致 1850 例假阳性召回和 628 例假阳性活检建议。年度筛查略微增加了可避免的估计死亡人数，但显着增加了估计的假阳性结果。 在这个基于模型的比较有效性分析中，在 DBT 筛查中添加针对致密乳房女性的补充 MRI 会带来更大的益处，但也增加了危害。当 MRI 针对约占人口 10% 的乳房极其致密的女性时，这种与补充 MRI 的权衡更有利。

Information on long-term benefits and harms of screening with digital breast tomosynthesis (DBT) with or without supplemental breast magnetic resonance imaging (MRI) is needed for clinical and policy discussions, particularly for patients with dense breasts.To project long-term population-based outcomes for breast cancer mammography screening strategies (DBT or digital mammography) with or without supplemental MRI by breast density.Collaborative modeling using 3 Cancer Intervention and Surveillance Modeling Network (CISNET) breast cancer simulation models informed by US Breast Cancer Surveillance Consortium data. Simulated women born in 1980 with average breast cancer risk were included. Modeling analyses were conducted from January 2020 to December 2023.Annual or biennial mammography screening with or without supplemental MRI by breast density starting at ages 40, 45, or 50 years through age 74 years.Lifetime breast cancer deaths averted, false-positive recall and false-positive biopsy recommendations per 1000 simulated women followed-up from age 40 years to death summarized as means and ranges across models.Biennial DBT screening for all simulated women started at age 50 vs 40 years averted 7.4 vs 8.5 breast cancer deaths, respectively, and led to 884 vs 1392 false-positive recalls and 151 vs 221 false-positive biopsy recommendations, respectively. Biennial digital mammography had similar deaths averted and slightly more false-positive test results than DBT screening. Adding MRI for women with extremely dense breasts to biennial DBT screening for women aged 50 to 74 years increased deaths averted (7.6 vs 7.4), false-positive recalls (919 vs 884), and false-positive biopsy recommendations (180 vs 151). Extending supplemental MRI to women with heterogeneously or extremely dense breasts further increased deaths averted (8.0 vs 7.4), false-positive recalls (1088 vs 884), and false-positive biopsy recommendations (343 vs 151). The same strategy for women aged 40 to 74 years averted 9.5 deaths but led to 1850 false-positive recalls and 628 false-positive biopsy recommendations. Annual screening modestly increased estimated deaths averted but markedly increased estimated false-positive results.In this model-based comparative effectiveness analysis, supplemental MRI for women with dense breasts added to DBT screening led to greater benefits and increased harms. The balance of this trade-off for supplemental MRI use was more favorable when MRI was targeted to women with extremely dense breasts who comprise approximately 10% of the population.