RE: Supplemental magnetic resonance imaging plus mammography compared with magnetic resonance imaging or mammography by extent of breast density

The article in the Journal by Kerlikowske et al. (1) evaluates outcomes of breast cancer screening for supplemental magnetic resonance imaging (MRI) plus mammography, MRI-only, and mammography-only screening categories. The study included 52 237 women aged 40-79 years and compared rates of cancer detection, false-positive recalls, and biopsy recommendations across breast density categories, adjusting for age, race, ethnicity, family history of breast cancer, previous breast biopsy, examination year, and Breast Cancer Surveillance Consortium (BCSC) recruitment registry. Results showed that MRI, with or without mammography, increased detection rates of early stage breast cancer but did not statistically significantly affect the incidence of advanced or interval cancers. Further, MRI with or without mammography led to higher false-positive recall rates compared with mammography alone. The study suggests that although MRI can improve early cancer detection, it does not affect the incidence of advanced cancers and increases false-positive recalls.

Findings from this large community-based study underscore the unmet clinical need for a paradigm shift to more holistic risk-based surveillance because of the heterogeneous genomic architecture of invasive breast cancer in diverse settings (2-4). The study was performed in a population at intermediate risk for breast cancer, but nearly one-third of the population was deemed at high or very high risk for the disease based on BCSC cutoffs. Debates about supplemental breast cancer surveillance over the past 2 decades have emphasized conventional risk factor profiles and breast density. However, individual risk profiles vary based on genomic and nongenomic factors, which may be ethnic specific and associated with molecular subtypes of breast cancer in diverse populations.

Supplemental MRI may benefit women at the highest risk of aggressive interval cancers, but the indiscriminate use in average-risk women may do more harm than good. A randomized controlled trial showed that in patients with a cumulative lifetime breast cancer risk of 20% and above, annual MRI statistically significantly downstaged detected cancers (5). Further, a prospective screening cohort of genomically defined high-risk women showed that biannual MRI downstages aggressive breast cancers, particularly among BRCA1 mutation carriers, with 1.7 biopsies needed to detect 1 cancer using biannual MRI, while maintaining the overall recall rate below 5% (6). An interesting follow-up analysis of the results by Kerlikowske et al. (1) might evaluate screening outcomes in the study subpopulation deemed high and/or very high risk based on BCSC 5-year risk score cut-off of at least 2.5. Proof of benefit in this subgroup might further support risk-stratified supplemental MRI screening.

One way to achieve the clinical benefit and efficiency that has been obtained with age-based screening mammography is to considerably shorten screening breast MRI protocols by decreasing image acquisition time. Our ongoing Chicago Alternative Prevention Study (clinicaltrials.gov: NCT04877912) will use the most updated ethnic-specific polygenic risk score developed by the African American Breast Cancer Consortium and the WISDOM (Women Informed to Screen Depending on Measures of Risk) study, along with high-penetrance gene mutations and known environmental and lifestyle factors, to identify women in the highest risk category. Integrating artificial intelligence tools in such studies will impact the quality of care and dissemination of MRI screening in community practices.

Data availability

Not applicable.

Author contributions

Olasubomi Jimmy Omoleye, MD (Writing—original draft; Writing—review & editing), Laura Esserman, MD, MBA (Conceptualization; Writing—original draft; Writing—review & editing), and Olufunmilayo Olopade, MD (Conceptualization; Resources; Supervision; Writing—original draft; Writing—review & editing).

Funding

Not applicable.

Conflicts of interest

OJO: None. LJE: None. OIO: Grant/research support from Hoffman La Roche and Color Genomics. Stockholder in CancerIQ, Tempus, 54gene, and Healthwell Corp.