Response to Omoleye, Esserman, Olufunmilayo

Our study evaluated screening outcomes of breast magnetic resonance imaging (MRI) with or without mammography by breast density in the Breast Cancer Surveillance Consortium (BCSC) cohort (1). We found that MRI screening with or without mammography increased rates of screen-detected early-stage cancer and false-positive recalls for women with dense breasts without a concomitant decrease in advanced or interval cancers, suggesting that breast density alone should not be the sole criterion by which to recommend supplemental breast MRI. We agree with Omoleye et al. (2) that breast cancer screening and supplemental imaging should be based on a woman’s breast cancer risk to maximize screening benefits and minimize screening harms. When comparing results by BCSC 5-year invasive cancer risk (Table 1), as suggested by Omoleye et al. (2), we found that the rate of screen-detected early-stage cancer was higher in the MRI group than in the mammography alone group for women at high risk vs low risk, with the MRI vs mammography difference not significantly higher in the high-risk vs low-risk group. Increases in false-positive recalls with MRI vs mammography were borderline significantly lower in the high-risk vs low-risk group. There were no significant differences in advanced or interval cancer rates within BCSC 5-year risk groups.

Table 1.

Rate per 1000 screening exams of screening benefits, failures, harms, and other outcomes for MRI vs mammography alone, by BCSC 5-year invasive breast cancer risk

	Low BCSC 5-year risk <2.5%a			High BCSC 5-year risk ≥2.5%b
	MRI, rate per 1000 screening exams (95% CI)	Mammography, rate per 1000 screening exams (95% CI)	Difference (95% CI)c	MRI, rate per 1000 screening exams (95% CI)	Mammography, rate per 1000 screening exams (95% CI)	Difference (95% CI)c	Difference of differences (95% CI)	P
Screening benefits
Early-stage invasive cancer detection	9.8 (7.3 to 12.3)	2.7 (2.2 to 3.1)	7.1 (4.6 to 9.7)	14.1 (9.9 to 18.3)	4.2 (3.3 to 5.1)	9.9 (5.6 to 14.1)	‒2.8 (‒7.7 to 2.2)	.28
Screening failures
Interval invasive cancer	1.1 (0.3 to 2.0)	0.7 (0.4 to 0.9)	0.4 (‒0.4 to 1.3)	1.3 (0.0 to 2.6)	1.8 (1.2 to 2.4)	‒0.5 (‒1.9 to 0.9)	0.9 (‒0.7 to 2.6)	.27
Advanced cancer	0.5 (0 to 1.0)	0.5 (0.3 to 0.7)	0.0 (‒0.6 to 0.6)	1.6 (0.2 to 3.0)	0.9 (0.5 to 1.3)	0.7 (‒0.7 to 2.2)	‒0.7 (‒2.3 to 0.8)	.36
Screening harmsd
False-positive recall	249.2 (237.5 to 261.0)	104.9 (101.8 to 107.9)	144.4 (132.2 to 156.5)	212.3 (197.0 to 227.6)	87.5 (83.6 to 91.4)	124.8 (108.9 to 140.6)	19.6 (‒0.1 to 39.3)	.05
False-positive biopsy recommendation	103.4 (95.7 to 111.2)	16.5 (15.3 to 17.8)	86.9 (79.1 to 94.7)	105.2 (94.2 to 116.2)	18.0 (16.1 to 19.9)	87.2 (76.1 to 98.4)	‒0.3 (‒13.9 to 13.2)	.96
Other outcomes
Ductal carcinoma in situ detection rate	6.5 (4.4 to 8.5)	1.5 (1.2 to 1.9)	4.9 (2.9 to 7.0)	6.6 (3.7 to 9.4)	2.8 (2.1 to 3.6)	3.7 (0.8 to 6.7)	1.2 (‒2.4 to 4.8)	.52

a

N = 49 711 examinations. CI = confidence interval; MRI = magnetic resonance imaging; BCSC = Breast Cancer Surveillance Consortium.

b

N = 23 976 examinations.

c

Values in bold are significant differences between MRI and mammography.

c

Not ductal carcinoma in situ or invasive cancer.

Two studies have found a statistically significant reduction in advanced breast cancer incidence among women undergoing MRI (3,4), and both were in women at high risk for breast cancer. The average 5-year BCSC risk among high-risk women (one-third of the study population) was 3.7%, which may limit our statistical power for detecting a reduction in advanced cancer incidence; also, 3.7% risk may not be sufficiently high to observe a reduction in advance cancer among women undergoing MRI.

Risk assessment conducted in the BCSC has gone beyond predicting invasive breast cancer risk to predicting advanced breast cancer risk, defined as prognostic pathologic stage II or higher (3). Prognostic pathologic stage predicts breast cancer death by combining markers associated with tumor aggressiveness (estrogen receptor, progesterone receptor, HER2 status, tumor grade) with anatomic staging elements (tumor size, lymph node status, presence of metastatic disease). Given the that goal of screening is to reduce breast cancer mortality and that advanced cancer is a surrogate for breast cancer mortality (5), advanced cancer risk is a clinically relevant outcome when discussing supplemental imaging with women. The BCSC created an advanced cancer risk model that includes age, race and ethnicity, breast density, first-degree family history of breast cancer, history of benign breast biopsy, body mass index, and menopausal status; this model can be used to determine a screening interval and decide whether supplemental imaging should be considered for those at high risk for advanced cancer (6). Age 60 years and older, Black race, dense breasts, obesity (body mass index >30), and proliferative breast disease with atypia are associated with increased advanced cancer risk (6). In addition to clinical risk factors, an artificial intelligence–based detection system (Transpara) has been shown to be associated with advanced cancer risk and could improve accuracy of risk prediction (7).

It is important to examine level of breast cancer risk in association with breast cancer mortality, an area of ongoing research within the BCSC.

Data availability

Data will be available after study aims of funded grants have been addressed and following approval by the BCSC Steering Committee ([email protected]).

Author contributions

Karla Kerlikowske, MD (Conceptualization; Data curation; Funding acquisition; Investigation; Methodology; Writing—original draft; Writing—review & editing), Weiwei Zhu, MS (Formal analysis; Methodology; Writing—review & editing), Yu-Ru Su, PhD (Data curation; Formal analysis; Methodology; Writing—review & editing), Brian Sprague, PhD (Data curation; Funding acquisition; Methodology; Writing—review & editing), Ellen O’Meara, PhD (Data curation; Project administration; Writing—review & editing), Anna Tosteson, ScD (Conceptualization; Data curation; Funding acquisition; Methodology; Writing—review & editing), Karen Wernli, PhD (Data curation; Funding acquisition; Methodology; Writing—review & editing), Diana Miglioretti, PhD (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Writing—review & editing).

Funding

Research reported in this work was funded through a Patient-Centered Outcomes Research Institute (PCORI) award (PCS-1504-30370). The BCSC additionally supported data collection for this research with funding from the National Cancer Institute (P01CA154292, U54CA163303), the Agency for Health Research and Quality (R01 HS018366-01A1), and the University of Vermont Cancer Center, with funds generously awarded by the Lake Champlain Cancer Research Organization (grant No. 032800). Cancer and vital status data collection was supported by several state public health departments and cancer registries (http://www.bcsc-research.org/work/acknowledgement.html).

Conflicts of interest

The authors report no potential conflicts of interest.

Acknowledgements

We thank the participating women, mammography facilities, and radiologists for the data they have provided for this study.

The views in this work are solely the responsibility of the authors and do not necessarily represent the views of the National Cancer Institute or Patient-Centered Outcomes Research Institute, its Board of Governors, or its Methodology Committee. Neither the National Cancer Institute nor the Patient-Centered Outcome Research Institute had a role in the design or conduct of the study or the reporting of results.