Abstract No. 
2019 San Antonio Breast Cancer Symposium
December 10-14

Lack of background parenchymal enhancement suppression in breast MRI during neoadjuvant chemotherapy may be associated with inferior treatment response in hormone receptor positive breast cancer

Onishi N, Li W, Newitt DC, Harnish R, Gibbs J, Jones EF, Nguyen A, Wilmes L, Joe BN, Campbell MJ, Basu A, van 't Veer LJ, DeMichele A, Yee D, Berry DA, Albain KS, Boughey JC, Chien A, Clark AS, Edmiston OKELY DOKELY, Elias AD, Ellis ED, Euhus DM, Han HS, Isaacs C, Khan QJ, Lang JE, Lu J, Meisel JL, Mitri Z, Nanda R, Northfelt DW, Sanft T, Stringer-Reasor E, Viscusi RK, Wallace AM, Yung R, Melisko M, Perlmutter J, Rugo HS, Schwab R, Symmans W, Asare SM, Yau JE, Yau C, Esserman LJ, Hylton NM


In breast MRI, contrast enhancement of normal fibroglandular tissue is referred to as background parenchymal enhancement (BPE). Hormonal status significantly affects the degree of BPE, potentially due to the association with mammary vascularity and activity1-5. Studies have shown that BPE may be associated with breast cancer surviva6l , treatment response to neoadjuvant chemotherapy (NAC7),8 and future breast cancer risk9. In most patients undergoing NAC, BPE is suppressed by the nonspecific anti-proliferative effects of chemotherapy on normal breast and/or ovary5,10. However, some patients exhibit equivalent or even stronger BPE post-NAC compared to pre-NAC. We hypothesized that non-suppressed BPE in post-NAC MRI may be associated with inferior treatment response. This study aimed to investigate the association between BPE suppression and treatment response as defined by pathologic complete response (pCR).


This study included patients with stage II/III breast cancer enrolled in the I-SPY 2 TRIAL being treated with standard NAC with or without investigational agents. The whole cohort was split into two subgroups based on hormone receptor status (HR+, n= 536; HR-, n=452). Patients underwent dynamic contrast enhanced MRIs at four time points during NAC: baseline (T0), after 3 weeks of the first regimen (T1), inter-regimen (T2), and pre-surgery (T3). Using in-house software, the contralateral breast parenchyma was automatically segmented for the entire breast volume. Quantitative BPE (qBPE) was calculated as the mean early (~150s post-contrast injection) percent enhancement of the central 50% of the axial slices. A breast radiologist reviewed all exams and excluded those where automated segmentation failed to accurately define tissue. For T1, T2 and T3, BPE was categorized based on the change from T0 as suppressed (qBPE < qBPE[T0]) or nonsuppressed (qBPE ≥ qBPE[T0]). Chi-squared test was used to examine the association between BPE suppression and pCR, with p<0.05 considered statistically significant.


HR+ cohort: pCR rates were lower for patients with non-suppressed BPE than those with suppressed BPE at every visit (T1-T3) (Table 1). The difference was statistically significant at T2 (p=0.04) and T3 (p=0.01). HR- cohort: pCR rates were slightly lower for the non-suppressed BPE group, but no statistically significant association was found (Table 2).


In HR+ breast cancer, lack of BPE suppression may indicate inferior treatment response. The contrasting results in HR+ and HR- cohorts are noteworthy in terms of the possible relationship between suppression of normal mammary and ovarian activity and treatment response in HR+ cancer.


1 Radiographics 2014; 34: 234-47.

2 Radiology 1997; 203: 137-44.

3 Radiology 1997; 203: 145-9.

4 Breast J 2005; 11: 236-41.

5 AJR Am J Roentgenol 2015; 204: 669-73.

6 Eur Radiol 2018; 28: 4705-16.

7 Eur Radiol 2016; 26: 1590-6.

8 Transl Oncol 2015; 8: 204-9.

9 J Clin Oncol 2019; : JCO1800378.

10 Radiology 2015; 277: 687-96.  

View original