Background parenchymal enhancement and breast cancer: a review of the emerging evidences about its potential use as imaging biomarker

Abstract

Objectives:

To conduct a systematic review of evidences about the relationship between background parenchymal enhancement (BPE) of the contralateral healthy breast and breast cancer: its association with clinicopathological breast cancer characteristics, its potential as predictive and prognostic biomarker and the biological linkage between BPE and breast cancer.

Methods:

A computerized literature search using PubMed and Google Scholar was performed up to June 2020. Two authors independently conducted search, screening, quality assessment, and extraction of data from the eligible studies. Studies were assessed for quality and risk of bias using the revised Quality Assessment of Diagnostic Accuracy Studies tool.

Results:

Of the 476 articles identified, 22 articles met the inclusion criteria. No significant association was found between BPE and invasiveness, histological cancer type, T- and N-stage, multifocality, lymphatic and vascular invasion and histological tumour grade while the association between BPE and molecular subtypes is still unclear. As predictive biomarker, a greater decrease in BPE during and after neoadjuvant chemotherapy was associated with pathological complete response. Results about the role of BPE as prognostic factor were inconsistent. An association between high BPE and microvessel density, CD34 and VEGF (histological markers of vascularization and angiogenesis) was found.

Conclusions:

BPE of the contralateral breast is associated with breast cancer in several aspects, therefore it has been proposed as a tool to refine breast cancer decision-making process.

Advances in knowledge:

Additional researches with standardized BPE assessment are needed to translate this emerging biomarker into clinical practice in the era of personalized medicine.

Introduction

Breast cancer is the most common cancer in females with 127.5 new cases per 100,000 females per year and the second leading cause of cancer-related death among females with 20.6 of deaths per 100,000 females per year.¹ Such statistics highlight the importance of optimizing the prevention and treatment of this disease.

Background parenchymal enhancement (BPE) is defined as the enhancement of breast fibroglandular tissue after i.v. administration of a contrast agent on MRI² and is a promising tool to refine decision-making process in breast cancer. In recent years, BPE has gained great interest and was included in the last edition of the Breast Imaging Reporting and Data System (BI-RADS) MRI lexicon (Figure 1).

Figure 1.

Degrees of background parenchymal enhancement (BPE). Axial contrast-enhanced magnetic resonance images (T₁-weighted fat-suppressed subtraction maximum intensity projection images) shows minimal (a), mild (b), moderate (c) and marked (d) BPE, as described in the Breast Imaging-Reporting and Data System (BI-RADS) lexicon.

In particular, several aspects of the clinical significance of BPE in breast cancer have been investigated. First, BPE has been reported to be a factor associated with breast cancer risk. Second, the relationship between BPE reduction and therapy efficacy has been analysed after neoajuvant chemotherapy (NAC), during tamoxifen treatment for primary breast cancer prevention, and in the adjuvant setting of oestrogen receptor (ER)-positive breast cancer. Third, the association between BPE and breast cancer prognosis has also been investigated, but the results remain still controversial.

Several studies focused on tumour-induced changes in the surrounding-tumour parenchyma or evaluated BPE of the ipsilateral breast, which could be affected by the increased vascularization due to breast cancer presence.^3–5 However, given the symmetry between the two breasts, the healthy contralateral breast could be considered comparable to the ipsilateral breast before tumorigenesis and it was hypothesized that properties of the healthy parenchyma could give useful information about breast cancer. Therefore, we sought to conduct a systematic review to provide a summary of evidences about the relationship between BPE of the contralateral healthy breast and breast cancer: its association with clinicopathological breast cancer characteristics and its potential as predictive and prognostic biomarker, also exploring the biological linkage between BPE and breast cancer in order to identify new biological avenues.

Methods and materials

We followed the guidelines defined by the Preferred Reporting Items for Systematic Reviews and Meta-analyses.⁶

For the present review, a computerized literature search using PubMed (http://www.pubmed.org) and Google Scholar (https://scholar.google.com) was performed up to June 2020. A manual revision of the reference lists was also performed to integrate the initial search with additional studies.

The search strategy included various combinations of the following terms: “background parenchymal enhancement” OR “parenchymal enhancement” OR “BPE” AND “magnetic resonance imaging” OR “MRI” OR “contrast-enhanced MRI” OR “dynamic contrast-enhanced MRI” AND “breast cancer” OR “neoadjuvant chemotherapy” OR “outcome” OR “prognosis”.

Only articles in English and on human subjects were included. Excluded were (1) review articles, case reports or case series, replies to study authors, studies published only in abstract form; (2) duplicate publication; (3) on non-human subjects. No publication date restriction was used.

Titles and abstracts of search results were examined. When considered suitable, the full text was reviewed. The reference section of retrieved studies was examined to identify additional papers.

Since a systematic review and meta-analysis analysing the association between BPE at breast MRI and breast cancer risk was recently published,⁷ studies evaluating only the association between BPE and breast cancer risk were excluded.

Two authors (RR, AC) independently conducted search, screening, quality assessment, and extraction of data from the eligible studies. Disagreements arising during each phase of the study selection or during the quality assessment of the studies were resolved in consensus. If consensus could not be reached, a clinical expert (PB, with more than 15 years of experience in breast imaging) was asked to resolve any disagreement. Required data to be obtained from each eligible study were as follows: demographic data, study design, dynamic contrast-enhanced MRI characteristics, BPE assessment, breast cancer characteristics, tumour response to NAC and outcome (disease/recurrence free survival and overall survival).

The quality assessments of the eligible study were evaluated independently by two authors (RR, AC) using a modified Quality Assessment of Studies of Diagnostic Accuracy Studies (QUADAS-2) checklist, which comprised four domains: patient selection, index test, reference standard and flow and timing.⁸ The answers “yes”, “no” or “unclear” to the standard questions of each domain represent the judgement regarding bias and applicability: low risk of bias, high risk of bias and insufficient data to permit a judgement (unclear), respectively.

Results

Figure 2 shows the flowchart of the selection of studies. The initial database search identified 476 articles. A total of 61 full-text articles were assessed after removal of duplicates (n = 42), review articles (n = 48) or case report/comments (n = 60), non-English and non-human articles (n = 79) and original articles not in the field of interest on the basis of title and abstract (n = 186). From the 61 full-text articles, 19 studies were excluded because they analysed only the association between BPE and breast cancer risk; 20 studies were excluded because laterality of BPE assessment was not specified (n = 10), BPE of the ipsilateral breast of the tumour or enhancement of the stromal tissue surrounding breast tumour were evaluated (n = 8) or both breasts were evaluated to assess BPE (n = 2). After hand-searching of references, two additional papers were identified but subsequently excluded because one did not specify the laterality of BPE assessment and the other one evaluated stromal tissue surrounding breast tumour.

Figure 2.

Flowchart of the selection of studies.

A total of 22 articles were finally included in the systematic review. Table 1 showed the characteristics of the six included studies that assessed the association between BPE and clinicopathological breast cancer characteristics. Tables 2 and 3 showed the characteristics of the included studies that evaluated BPE as predictive biomarker (n = 7) and BPE as prognostic biomarker (n = 7), respectively. Two studies analysed pathobiological links between contralateral BPE and breast cancer.^29,30 The study by Moliere et al¹⁸ evaluated BPE both as predictive and prognostic biomarker and the study by van der Velden et al³¹ evaluated BPE as prognostic biomarker and analysed a possible biological process.

Table 1.

Overview of the studies that assessed the association between BPE and clinicopathological breast cancer characteristics

Study		Subjects				BPE Assessment
Reference	Design	Number (BC)	Mean age ±SD, years (range)	Menopausal status: number (percentage)	Week of menstrual cycle for breast MRI (only pre-menopausal)	Contrast Media (commercial name)	Method	Sequences for qualitative assessment or Method for quantitative assessment
Vreemann et al.9	Retrospective cohort, patients at increased breast cancer risk (>20–25% lifetime risk) with unilateral BC screen-detected on breast MRI	76 (77)	48 ± 9.8 (24–76)	Pre 36 (47) Post 41 (53)	2°/3° week	Different types	Qualitative	Post-contrast images and/or subtracted images
Kim et al.10	Retrospective cohort, patients with unilateral invasive BC who underwent preoperative breast MRI	178 (178)	NR	Pre 91 (51.1) Post 87 (48.9)	NR	Dotarem	Qualitative	MIP
Ha et al.11	Retrospective cohort, patients with unilateral invasive BC who underwent preoperative breast MRI	328 (328)	54.4 (21–86)	Pre 112 (34.1) Post 216 (65.9)	NR	Multihance	Qualitative	Unenhanced, post-contrast images and subtracted images
Dilorenzo et al.12	Retrospective cohort, patients with unilateral invasive BC who underwent preoperative breast MRI	82 (82)	NR	Pre 34 (41.5) Post 48 (58.5)	2° week	Gadovist	Qualitative	Post-contrast images and/or subtracted images
Mema et al.13	Retrospective case-control, 225 cases with triple negative BC matched 1:1 to cases with non-triple negative BC who underwent breast MRI before surgery	450 (450)	60 (23–96)	NR	NR	NR	Qualitative	Pre- and post-contrast images and subtracted images
Li et al.14	Retrospective cohort, patients with invasive BC who underwent preoperative breast MRI	163 (164)	47 (22–69)	Pre 98 (59.8) Post 66 (40.2)	NR	Magnevist	Qualitative	NR

BC, breast cancer; BPE, background parenchymal enhancement; MRI, magnetic resonance imaging; NR, not reported.

Table 2.

Overview of studies on association between background parenchymal enhancement and response to neoadjuvant chemotherapy

Study		Subjects			BPE assessment			pCR versus non-pCR			Additional relevant findings
Reference	Design	Number (BC)	Mean age ±SD, years (range)	Menopausal status: number (percentage)	Contrast Media (commercial name)	Method	Sequences for qualitative assessment or method for quantitative assessment	Baseline BPE	Post-NAC BPE	BPE change (post-NAC vs pre-NAC)
Chen et al.15	Retrospective cohort, patients with invasive BC who underwent MRI before NAC, after one or two cycles of AC and after four cycles of AC or after two cycles AC plus three weekly second-line taxane-based regimen, before surgery	46 (46)	50 ± 11 (31-77)	Pre 32 (69.6) Post 14 (30.4)	Omniscan	Quantitative	FGT segmentation	19.4 vs 15.8% (p = 0.31)	NR	NR	In pCR group BPE significantly decreased at first (15.6%, p = 0.02) and second (12.5%, p = 0.006) follow-up during NAC while in non-pCR group the BPE was not decreased at first follow-up but became significantly decreased at second follow-up (p = 0.04). This difference is more evident in pre-menopausal females and ER-negative cancers.
You et al.16	Retrospective cohort, patients with invasive BC who underwent MRI before NAC, after two, four and six cycles of NAC, before surgery	90 (90)	49.84 ± 10.04 (28–69)	Pre 50 (55.6) Post 40 (44.4)	Magnevist	Quantitative	FGT segmentation	NR	NR	NR	Decrease of BPE during NAC was associated with pCR in the three monitoring points during NAC (p = 0.002 for second NAC, 0.007 for fourth NAC and 0.036 for sixth NAC, respectively). Changes of BPE after second (p = 0.001) and fourth (p = 0.005) NAC were independent variables correlated with better tumour response in the multivariate analysis.
Oh et al.17	Retrospective cohort, patients with invasive BC who underwent pre- and post-NAC (after completion of four, six, or eight cycles of NAC) MRI, before surgery	186 (186)	45 (25–81)	Pre 128 (68.8) Post 58 (31.2)	Dotarem or Magnevist	Qualitative	Post-contrast images and MIP	2.71 ± 1.04 vs 2.51 ± 0.83 (p = 0.2019)	1.21 ± 0.14 vs 1.59 ± 0.62 (p < 0.001*)	1.50 ± 1.13 vs. 0.91 ± 0.9 (p = 0.0008*) (decrease)
Moliere et al.18	Retrospective cohort, patients with unilateral invasive BC who underwent pre- and post-NAC (after the last cycle of NAC, mean time to surgery: 15 days, range 2–35) MRI before surgery	102 (102)	49.8	Pre 55 (53.9) Post 47 (46.1)	Dotarem	Quantitative	FGT segmentation	8.1 ± 1.8 vs 8.2 ± 1.7 (p = NS)	5.4 ± 2.7 vs 4.8 ± 2.0 (p = NS)	NR
Rella et al.19	Retrospective cohort, patients with unilateral invasive BC who underwent MRI before, during (after the fourth NAC cycle) and after NAC (within 2 weeks after the completion of NAC), before surgery	228 (228)	47.6 ± 10 (24-74)	Pre 139 (61) Post 89 (39)	MultiHance	Quantitative	ROI	12.73 ± 9.38 vs 13.13 ± 10.8 (p = 0.9704)	7.18 ± 5.21 vs 7.56 ± 6.89 (p = 0.9022)	−5.85 ± 9.36 vs -5.71 ± 10.98 (p = 0.254)	Early BPE change (defined as BPE in the MRI after the fourth cycle of NAC minus BPE at the baseline MRI) resulted significantly higher in non-pCR group if compared to pCR group in Stages 3 and 4 breast cancers (p = 0.019) and HER2- tumour phenotype (p = 0.02).
You et al.20	Retrospective cohort, patients with unilateral invasive HER2+ BC who underwent MRI before and after 2 cycles of NAC, before surgery	71 (71)	47.65 ± 10.10 (26–71)	Pre 37 (52.1) Post 34 (47.9)	Magnevist	Qualitative	Post-contrast images	p = 0.287	NR	NR	Moderate or marked BPE after the second NAC and the reduction in BPE (between baseline and after the second NAC cycle) were significantly associated with pCR (OR = 43.3, p < 0.001 and OR = 0.219, p = 0.003, respectively)
Dong et al.21	Retrospective cohort, patients with invasive HER2+ BC who underwent pre- and post-NAC (timing not specified) MRI, before surgery	51 (51)	46.24 ± 8.79 (27-63)	Pre 27 (52.9) Post 24 (47.1)	Magnevist	Qualitative	NR	p = 0.894	p = 0.021*	p = 0.043* (decrease)

AC, doxorubicin and cyclophosphamide; BPE, background parenchymal enhancement; ER, estrogen receptors; FGT, fibroglandular tissue; HER2, human epidermal growth factor receptor 2; MRI, magnetic resonance imaging; NAC, neoadjuvant chemotherapy; NR, not reported; NS, not significant (p > 0.05); pCR, pathological complete response; ROI, region of interest; SD, standard deviation.

Table 3.

Overview of studies on association between background parenchymal enhancement and breast cancer outcome

Study	Reference	Van der Velden et al.22	Lim et al.23	Choi et al.24	Moliere et al.18	Van der Velden et al.25	van der Velden et al.26	Shin et al.27	Ragusi et al.28
	Design	Retrospective cohort, patients with unilateral invasive BC who underwent preoperative breast MRI	Retrospective cohort, patients with unilateral invasive BC who underwent preoperative breast MRI	Retrospective cohort, patients with unilateral invasive BC who underwent pre- and post-NAC MRI before surgery	Retrospective cohort, patients with unilateral invasive BC who underwent pre- and post-NAC (after the last cycle of NAC, mean time to surgery: 15 days, range 2–35) MRI before surgery	Retrospective cohort, patients with unilateral invasive ER+ HER2- BC who underwent preoperative breast MRI	Retrospective cohort, patients with unilateral invasive ER+ HER2- BC who underwent preoperative breast MRI	Retrospective cohort, patients with unilateral invasive BC (ER+ HER2-; node-negative; > 5 mm in size) who underwent preoperative breast MRI	Retrospective cohort, patients with unilateral invasive BC (ER+ HER2-) who underwent pre-NET and post-NET MRI (after 3 and 6 months of NET)
Subjects	Number	531 (531)	804 (804)	93 (93)	102 (102)	415 (415)	302 (302)	289 (289)	40 (40)
	Mean age ± SD, years (range)	56 ± 10	49.9 (22-88)	47 (38–54)	49.8	NR	48 (42-57)	49 (30–80)	61 (52-69)
	Menopausal Status	NR	Pre 346 (43.0) Post 458 (57.0)	Pre 43 (46.2) Post 50 (53.8)	Pre 55 (53.9) Post 47 (46.1)	NR	Pre 166 (55.0) Post 136 (45.0)	Pre 162 (56.0) Post 127 (44.0)	Pre 10 (25.0) Post 30 (75.0)
	Clinical scenario	Newly diagnosed BCs	Newly diagnosed BCs	NAC	NAC	Newly diagnosed ER+ HER2- BCs	Newly diagnosed ER+ HER2- BCs	Newly diagnosed ER+ HER2- Node-Negative	NAC(only NET, ER+ HER2− BCs)
BPE assessment	Contrast Media (Commercial Name)	ProHance	Gadovist	Gadovist	Dotarem	Prohance	Magnevist	Dotarem	Dotarem
	Method	Quantitative	Qualitative	Qualitative	Quantitative	Quantitative	Quantitative	Quantitative	Quantitative
	Sequences for qualitative assessment ormethod for quantitative assessment	FGT segmentation	Combination of Post-contrast images and/or subtracted images	Combination of pre- and post-contrast, subtracted and MIP images	FGT segmentation	FGT segmentation	FGT segmentation	ROI and FGT segmentation	FGT segmentation
	Explored variable	BPE at diagnosis	BPE at diagnosis	pre-NAC BPE	pre-NAC BPE post-NAC BPE	BPE at diagnosis	BPE at diagnosis	BPE at diagnosis	pre-NET BPE change in BPE
OS	N events	51	//	//	//	34	15	//	The authors evaluated PEPI (preoperative endocrine prognostic index, derived from the surgical excision specimen after NET and is based on pT- and pN-stage, Ki67 index, and ER-status) as a surrogate endpoint of prognosis: High pre-NAC BPE and a decrease in BPE during NET were associated with a higher PEPI-group (poor prognosis)
	HRs (high vs. low BPE)	ER+ HER2-: 0.03 Adjuvant ET: 0.00 ER+ HER2- and AST: 0.00	//	//	//	0.35	0.22 Adjuvant ET: 0.18	//
	95% CI	ER+ HER2-: 0.00-0.44 Adjuvant ET: 0.00-0.08 ER+ HER2- and AST: 0.00-0.06	//	//	//	0.05–0.99	0.00-0.83 Adjuvant ET: 0.00-0.76	//
	P value	ER+ HER2-: 0.011 Adjuvant ET: <0.001 ER+ HER2- and AST: <0.001	//	//	//	0.048	0.032 Adjuvant ET: 0.024	//
DFS	N events	73	75	23	15	42	37	17
	HR (high vs. low BPE)	ER+ HER2-: 0.19 Adjuvant ET: 0.01 ER+ HER2- and AST: 0.01	R1: 3.086* R2: 2.221*	3.851	0.038 6.38	0.42	0.27 Adjuvant ET: 0.26	NS
	95% CI	ER+ HER2-: 0.03-1.35 Adjuvant ET: 0.00-0.17 ER+ HER2- and AST: 0.00-0.17	R1: 1.452-6.559* R2: 0.923-5.346*	1.459–10.166	−0.044-0.12 0.71-12.06	0.11–1.03	0.05-0.68 Adjuvant ET: 0.05-0.71	NS
	P value	ER+ HER2-: 0.096 Adjuvant ET: 0.002 ER+ HER2- and AST: 0.002	R1: 0.003* R2: 0.075*	0.006	0.19 0.027	0.058	0.004 Adjuvant ET: 0.009	NS

AST, adjuvant systemic therapy; BC, breast cancer; BPE, Background parenchymal enhancement; DFS, disease free survival; ER, estrogen receptors; ET, endocrine therapy; FGT, Fibroglandular tissue; HER2, Human epidermal growth factor receptor 2; HR, hazard ratio; NAC, neoadjuvant chemotherapy; NET, neoadjuvant endocrine therapy; NR, not reported; NS, not significant (p > 0.05); OS, overall survival; pCR, pathological complete response; ROI: Region of interest; SD, standard deviation.

BPE and clinicopathological breast cancer characteristics

A recent systematic review and meta-analysis analysed the association between BPE at breast MRI and breast cancer risk.⁷ A higher level of BPE was observed more often in breast cancer cases than in control participants in females with high risk, either at qualitative (at least mild BPE: odds ratio [OR], 2.1; p = 0.001; at least moderate BPE: OR, 1.6; p = 0.04) or quantitative BPE assessment (standardized mean difference, 0.6; 95% Confidence Interval [CI] 0.3–0.9; p = 0.001). On the contrary, no significant association was found between BPE and breast cancer in females with average risk.⁷

So we will focus on the association between BPE and clinicopathological breast cancer characteristics. No significant association was found between BPE and invasiveness,⁹ histological cancer type,⁹ T-stage,^10–12,14 N-stage,^10,11,14 multifocality,¹⁰ lymphatic and vascular invasion¹⁰ and histological tumour grade.^10,11 Regarding the latter, only Vreeman et al⁹ found a significant negative association between BPE and tumour grade at qualitative BPE evaluation by two independent readers (reader 1: p = 0.016 and reader 2: p = 0.003 for all cancers) in females with intermediate or high risk who participated to an MRI screening program, but only in non-BRCA patients. Assuming that BPE could be considered a marker of increased hormone levels, the authors hypothesized that this association could be related to the different pathogenesis for high-grade tumours, in which hormonal stimulation is less important; this could also explain why this correlation was found only in non-BRCA patients (frequently sporadic tumours), as hormonal stimulation pathways might be different in BRCA tumour development.

About BPE and hormonal receptor status, it was hypothesized an association between BPE and ER and Progesterone Receptor (PR) statuses in patients with breast cancer. This hypothesis is based on the assumption that ER and PR expressions increase correspondingly with increasing levels of oestrogen and progesterone,³¹ and oestrogen and progesterone levels are associated with BPE.^32,33 Nevertheless, published results about this association are inconsistent.^9,10,14 Similarly, association between BPE and molecular subtypes is still unclear.^{9–12,15–18}

Table 1 summarized the currently available studies analysing the association between BPE and clinicopathological breast cancer characteristics.

BPE as predictive biomarker

A predictive biomarker provides information on the response to a specific therapy. It can be distinguished between upfront and early predictive markers: the first can be used for patient selection while the second provides information early during therapy. In this perspective, BPE evaluation could allow to tailor treatments, helping in the decision about neoadjuvant or adjuvant therapy in breast cancer patients or in determining eligibility for preventive therapy in high-risk females, or could be an early indicator of lack of response to NAC.

Primary and secondary prevention settings (endocrine therapies)

Oestrogen exposure during life has consistently been demonstrated to elevate breast cancer risk; then endocrine therapy (ET) may be used both to reduce the risk of contracting (chemoprevention)³⁴ as well as to attenuate the risk of recurrence of ER-positive breast cancers.³⁵

As previously described, BPE demonstrated utility in predicting the initial risk of breast cancer in females with high risk. It is now widely accepted that BPE is affected by the hormonal status: higher BPE levels have been documented in pre-menopausal women³⁶ as well as in post-menopausal females receiving exogenous oestrogen as a replacement therapy^36,37; moreover, serum oestrogen concentrations resulted positively associated with BPE, providing further evidence of its hormone-sensitive nature.³⁸ Thus, an intrinsic causal link appears reasonable within this triad of higher oestrogenic levels, high BPE and increased breast cancer risk. This link allows the potential use of BPE as imaging biomarker: it could be evaluated through the ET period and its reduction during therapy could indicate ET efficacy, both in the preventive (in females with high risk) and adjuvant settings.

Several publications recorded BPE changes across periods of ET, both with Selective Oestrogen Receptor Modulators (SERMs)^36,37,39 and Aromatase Inhibitors (AIs).^39,40 These studies reported an important decrease of BPE due to SERMs therapy, with a shift to minimal BPE in the majority of investigated females while the effect of AIs on BPE is less evident. The previously mentioned studies^36–40 evaluated BPE change due to ET in the adjuvant setting while only one study⁴¹ analysed BPE change in a preventive scenario. To date, no literature has directly linked ET-induced BPE reduction to the subsequent risk of developing breast cancer both in the prevention and in the adjuvant settings.

Response to neoadjuvant chemotherapy

Several studies evaluated the association between BPE and response to NAC (Table 2): a greater decrease in BPE after NAC was associated with pathological Complete Response (pCR) in all breast cancer subtypes and specifically in Human Epidermal growth factor Receptor 2 (HER2)-positive breast tumours. In particular, Oh et al¹⁷ described an average decrease in BPE of 1.5 ± 1.13 in pCR group and of 0.91 ± 0.9 in non-pCR group (p < 0.001) in 186 invasive breast cancers while Dong et al²¹ demonstrated that BPE decrease and pCR were significantly correlated (r = 0.285, p = 0.043) in 51 patients with HER2-positive tumours.

Moreover, these studies demonstrated that decrease of BPE during NAC was positively correlated with pCR: the authors evaluated BPE at varying monitoring points during NAC and found that the reduction of BPE can serve as an early prediction tool for tumour response.^17,21

Conversely, the study by Rella et al¹⁹ reported a significant association between early BPE change (defined as the enhancement rate identified in the MRI after the fourth cycle of NAC minus the enhancement rate identified at the baseline MRI) and pCR in females with Stages 3 and 4 breast cancers (p = 0.019) and who were diagnosed with a HER2-negative tumour phenotype (p = 0.02): early BPE change resulted significantly higher in non-pCR group if compared to pCR group. The authors conclude that changes in BPE during NAC may serve as an early outcome predictor in locally advanced and metastatic breast cancers and in females who were diagnosed with a HER2-negative breast cancer phenotype.

About post-NAC BPE, while the previously mentioned studies reported a significantly lower post-NAC BPE in pCR group (1.21 ± 0.47) than in non-pCR group (1.59 ± 0.62) (p = 0.0004)¹⁷ and a statistically significant correlation between post-NAC BPE levels and pCR (r = 0.322, p = 0.02),²¹ a recent study by Moliere et al¹⁸ did not find significant difference between complete responders and non-complete responders in terms of post-therapeutic BPE at quantitative BPE analysis (fibroglandular tissue segmentation).

All published studies did not found association between baseline BPE and pCR, both at qualitative and quantitative BPE assessment, so BPE evaluation before NAC beginning is not able to identify patients who will not benefit from the treatment.^15,17–21

BPE as prognostic biomarker

Only a limited number of published studies evaluated the role of BPE as a prognostic factor, and the results were inconsistent and controversial (Table 3).

Concerning the relationship between BPE and prognosis in newly diagnosed breast cancers, van der Velden et al²² found that patients with low BPE (evaluated as LE90, mean of the top 10% of late enhancement in contralateral breast) showed worse prognosis in terms of Overall Survival (OS) and Disease Free Survival (DFS), especially in ER-positive, HER 2-negative breast cancers (Hazard Ratio [HR], 0.03; CI 0.00–0.44; p = 0.011 for OS and HR, 0.19; CI 0.03–1.35;p = 0.096 for DFS). Moreover, BPE resulted associated with long-term outcomes in patients who underwent endocrine adjuvant therapy, with lower BPE of the contralateral breast showing potential as a predictive biomarker for relatively poor outcome in this subgroup of patients. This result was confirmed by subsequent studies conducted by the same authors^25,26 in the subgroup of patients with ER-positive, HER2-negative breast cancers. Furthermore, being a character of the “breast” rather than of the “tumour” with pathogenic links to breast cancer risk, BPE may contain complementary information. Thus, the same group of authors^25,26 tried to incorporate BPE into prognostic models derived from the primary tumour: in both the biomarker-discovery study²⁵ and the subsequent biomarker-assessment study,²⁶ BPE showed complementary ability to improve the risk stratification of routine prognostic models. In particular, they investigated the association between BPE and survival in patients at high risk according to the Nottingham Prognostic Index (NPI) and according to PREDICT, considering at high risk patients with an NPI > 3.4 and/or with a 10-year overall survival <85% according to PREDICT. In patients at high risk as estimated by both tested prognostic models, those with high BPE had a significantly better outcomes compared to the patients with low BPE, either in terms of DFS (HR, 0.15; CI 0.02–0.41;p < 0.001 according to NPI and HR, 0.16; CI 0.02–0.45;p < 0.001 according to PREDICT) and OS (HR, 0.20; CI 0.00–0.83;p = 0.029 according to NPI and HR, 0.17; CI 0.00–0.76;p = 0.021 according to PREDICT).³¹

The study by Shin et al²⁷ did not confirmed this results in 289 consecutive patients with the same breast cancer molecular subtype (ER-positive, HER2-negative breast cancers): quantitative assessments of contralateral BPE (evaluated using LE90, the same method of the previous described studies)^22,25,26 did not show an association with DFS while a high Ki-67 expression level was the only independent factor associated with worse outcome. However, these contrasting results may be related to the different study populations: Shin et al²⁷ evaluated only node-negative breast cancers while the studies by Van der Velden et al^22,25,26 included 55–66% of patients with nodal involvement. So probably in the subgroup of patients studied by Shin et al,²⁷ who already have a good prognosis, contralateral BPE may provide less additional information after consideration of tumour proliferation activity (Ki-67 level). Moreover, the two study populations underwent different treatment methods, including type of surgery, radiation therapy, and adjuvant chemotherapy; in particular, all patients included in the study by Shin et al²⁷ received adjuvant endocrine therapy and 52.2% adjuvant chemotherapy (versus 93 and 65.9%, respectively, in the study by van der Velden et al.)²⁶ These differences may have an impact on survival outcome and influence the different prognostic value of BPE. Furthermore, Lim et al²³ reported a relationship between high BPE and lower DFS in HER2-positive (HR, 6.84; CI 1.32–35.31; p = 0.022) and triple negative (HR, 2.23; CI 1.10–4.53;p = 0.026) breast cancer subtypes.

In an NAC setting, conflicting results have been published about the association between pre-NAC BPE of the contralateral breast and outcome^18,24 while a strong positive association was found between post-NAC BPE and risk of recurrence (HR, 6.38; CI 0.71–12.06;p < 0.05).¹⁸ As previously described, post-NAC BPE reduction was associated with pCR^17,21 so the decreased intensity of BPE after NAC may be an indirect marker of chemotherapy action.

Insights from the pathobiological links between BPE and breast cancer

It is also worth considering the biology behind the association between BPE and breast cancer. About the relationship between BPE and breast cancer risk, the observation of increased risk only in high-risk population leads the authors to hypothesize that in breast tissues that lacks of repair mechanisms (due to genetic defects or to genetic alterations result of radiation exposure), tissue activation could lead to malignant transformation. In this scenario, BPE may be considered a measure of tissue activity. However, precise mechanisms of tumorigenesis interlinked with BPE (including for example angiogenesis and inflammation) need to be investigated to further validate BPE as intrinsic to breast cancer development. For example, recently Sung et al²⁹ found a strong correlation between high BPE (qualitatively evaluated using a combination of pre-contrast, the first post-contrast and subtraction images) and microvessel density, CD34, and VEGF, three histological markers of vascularization and ongoing angiogenesis. These results are strongly supportive of the common assumption that BPE is a measure of microvessels. Moreover, VEGF is a potent mitogen for micro- and macro-vascular endothelial cells and is associated with increased vascular permeability. A high correlation between VEGF expression and microvessel density was previously noted in breast cancer samples⁴² so it could be one of the biological links between BPE and breast cancer development.

Likewise, it is also important to investigate the biology behind the association between BPE and breast cancer outcome, both for recurrence and mortality.

Considering high BPE as a marker of higher parenchymal perfusion, as discussed above, it could facilitate drug-transport, influencing breast cancer prognosis. Moreover, the described association between BPE and VEGF,²⁹ a key regulator of vascular permeability, could underpin the link between BPE and capillary permeability, that could also be linked to better pharmacokinetics in both neoadjuvant and adjuvant settings.

It was also proposed a relationship between BPE and development of breast cancer with higher hormone sensitivity (and so better prognosis) but no significant correlation was found between ER- and PR-percentage of staining at pathology and BPE²⁶ nor with genomic ER-pathway activity.³⁰

According to another hypothesis, BPE could be linked to immune response, but specific involved mechanisms have not yet been tested.

Finally, BPE change in the adjuvant setting could be an imaging biomarker for favourable pharmacokinetics such as increased drug activation, longer retention or reduced deactivation but the specific involved enzymes need to be studied.

Risk of bias

Assessment of the methodological quality of the included studies by the modified QUADAS-2 tool is depicted in Table 4.

Table 4.

Risk of bias table demonstrating the overall risk of bias for each of the domains of patient selection, index test, reference standard, and flow and timing

Study	Patients selection	Index test	Reference standard	Flow and timing
Vreeman et al.9	Low	Low	Low	Low
Kim et al.10	Low	Low	Low	Low
Ha et al.11	Unclear	Low	Low	Unclear
Dilorenzo et al.12	Low	Unclear	Low	Low
Mema et al.13	High	Low	Low	Unclear
Li et al.14	Unclear	Unclear	Low	Low
Chen et al.15	Low	High	Low	High
You et al.16	Low	Low	Low	Low
Oh et al.17	Low	Unclear	Low	Unclear
Moliere et al.18	Low	Unclear	Low	Unclear
Rella et al.19	Low	Unclear	Low	Low
You et al.20	Low	Unclear	Low	Low
Dong et al.21	Low	Unclear	Low	Unclear
Van der Velden et al.22	Low	Low	Low	Low
Lim et al.23	Low	Unclear	Low	Low
Choi et al.24	Low	Low	Low	Low
Van der Velden et al.25	Low	Low	Low	Low
Van der Velden et al.26	Low	Low	Low	Low
Shin et al.27	Low	Low	Low	Unclear
Ragusi et al.28	Low	Low	Low	Low
Sung et al.29	Low	Unclear	Low	Low
Van der Velden et al.30	Low	Low	Low	Low

The domain of “patient selection” was unclear in the studies of Ha et al¹¹ and Li et al.¹⁴ Mema et al¹³ conducted a case-control study comparing triple negative breast cancer and non-triple negative breast cancer patients matched with a propensity score generated using a logistic regression model based on the patients’ age at diagnosis; however, other confounding variables (first of all menopausal status) were not balanced and this could introduce a selection bias. The domain of “index test” was described in detail in most of the studies, except the studies by Li et al¹⁴ and Dong et al²¹ where sequences for qualitative BPE assessment were not specified. A risk of bias was judged in the study of Chen et al¹⁵ for the low quality of MRI examinations where the BPE assessment was performed and in the studies where it was not reported if the readers were blinded to all other information when qualitatively evaluating BPE.^{12,17–20,23} About the domain of “flow and timing”, several studies did not specify the timing of breast MRI or just reported the terms “pre-operative”, “before NAC” or “after NAC”, without describing the exact time interval. Moreover, in the study by Chen et al¹⁵ the first follow-up MRI was performed after one cycle or two cycles of doxorubicin and cyclophosphamide and the second follow-up MRI was performed after receiving four cycles of doxorubicin and cyclophosphamide or after two cycles of doxorubicin and cyclophosphamide plus three weekly second-line taxane-based regimen; these differences in the timing of MRI examinations could affect BPE evaluation. Figure 3 shows the grouped bar graph of risk of bias for the included studies, as determined using the QUADAS-2.

Figure 3.

Grouped bar graph of risk of bias for the included studies, as determined using the revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). Gray bars denote low risk, white bars denote high risk and black bars denote unclear risk.

Discussion

We conducted a systematic review of the currently available literature on the relationship between BPE of the contralateral healthy breast and breast cancer. In the last years, interest in BPE has steadily increased, also due to an expanding use of MRI for breast cancer screening, diagnosis and staging. In particular, specific interest was directed to the link between BPE and breast cancer, as the consequence of its relationship with breast cancer risk and its potential use as imaging biomarker. As breasts are symmetric organs, BPE of the contralateral breast can be considered as an intrinsic feature of breast parenchyma, specific of individual females due to several factors, and may potentially contain complementary information about the risk of breast cancer and treatment outcomes.

In particular, high BPE has been recognized as a breast cancer risk factor in high-risk females but the association between BPE and specific breast cancer characteristics (such as tumour grade, histology, molecular subtype) need to be clarified, due to their important prognostic impact. It still needs to be clarified whether contralateral BPE is associated with breast cancer subtypes with poor prognostic outcome, especially in high-risk females with high BPE levels. This might support the creation of dedicated screening programs and adapted intervals for surveillance by breast MRI. To date, the published studies on this topic are extremely heterogeneous in terms of populations, MRI protocols and BPE assessment, as highlighted by Table 1.^9–14 This could affect BPE data and could be responsible of the contrasting results. Moreover, most of the studies included patients with already diagnosed cancers, but do not deal with the prospective risk of developing certain types of cancers, so further longitudinal studied are required.

BPE has been proposed as predictive biomarker: it could be of value before treatment beginning to predict response to therapy or it could be monitored during treatment to evaluate therapy effectiveness. In this perspective, BPE could be a perfect candidate thanks to the possibility to be measured non-invasively and to its characteristic of being a marker of the “breast” (not of the tumour) that consequently could be monitored also after breast cancer resection. To date, regarding prediction of response to treatments, BPE evaluation during NAC was found to be associated with tumour response: its evaluation early during therapy seems to provide useful information about patients who are more likely to achieve pCR versus patients who will not benefit from NAC and who can be promptly sent to surgery. It could be argued that the BPE decrease during NAC could indicate a better delivery of chemotherapy agents into the normal breast tissue. Indeed, BPE can be considered a marker of parenchymal perfusion and, since chemotherapy agents exert effects on tumour vessels by loss of pro-angiogenic support,⁴³ its reduction could be a sign of more severe vascular injury. Conversely, the association between post-NAC BPE and pCR needs to be cleared up. Post-NAC BPE evaluation could be useful in the cases where diagnostic accuracy of breast MRI in assessing residual disease or pCR after NAC is known to be less effective. In particular, it was previously demonstrated that breast cancer subtype and treatment regimen could affect MRI estimation of pCR: lower performance has been associated with the Luminal phenotype vs triple-negative and HER2-positive breast cancer subtypes⁴⁴ and with chemotherapy regimen including docetaxel (due to its antivascular effect resulting in less tumour enhancement).⁴⁵ An accurate evaluation of presence and size of residual disease after NAC is essential for the planning of subsequent surgery and to avoid positive resection margins. Moreover, it was recently proposed to omit breast surgery in patients who achieved pCR after NAC, treating them with radiotherapy alone.⁴⁶ The main drawback to potential omission of surgery is that imaging modalities are insufficiently accurate to predict pCR.^47,48 From this perspective, post-NAC BPE evaluation could be an additional tool to evaluate response to NAC and to select patients who could avoid breast surgery.

On the contrary, no literature has directly linked ET-induced BPE reduction to the subsequent risk of developing breast cancer both in the prevention and in the adjuvant settings, so further studies are needed to determine whether the reduction of BPE is indicative of a concomitant reduction of breast cancer risk or tumour recurrence. Immunohistochemical biomarkers or multigene tests (such as Mammaprint, Oncotype DX) are indeed helpful to identify patients who may benefit from ET. Nonetheless, such biomarkers are not useful as monitoring tools, as they are inherent characteristics of the initial tumour and their changes cannot be measured after tumour excision to reflect ET action.⁴⁹ Conversely, BPE can be measured through the treatment period, with its change potentially reflecting ET efficacy, and maintaining utility also after surgical excision of the cancerous breast tissue. Thus, BPE is intuitively a compelling predictive biomarker, analogously to mammographic breast density.⁵⁰ In particular, to validate BPE reduction as “predictive”, correlation with outcome needs to be evaluated in prospective clinical trials with the biomarker study forming part of the prospective design.

Finally, BPE could play a role as prognostic biomarker, providing information about patients’ cancer outcome. Regarding this aspect, the transformation of a routine clinical MRI examination from a diagnostic tool to one with prognostic implications could increase the utility of pre-treatment imaging. To date, the two main prognostic studies focused on ER-positive HER2-negative tumours,^26,27 which constitute the largest subset of breast cancer patients. This subset of breast cancer patients has a significant clinical benefit from adjuvant endocrine therapy⁵¹ while the benefit of systemic chemotherapy for the individual female remains uncertain. Clinical and pathological factors (such as patient age, tumour size, tumour receptor status and nodal involvement) are used to select chemotherapy using prognostic models and, more recently, further individualization of chemotherapy has been accomplished using molecular assays.⁴⁹ However, these markers all focus on the tumour. If confirmed, the evaluation of BPE of the contralateral healthy breast could represent a marker of the stroma to complement prognostic models and further stratify patients, with a potential reduction of overtreatment.

About the pathobiological links between BPE and breast cancer, to date, only a limited number of known mechanisms of pathogenic importance to breast cancer have been found to be interlinked with BPE, including angiogenesis and inflammation, which further validate BPE as associated to breast cancer development and prognosis. The identification of specific molecules within these functional domains may indicate that these molecules are particularly important to oncogenesis and thereby, could elevate them to suitable drug targets.

Despite the great interest and existing evidences about the link between BPE and breast cancer, a number of limitations remain. First, there is a variability in the designs and methods of the included studies. Second, an extreme heterogeneity in BPE assessment is evident, as recognized by recent reviews.^52,53 This could be, at least in part, responsible of the conflicting results of the published articles. Indeed, BPE can be qualitatively assessed using the BIRADS scores or quantitatively evaluated. In the first case, a huge heterogeneity in the MRI sequences evaluated to assess BPE was evident and this method is limited by readers’ subjectivity, with a wide variability of inter-reader agreement. Similarly, numerous methods are used for quantitative evaluation: region of interest-based methods, fibroglandular tissue segmentation and fully-automated methods. Moreover, BPE calculation is made using different formulae, ultimately based on two different methods: volume-based measurements, that need the selection of early or late post-contrast phase and the choice of a threshold for the derived measures (e.g., volume of breast with percent enhancement >threshold) versus absolute signal enhancement or its ratio of background parenchyma, using both pre-contrast, early and late post-contrast phases, and its derived measures (e.g., LE90).

Furthermore, it is now widely accepted that the enhancement of normal breast tissue is affected by the hormonal status: BPE varies with age, week of menstrual cycle and menopausal status.^32,33 So another important factor that could be responsible of the conflicting published results is the different proportion of premenopausal patients in each population, and the timing of the MRI in the menstrual cycle in premenopausal patients. The latter was unknown or not available in the majority of studies especially regarding the association between BPE and clinicopathological breast cancer characteristics, as shown in Table 1.

Finally, also the contrast media used is extremely composite: a higher enhancement of breast parenchyma has been reported with the use of high-relaxivity contrast media (such as gadobenato dimeglumina).⁵⁴

BPE of the contralateral healthy breast resulted to be associated with breast cancer in several aspects, therefore it has been proposed as a tool to refine the breast cancer decision-making process. In particular, it was found to be associated with breast cancer risk in females with high risk (not in females with average risk) but its relationship with specific breast cancer subtypes needs to be cleared up. BPE shows potential as predictive biomarker, both for preventative and adjuvant breast cancer endocrine therapies and to evaluate response to NAC. It has also potential as a prognostic biomarker, to tailor the type and the rate of post-operative follow-up. However, while the existing evidences for BPE are compelling, to date its clinical value is still limited, especially due to disparate methods for BPE assessment. Objectivity and reliability are essential to foresee its employment in the field of personalized medicine. Thus, additional researches are needed to clarify how to best evaluate BPE, and a standardized protocol for its measurement is required to improve reliability. This is essential to translate this emerging biomarker into clinical practice in the era of personalized medicine.