TY - JOUR
T1 - Breast Cancer Screening Using Mammography, Digital Breast Tomosynthesis, and Magnetic Resonance Imaging by Breast Density
AU - Stout, Natasha K.
AU - Miglioretti, Diana L.
AU - Su, Yu Ru
AU - Lee, Christoph I.
AU - Abraham, Linn
AU - Alagoz, Oguzhan
AU - de Koning, Harry J.
AU - Hampton, John M.
AU - Henderson, Louise
AU - Lowry, Kathryn P.
AU - Mandelblatt, Jeanne S.
AU - Onega, Tracy
AU - Schechter, Clyde B.
AU - Sprague, Brian L.
AU - Stein, Sarah
AU - Trentham-Dietz, Amy
AU - van Ravesteyn, Nicolien T.
AU - Wernli, Karen J.
AU - Kerlikowske, Karla
AU - Tosteson, Anna N.A.
N1 - Publisher Copyright:
© 2024 American Medical Association. All rights reserved.
PY - 2024/8/26
Y1 - 2024/8/26
N2 - Importance: 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. Objective: 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. Design, Setting, and Participants: 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. Intervention: 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. Main outcomes and Measures: 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. Results: 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. Conclusions and relevance: 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.
AB - Importance: 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. Objective: 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. Design, Setting, and Participants: 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. Intervention: 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. Main outcomes and Measures: 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. Results: 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. Conclusions and relevance: 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.
UR - http://www.scopus.com/inward/record.url?scp=85206017698&partnerID=8YFLogxK
U2 - 10.1001/jamainternmed.2024.4224
DO - 10.1001/jamainternmed.2024.4224
M3 - Comment/Letter to the editor
C2 - 39186304
AN - SCOPUS:85206017698
SN - 2168-6106
VL - 184
SP - 1222
EP - 1231
JO - JAMA Internal Medicine
JF - JAMA Internal Medicine
IS - 10
ER -