Reviews and CommentaryFree Access

Supplemental MRI in Extremely Dense Breasts: Sharp Reduction in False-Positive Rate in the Second Screening Round of the DENSE Trial

Published Online:https://doi.org/10.1148/radiol.2021210265

See also the article by Veenhuizen et al in this issue.

Dr Moy is a professor of radiology at New York University. She is the senior deputy editor for Radiology and deputy editor of Breast Imaging for Radiology. Her research focuses on diagnostic oncologic imaging, with an emphasis on breast cancer. She is an NIH-funded investigator with applications on multiparametric breast MRI and artificial intelligence. She collaborates with the NYU Center for Data Science to investigate deep learning tools for multitask learning across modalities.

Dr Moy is a professor of radiology at New York University. She is the senior deputy editor for Radiology and deputy editor of Breast Imaging for Radiology. Her research focuses on diagnostic oncologic imaging, with an emphasis on breast cancer. She is an NIH-funded investigator with applications on multiparametric breast MRI and artificial intelligence. She collaborates with the NYU Center for Data Science to investigate deep learning tools for multitask learning across modalities.

Dr Gao is an assistant professor in the breast imaging section of the Department of Radiology at the NYU Grossman School of Medicine. She is the associate fellowship program director for breast imaging at NYU. Her research interests focus on advanced imaging techniques and improving breast imaging outcomes. She serves as a consulting editor for the AJR. She is a fellow of the Society of Breast Imaging.

Dr Gao is an assistant professor in the breast imaging section of the Department of Radiology at the NYU Grossman School of Medicine. She is the associate fellowship program director for breast imaging at NYU. Her research interests focus on advanced imaging techniques and improving breast imaging outcomes. She serves as a consulting editor for the AJR. She is a fellow of the Society of Breast Imaging.

Supplemental breast MRI screening in women at high risk for breast cancer has the highest cancer detection rate (CDR) among all breast imaging modalities (1). There is growing interest in performing MRI in women with dense breast tissue considered at intermediate risk for breast cancer. In these women, screening mammography has a lower sensitivity due to the masking effect of dense breast tissue. In addition, dense breast tissue portends a higher risk of breast cancer. Compared to women with fatty breasts, women with extremely dense breasts are 17.8 times more likely to have interval cancers after a negative mammogram (2). Although digital breast tomosynthesis has improved sensitivity and specificity in mammography screening, there has been no benefit to women with extremely dense breasts regardless of age or round of screening (3). Therefore, it is of great interest that the initial results of the Dutch Dense Tissue and Early Breast Neoplasm Screening (DENSE) trial showed that supplemental MRI screening in women with extremely dense breasts not only yielded 16.5 cancers not detected at mammography per 1000 examinations but also significantly lowered interval cancer rate (4). However, despite these promising early findings, several key questions remain unanswered.

The ability to maintain the CDR and interval cancer rate over multiple rounds of screening is currently unknown. The prevalence screening round with MRI results in a surplus of cancer yield due to detection of mammographically occult cancers. Therefore, incremental CDRs in subsequent incidence rounds of screening are expected to be lower but also better reflect the actual potential benefit of ongoing screening. A clearer understanding of results across multiple rounds of screening breast MRI is therefore essential.

Similarly, the clinical impact of supplemental MRI screening in women with dense breasts without other risk factors is not well understood. There are two major questions. First, does the higher cancer detection with MRI translate into reduced mortality from breast cancer? Second, is the number of aggressive subtypes of cancer detected with MRI over mammography large enough to justify potential trade-offs, such as higher false-positive rates? These issues are important in considering whether to offer density-based supplemental MRI screening in a substantial segment of the population.

In this issue of Radiology, Veenhuizen et al (5) report their results from the second screening round of the DENSE trial, directly addressing these questions. The study included women aged 50–75 years with extremely dense breasts as quantified by Volpara imaging software and who undergo biennial screening in the Netherlands. These women were randomized to receive screening mammography plus screening MRI versus screening mammography alone following a negative mammogram (4). Those with negative results on the first round of MRI and mammographic screening without cancer diagnosis in the interim were invited for the second round of MRI screening. The authors found that the incremental CDR of MRI over mammography in the incidence round (second round) was lower than that in the prevalence round (first round) (5.8 vs 16.5 per 1000 screening examinations). The number of false-positive results was also lower (26.3 vs 79.8 per 1000 screening examinations) (5).

This study provides high-quality data from a multicenter randomized controlled trial. This approach reflects a range of clinical practice and breast MRI screening outcomes in women with extremely dense breasts, where, to our knowledge, no prior data exist. The observed lower CDR from the prevalence to incidence round is expected (16.5 vs 5.8 per 1000 screening examinations). These findings are consistent with those of Kuhl et al (6), who performed the only other MRI screening study in average-risk women (22.6 vs 6.9 per 1000 screening examinations). Despite this reduction, the incidence round CDR remains approximately 6 per 1000 screening examinations. This rate compares favorably to the incremental yield of digital breast tomosynthesis or US in women with dense breasts relative to digital mammography (2.8 and 4.9 per 1000 screening examinations, respectively), as reported in the Adjunct Screening with Tomosynthesis or Ultrasound in Women with Mammography-Negative Dense Breasts trial, or ASTOUND-2 (7). The DENSE trial has a third round of screening with MRI planned, which will be of continued interest.

The false-positive findings frequently encountered with supplemental screening modalities is a critical factor that limits widespread clinical implementation. For example, US is frequently performed as a supplemental examination in women with dense breasts but often criticized for its limited specificity. In the ASTOUND-2 trial, although supplemental US provided a higher incremental CDR than digital breast tomosynthesis as compared with digital mammography (4.9 vs 2.8 per 1000 screening examinations), it was also associated with a more than threefold higher rate of false-positive findings (7). For similar reasons, the significantly lower false-positive rate (a more than 65% reduction) in the DENSE trial during the incidence round of screening is of clinical significance (5). This finding is in line with prior prospective MRI screening studies across multiple rounds in the high-risk population. For example, in their prospective study of multimodality screening using digital mammography, US, and MRI, Riedl et al (8) found that the greatest improvement in specificity in incidence rounds of screening was seen with MRI (91.8% vs 84.6% in the first round). These findings suggest that limited specificity may be less problematic in the setting of ongoing MRI screening. Moreover, additional details on false-positive rates based on imaging features of lesions that underwent biopsy in both screening rounds, radiologist characteristics, and practice type may be helpful to further lower the false-positive rates.

Interval cancer rate has become a powerful indicator of the robustness of a screening tool. It is therefore highly promising that the interval cancer rate was reduced by more than half in women with extremely dense breasts who underwent supplemental MRI in the prevalence screening round of the DENSE trial (MRI-performed vs MRI-invited vs mammography only, 0.8 vs 2.5 vs 5.0 per 1000 screening examinations) (4). This result is consistent with the MRI screening outcomes reported by Kuhl et al (6) and Comstock et al (9), who found interval cancer rates of zero. Nevertheless, incidence round data on interval cancer rates would be of great interest, although not yet available.

Tumor characteristics offer another glimpse of the potential clinical impact of supplemental MRI screening. In the DENSE trial, MRI-detected invasive cancers were small (median, 9.5 mm in the prevalence round) and largely node negative (89% in the prevalence round, 100% in the incidence round). However, the vast majority of these invasive cancers were the indolent hormone receptor–positive subtypes (100% in the incidence round). Interestingly, the proportion of ductal carcinoma in situ (DCIS) went from 19% in the prevalence round to 30% in the incidence round, although all cases of DCIS in the incidence round were intermediate to high grade. A key strength of MRI as compared with mammography has been the relative predilection to identify more aggressive invasive cancers and higher-grade cases of DCIS (10), potentially mitigating overdiagnosis. This expectation is not entirely borne out in the current data. Compared with the findings of Kuhl et al (6) in average-risk women, where 41.7%–46% of MRI-detected cancers were high grade, only 6.7%–6.8% cancers in the DENSE trial were high grade (5). Early detection of small, node-negative invasive cancers of good prognosis is clinically important, especially in women with extremely dense breasts where mammography sensitivity is limited. On the other hand, the question of whether ongoing supplemental MRI screening in this population will improve mortality remains unanswered, further underscoring the need for continued data on interval cancer rates.

Some practical considerations are in order. Despite the promising results of the DENSE trial, its generalizability may be limited in many countries where more heterogeneous populations undergo annual screening. The hybrid bitemporal dynamic acquisition imaging protocol in the DENSE trial streamlines imaging and simulates abbreviated MRI, yet only 59% of women accepted MRI invitation (5). A better understanding of barriers would be paramount, as breast MRI adherence is suboptimal even among currently eligible women. Finally, as MRI technique continues to evolve, cost-effectiveness analysis is necessary to understand how to optimize supplemental screening approaches in women with dense breasts.

In summary, Veenhuizen et al presented the first round of follow-up data to the initial results of the DENSE trial, addressing the key issue of sustainability of benefit in ongoing supplemental MRI screening of women with extremely dense breasts. The current data also posed questions as to the types of cancers detected and potential overlap with mammography. Furthermore, would it be possible to substitute mammography with MRI altogether? Ultimately, what is the impact of screening MRI on the breast cancer mortality rate? We look forward to finding the answers in the next round of data.

Disclosures of Conflicts of Interest: L.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: is part of the advisory board for iCAD; has stock or stock options in Lunit; institution received compensation from Siemens for expert testimony. Other relationships: disclosed no relevant relationships. Y.G. disclosed no relevant relationships.

References

  • 1. Mann RM, Cho N, Moy L. Breast MRI: State of the Art. Radiology 2019;292(3):520–536. LinkGoogle Scholar
  • 2. Boyd NF, Guo H, Martin LJ, et al. Mammographic density and the risk and detection of breast cancer. N Engl J Med 2007;356(3):227–236. Crossref, MedlineGoogle Scholar
  • 3. Lowry KP, Coley RY, Miglioretti DL, et al. Screening performance of digital breast tomosynthesis vs digital mammography in community practice by patient age, screening round, and breast density. JAMA Netw Open 2020;3(7):e2011792. Crossref, MedlineGoogle Scholar
  • 4. Bakker MF, de Lange SV, Pijnappel RM, et al. Supplemental MRI screening for women with extremely dense breast tissue. N Engl J Med 2019;381(22):2091–2102. Crossref, MedlineGoogle Scholar
  • 5. Veenhuizen SGA, de Lange SV, Bakker MF, et al. Supplemental breast MRI for women with extremely dense breasts: results of the second screening round of the DENSE trial. Radiology 2021. https://doi.org/10.1148/radiol.2021203633. Published online March 16, 2021. LinkGoogle Scholar
  • 6. Kuhl CK, Strobel K, Bieling H, Leutner C, Schild HH, Schrading S. Supplemental Breast MR Imaging Screening of Women with Average Risk of Breast Cancer. Radiology 2017;283(2):361–370. LinkGoogle Scholar
  • 7. Tagliafico AS, Mariscotti G, Valdora F, et al. A prospective comparative trial of adjunct screening with tomosynthesis or ultrasound in women with mammography-negative dense breasts (ASTOUND-2). Eur J Cancer 2018;104:39–46. Crossref, MedlineGoogle Scholar
  • 8. Riedl CC, Luft N, Bernhart C, et al. Triple-modality screening trial for familial breast cancer underlines the importance of magnetic resonance imaging and questions the role of mammography and ultrasound regardless of patient mutation status, age, and breast density. J Clin Oncol 2015;33(10):1128–1135. Crossref, MedlineGoogle Scholar
  • 9. Comstock CE, Gatsonis C, Newstead GM, et al. Comparison of abbreviated breast MRI vs Digital Breast Tomosynthesis for breast cancer detection among women with dense breasts undergoing screening. JAMA 2020;323(8):746–756. Crossref, MedlineGoogle Scholar
  • 10. Sung JS, Stamler S, Brooks J, et al. Breast cancers detected at screening MR imaging and mammography in patients at high risk: Method of detection reflects tumor histopathologic results. Radiology 2016;280(3):716–722. LinkGoogle Scholar

Article History

Received: Jan 29 2021
Revision requested: Feb 1 2021
Revision received: Feb 15 2021
Accepted: Feb 16 2021
Published online: Mar 16 2021
Published in print: May 2021