Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients

Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients

The Breast xxx (2015) 1e6 Contents lists available at ScienceDirect The Breast journal homepage: www.elsevier.com/brst Original article Mammograph...

542KB Sizes 0 Downloads 18 Views

The Breast xxx (2015) 1e6

Contents lists available at ScienceDirect

The Breast journal homepage: www.elsevier.com/brst

Original article

Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients S. Elsamany a, b, *, A. Alzahrani a, W.N. Abozeed c, d, A. Rasmy e, f, M.U. Farooq g, M.A. Elbiomy b, E. Rawah h, K. Alsaleh c, N.M. Abdel-Aziz c, i a

Oncology, King Abdullah Medical City, Makkah, Saudi Arabia Oncology, Oncology centre, Mansoura University, Mansoura, Egypt Medical Oncology, King Khaled Hospital, King Saud University, Riyadh, Saudi Arabia d Clinical Oncology, Mansoura University Hospital, Mansoura, Egypt e Oncology, King Fahd Specialist Hospital, Dammam, Saudi Arabia f Oncology, Zagazig University Hospital, Zagazig, Egypt g Research, King Abdullah Medical City, Makkah, Saudi Arabia h Radiology, King Abdullah Medical City, Makkah, Saudi Arabia i Medical Oncology, South Egypt Cancer Institute, Assuit University, Assuit, Egypt b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 January 2015 Received in revised form 8 May 2015 Accepted 20 May 2015 Available online xxx

Background: This study aims to evaluate the relation between mammographic breast density (BD) and pathological response to neoadjuvant chemotherapy. Methods: In this retrospective study, 241 breast cancer patients who received neoadjuvant chemotherapy were included. BD was assessed in mammograms already performed at diagnosis. Pathological complete response (pCR) and pathological stage were correlated with BD, tumour phenotype and other clinico-pathological factors. Results: Patients with low BD had better pCR compared to those with high density (30.5% vs 19.5% respectively, OR ¼ 1.8, 95% CI ¼ 0.98e3.3, p ¼ 0.056) which was more pronounced after adjustment with body mass index (BMI) (OR ¼ 2.4, 95% CI ¼ 1.2e4.8, p ¼ 0.011). HER2-positive disease (32.5% vs. 18.4%, OR ¼ 2.2, 95% ¼ 1.2e4.0, p ¼ 0.01), lower BMI (OR ¼ 1.1, 95% CI ¼ 1.03e1.15, p ¼ 0.004) and lower clinical stage (p ¼ 0.002) were significant predictors of pCR in univariate analysis. In multivariate analysis, low BD (OR ¼ 2.7, 95% CI ¼ 1.3e5.5, p ¼ 0.006) and lower BMI (OR ¼ 1.1, 95% CI ¼ 1.03e1.17, p ¼ 0.003) were independent predictors of better pCR, while early clinical stage (I, II) was of borderline significance (OR ¼ 2.6, 95% CI ¼ 0.99e6.7, p ¼ 0.052). High BD (OR ¼ 1.8, 95% CI ¼ 1.1e3.2, p ¼ 0.03), advanced clinical stage (III) (OR ¼ 1.5, 95% CI ¼ 1.03e2.1, p ¼ 0.03) and higher BMI (OR ¼ 1.06, 95% CI ¼ 1.02e1.11, p ¼ 0.006) were significant predictors of advanced pathological stage. Conclusion: Low mammographic BD, low BMI and early clinical stage were associated with improved pCR rate and lower pathological stage after neoadjuvant chemotherapy. BD had more pronounced association with response to chemotherapy after adjustment with BMI. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Mammogram Breast density Neoadjuvant chemotherapy Pathological response

Introduction Neoadjuvant chemotherapy was originally applied in locally advanced breast cancer patients with the aim of tumour downstaging to allow for surgical resection [1]. Evidence then emerged

* Corresponding author. Medical Oncology Department, Oncology Centre, King Abdullah Medical City, 2677 Al-Mashaeer District, Makkah 57657, Saudi Arabia. Tel.: þ966 2 5549999; fax: þ966 2 5532239. E-mail address: [email protected] (S. Elsamany).

that induction of pathological complete response (pCR) is predictive of long-term outcome with improved survival [2]. Several predictive factors for response to neoadjuvant chemotherapy have been studied [3], however the link between mammographic breast density (BD) and response to neoadjuvant chemotherapy is not yet clarified. Mammographic BD is expressed as the percentage of the mammogram occupied by radio-dense tissue relative to breast volume [4]. The radio-dense tissue, which consists of stromal and epithelial components, is assumed to reflect the target tissue of breast cancer [5]. Different amounts of epithelial, stromal cells,

http://dx.doi.org/10.1016/j.breast.2015.05.007 0960-9776/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007

2

S. Elsamany et al. / The Breast xxx (2015) 1e6

collagen and fat contribute to variations in BD which may influence the risk of breast cancer initiation and progression [6]. Several reports displayed that high BD is strongly associated with increased breast cancer risk [4]. In a meta-analysis involving thousands of breast cancer patients and healthy individuals from 42 studies, high BD was associated with increased risk of breast cancer [6]. Higher mammographic BD may increase the probability of cancer initiation through incorporating a larger amount of at risk-cells that may proceed to malignancy within the suitable microenvironment [7]. In addition, several reports indicate that the extracellular matrix (ECM) contributes to neoplastic progression and its disturbance may precede epithelial changes [8]. Interactions between the ECM and breast epithelial cells are required for tumorigenesis and there is a growing list of molecules that mediate the effect of the ECM during neoplastic progression [9,10]. In the present study, we assessed the utility of mammographic BD in predicting pathological response to neoadjuvant chemotherapy. This approach may provide a simple, non-costly parameter that can aid in the assessment of the likelihood of benefit from neoadjuvant chemotherapy for the individual patient. Patients and methods Study population We screened female patients with histologically confirmed breast cancer who received neoadjuvant chemotherapy and presented to four institutes; three in Saudi Arabia and one in Egypt from September 2012 to September 2013. We included patients with locally advanced disease or those who required downsizing to be eligible for conservative breast surgery. We selected patients who have received at least four cycles of neoadjuvant chemotherapy; anthracycline-based, taxane-based or both. The choice of chemotherapy was at the discretion of the treating physician. Available baseline mammogram before starting neoadjuvant chemotherapy is a must for inclusion in the study. Study design and procedures In this retrospective study, clinico-pathological data was collected including body mass index (BMI), pre-chemotherapy tumour size and lymph node (LN) status as assessed in baseline mammograms. Tumour phenotype was recorded from data derived from diagnostic biopsies. Oestrogen receptor (ER)/progesterone receptor (PR) were considered positive if >1% of cells showed positive staining. Hormonal receptor (HR) positivity was defined as ER and/or PR positive, while patients with negative staining for both ER and PR were considered HR-negative. ER positivity was quantified according to Allred score based on the percentage of positive cells and intensity of staining. HER2 status was assessed by immune-histochemistry (IHC) in addition to FISH confirmation in cases with (þ2) by IHC. Treatment data was recorded including type and number of chemotherapy cycles. Furthermore, pathological tumour size and nodal status were derived from pathological reports of definitive breast surgery. BD was assessed through visual assessment of screen films of two-view baseline mammograms already performed at the time of diagnosis. Density was semiquantitatively assessed by one on-site radiologist in each institute and was classified based on the method originally proposed by Wolfe [11] as follows: low BD indicating radio-dense fibroglandular tissue 25% of the breast and high BD described radio-dense tissue >25%. PCR and post-chemotherapy pathological stage were correlated with mammographic BD in addition to other parameters. PCR was defined as absence of any invasive carcinoma at the breast or axillary LNs at the time of definitive breast surgery. Meanwhile,

tumour stage was recorded according to TNM staging system of the American Joint Committee on Cancer (AJCC), 7th edition. Statistical analysis The data was analysed using SPSS version 16 (SPSS Inc., Chicago, IL, USA). We examined the distribution of patients and treatment characteristics in low compared to high BD patients using Chisquare test to identify possible associations with the two BD categories. A logistic and an ordinal regression models were constructed using pCR and pathological stage after surgery as the dependent variables respectively. The following parameters were entered into each model as explanatory variables in a univariate manner: BD (mild vs. high), BMI (as a categorical and a continuous variable), clinical stage, other clinico-pathological parameters, tumour phenotype (HR-positive vs. HR-negative; Allred score 2e6 vs. Allred score 7e8; HER2-positive vs. HER2-negative), type and number of chemotherapy cycles (<6 vs. 6e8). BD was adjusted for BMI as a continuous variable and adjusted BD was correlated with pCR and pathological stage. For each dependent variable, a multivariate model was constructed that included all factors with significant association in univariate analysis. An alpha level of <0.05 was considered significant for all two tailed comparisons. In addition, receiver operating characteristic (ROC) analysis was conducted to estimate overall predictive accuracy of the parameters having significant association with pCR in univariate analysis. Results In this study, 255 breast cancer patients who received neoadjuvant chemotherapy were screened. Fourteen patients were excluded (7; no available baseline mammogram, 5; received <4 cycles of neoadjuvant chemotherapy, 2; refused surgery) (Fig. 1). Among the 241 patients included, BD at baseline mammogram was low in 34% and high in 66% of cases. Among our patients, 61.4% were premenopausal, 20.7% were younger than 40 years and 76.8% had clinical stage III at diagnosis. Only 3 patients with stage I disease who initially refused surgery were included. They have agreed later to go for definitive surgery after receiving primary chemotherapy course. Regarding tumour phenotype, 34% of patients were HER2-positive, 67.6% were HR-positive and 62.6% of them had Allred score 7e8. The majority of patients received 6e8 cycles of neoadjuvant chemotherapy (76.8%) and two thirds (67.2%) received both anthracycline and taxane-containing therapy (3e4 cycles of anthracycline-based followed by 3e4 cycles of taxane-based chemotherapy and trastuzumab was added to taxane in HER2positive patients) while 25.3% received anthracycline-only chemotherapy (FEC regimen; fluorouracil, epirubicin, cyclophosphamide). Only, 7.5% of patients received taxane-only chemotherapy (TCH; docetaxel, carboplatin, trastuzumab). Among our cohort, pCR rate was 23.2% while pathological stages I, II, and III were encountered in 8.7%, 26.6% and 41.5% of patients, respectively. Distribution of patients' characteristics according to breast density categories Patients with high BD were more likely to have younger age at diagnosis (p ¼ 0.01) with HR-positive phenotype (high; 71.7% vs. low; 59.8%, p ¼ 0.04). However, low and high BD patients had similar distribution of clinical stage at diagnosis (p ¼ 0.16), BMI (p ¼ 0.44) and HER-2 status (p ¼ 0.85). High BD patients received more commonly chemotherapy regimens containing both anthracyline and taxane (high; 71.1% vs. low; 59.8%, p ¼ 0.03) and higher number of cycles (6e8 cycles) (high; 82.4% vs. low 65.9%, p ¼ 0.006) (Table 1).

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007

S. Elsamany et al. / The Breast xxx (2015) 1e6

3

Fig. 1. Flow diagram of the study. BC; breast cancer, NAC; neoadjuvant chemotherapy, BD; breast density.

Predictors of response to neoadjuvant chemotherapy Low BD patients were more likely to have pCR compared to those with high BD, however, the difference was of borderline significance (30.5% vs 19.5% respectively, OR ¼ 1.8, 95% CI ¼ 0.98e3.3, p ¼ 0.056) (Table 2). Noteworthy, after adjustment with BMI as a continuous variable, low BD was more significantly associated with pCR (OR ¼ 2.4, 95% CI ¼ 1.2e4.8, p ¼ 0.011). In addition, certain subgroups of patients were more likely to have pCR including HER2-positive patients (32.5% vs 18.4%, OR ¼ 2.2, 95% CI ¼ 1.2e4.0, p ¼ 0.01), those with lower clinical stage at diagnosis (p ¼ 0.002) and low BMI (as a categorical variable) (p ¼ 0.003) (Table 2). With logistic regression analysis including BMI as a continuous variable, the likelihood of having pCR increased with the decline of BMI value (OR ¼ 1.1, 95% CI ¼ 1.03e1.15, p ¼ 0.004) indicating that with every one figure decrease in BMI value, the possibility of pCR increases by 10%. Using ordinal regression analysis, patients with higher clinical stage (p ¼ 0.04) and high BMI as a categorical variable (OR ¼ 1.64, 95% CI ¼ 1.23e2.19, p ¼ 0.001) were significantly associated with advanced pathological stage (stage III) after neoadjuvant chemotherapy (Table 3). Similarly, increasing BMI (as a continuous variable) was associated with increased pathological stage (OR ¼ 1.06, 95% CI ¼ 1.01e1.10, p ¼ 0.008). Unadjusted BD was not significantly associated with the pathological stage (p ¼ 0.10) (Table 3). Meanwhile, high BD was significantly associated with advanced

pathological stage (stage III) after BMI-adjustment (OR ¼ 2.4, 95% CI ¼ 1.1e5.2, p ¼ 0.022). Noteworthy, ER/PR status, Allred score, type of chemotherapy and number of cycles were not significantly associated with pCR (Table 2) or pathological stage (Table 3). In multivariate logistic regression analysis including BMI as a continuous variable, low breast density (OR ¼ 2.7, 95% CI ¼ 1.3e5.5, p ¼ 0.006) and lower BMI (OR ¼ 1.1, 95% CI ¼ 1.03e1.17, p ¼ 0.003) were independent predictors of pCR, while early clinical stage (I, II) was of borderline significance (OR ¼ 2.6, 95% CI ¼ 0.99e6.7, p ¼ 0.052). Similarly, in multivariate ordinal regression analysis, high BD (OR ¼ 1.8, 95% CI ¼ 1.1e3.2, p ¼ 0.03), advanced clinical stage (III) (OR ¼ 1.5, 95% CI ¼ 1.03e2.1, p ¼ 0.03) and higher BMI (OR ¼ 1.06, 95% CI ¼ 1.02e1.11, p ¼ 0.006) were independent predictors of advanced pathological stage after neoadjuvant chemotherapy. In the ROC curve analysis of predictors of pCR (including BMI as a continuous variable, BD, clinical stage and HER2 status), area under the curve and difference from line of zero discrimination was 0.64 for BMI (95% CI ¼ 0.56e0.72, p ¼ 0.002), 0.54 for clinical stage (95% CI ¼ 0.45e0.63, p ¼ 0.40), 0.59 for BD (95% CI ¼ 0.50e0.70, p ¼ 0.06) and 0.56 for HER2 status (95% CI ¼ 0.50e0.70, p ¼ 0.17) (Fig. 2). Discussion Neoadjuvant chemotherapy has been increasingly used in the current practice after displaying equivalent survival compared to

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007

4

S. Elsamany et al. / The Breast xxx (2015) 1e6

Table 1 Distribution of different parameters according to breast density categories. Parameters

Age 40 >40e55 >55 Menopause Premenopause Postmenopause Grade 1/2 3 Multicentric tumours Yes No LVIa Yes No Clinical stage I II IIIA IIIB BMIb 18e24.9 25e29.9 30e39.9  40 ER/PR status Negative Positive Allred score 2-6 7-8 HER2 Negative Positive Type of chemotherapy Anthracycline and taxane Anthracycline only Taxane only No. of cycles <6 6e8 a b

Low ¼ 82

High ¼ 159

N (%)

N (%)

Table 2 Pathological complete response (pCR) rate according to different parameters. p

Parameters

ORa (95% CIb)

p

12/50(24.0%) 44/191(23.0%)

1.06(0.51e2.19)

0.89

34/148(23.0%) 22/93(23.7%)

0.96(0.5e1.8)

0.90

31/137(22.6%) 25/104(24.0%)

0.92(0.51e1.69)

0.80

11/52(21.2%) 45/189(23.8%)

0.9(0.4e1.9)

0.74

24/110(21.8%) 32/131(24.4%)

0.9(0.5e1.6)

0.63

17/35(48.6%) 15/63(23.8%) 19/96(19.8%) 2/19(10.5%)

e

0.003

3/3(100%) 14/53(26.4%) 28/106(26.4%) 11/79(13.9%)

e

0.002

18/78(23.1%) 38/163(23.3%)

0.96(0.5e1.8)

0.90

14/60(23.3%) 22/98(22.4%)

1.1(0.5e2.4)

0.81

27/83(32.5%) 29/158(18.4%)

2.2(1.2e4.0)

0.01

19/79(24.1%) 37/162(22.8%)

1.1(0.6e2.0)

0.83

45/185(24.3%) 11/56(19.6%)

1.5(0.7e3.1)

0.34

25/82(30.5%) 31/159(19.5%)

1.8(0.98e3.3)

0.056

pCR n/N (%)

1(18.3%) 34(41.5%) 33(40.2%)

35(22.0%) 89(56.0%) 35(22.0%)

0.01

44(53.7%) 38(46.3%)

104(65.4%) 55(34.6%)

0.08

46(56.1%) 36(43.9%)

91(57.2%) 68(42.8%)

0.66

13(15.9%) 69(84.1%)

39(24.5%) 120(75.5%)

0.08

31(37.8%) 51(62.2%)

79(49.7) 80(51.3%)

0.08

2(2.4%) 17(20.7%) 30(36.6%) 33(40.3%)

1(0.7%) 36(22.6%) 76(47.8%) 46(28.9%)

0.16

9(14.5%) 15(24.2%) 31(50.0%) 7(11.3%)

27(17.9%) 49(32.4%) 64(42.4%) 11(7.3%)

0.44

33(40.2%) 49(59.8%)

45(28.3%) 114(71.7%)

0.04

20(41.7%) 28(58.3%)

40(36.4%) 70(63.6%)

0.58

54(65.9%) 28(34.1%)

104(65.4%) 55(34.6%)

0.85

49(59.8%) 29(35.4%) 4(4.8%)

113(71.1%) 32(20.1%) 14(8.8%)

0.03

28(34.1%) 54(65.9%)

28(17.6%) 131(82.4%)

0.006

LVI, lymphovascular invasion. BMI, body mass index (BMI data is available for 213 patients only).

Age 40 >40 Menopause Premenopause Postmenopause Grade 1/2 3 Multicentric tumours Yes No LVIc Yes No BMId 18e24.9 25e29.9 30e39.9 40 Clinical stage I II IIIA IIIB ER/PR status Negative Positive Allred score 2-6 7-8 HER2 Positive Negative Type of chemotherapy Anthracycline or taxane Anthracycline and taxane Chemotherapy cycles 6-8 <6 Breast density Low High a b c

adjuvant chemotherapy [12]. In a pooled analysis of data of 12 trials including almost 12,000 patients, achievement of pCR was associated with long term survival benefit in those with HER2-positive, triple negative breast cancer in addition to patients with luminal-B disease [13]. In addition, in patients who do not achieve pCR, lower pathological stage after neoadjuvant chemotherapy was associated with better long term survival [14]. In view of this, prediction of response to neoadjuvant chemotherapy is of paramount importance for proper selection of patients. In the present study, low BD assessed at diagnostic mammograms was associated with more likelihood of pCR and lower pathologic stage which were more pronounced after adjustment for BMI. Similarly, in a recent study conducted by the authors in metastatic breast cancer patients, low BD (<25%) was associated with better progression free survival [15]. Noteworthy, high mammographic BD was linked with four to six-fold risk of having breast cancer [4]. In addition, patients with high BD at diagnostic mammograms had increased risk of local recurrence after mastectomy [16]. In a recent study involving 267 patients, low BD was associated with improved disease-free survival in locally advanced patients treated with neoadjuvant chemotherapy [17]. However, Castaneda et al., 2014 [18] failed to show any correlation between mammographic BD and pCR. In that

d

OR, odds ratio. CI, confidence interval. LVI, lymphovascular invasion. BMI, body mass index (BMI data is available for 213 patients only).

study, pCR rate was only 15% and trastuzumab was not given to HER2-positive patients. In addition, they used Miller-Payne scale to assess pathological response to chemotherapy and defined pCR in the breast as grade 4 and 5 response where grade 4 refers to decreased cancer cellularity >90%. Suboptimal neoadjuvant chemotherapy, low pCR rate and definition of pCR including patients with residual tumour cells may account for lack of significant association of BD with pCR. The link of BD at prediagnostic mammograms with survival of breast cancer patients is less clear. In a Swedish trial, patients with high BD at screening mammograms had worse survival that was of borderline significance [19] which was not found in British and US cohorts [20,21]. BD is a modifiable trait affected by several genetic, hormonal, biological, and environmental factors that can change over time [22]. Prediagnostic images taken several years before actual diagnosis of cancer may not reflect the tumour microenvironment, biology and aggressiveness at the time of diagnosis. This can limit their utility as a prognostic factor and may explain the inconsistent association of BD with survival in trials that utilized screening mammograms to predict survival outcome.

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007

S. Elsamany et al. / The Breast xxx (2015) 1e6

5

Table 3 Distribution of post chemotherapy pathological stages according to different parameters. Parameters

Pathological stage 0

I

II

III

N (%)

N (%)

N (%)

N (%)

4(8.0%) 17(8.9%)

14(28.0%) 50(26.2%)

20(40.0%) 80(41.9%)

0.90

15(10.1%) 6(6.5%)

37(25.0%) 27(29.0%)

62(41.9%) 38(40.8%)

0.98

11(8.0%) 10(9.6%)

40(29.2%) 24(23.1%)

55(40.2%) 45(43.3%)

0.93

6(11.5%) 15(7.9%)

10(19.2%) 54(28.6%)

25(48.1%) 75(39.7%)

0.4

8(7.3%) 13(10.0%)

30(27.3%) 34(25.9%)

48(43.6%) 52(39.7%)

0.5

1(2.8%) 9(14.1%) 7(7.4%) 2(11.1%)

10(27.8%) 16(25.0)% 25(26.3%) 6(33.3%)

8(22.2%) 23(35.9%) 45(47.4%) 8(44.5%)

0.001

0 6(11.3%) 9(8.5%) 6(7.6%)

0 16(30.2%) 27(25.5%) 21(26.6%)

0 17(32.1%) 42(39.6%) 41(51.9%)

0.04

10(12.8%) 11(6.8%)

20(25.6%) 44(27.0%)

30(38.5%) 70(42.9%)

0.60

4(6.7%) 7(7.1%)

16(26.7%) 27(27.6%)

26(43.3%) 42(42.9%)

0.98

14(8.9%) 7(8.4%)

51(32.3%) 13(15.7%)

64(40.5%) 36(43.4%)

0.30

16(9.9%)

42(25.9%)

67(41.4%)

0.90

5(6.3%)

22(27.8%)

33(41.8%)

2(3.6%) 19(10.3%)

16(28.6%) 48(25.9%)

27(48.2%) 73(39.5%)

0.12

8(9.8%) 13(8.2%)

18(21.9%) 46(28.9%)

31(37.8%) 69(43.4%)

0.10

Age 40 12(24.0%) >40 44(23.0%) Menopause Premenopause 34(23.0%) Postmenopause 22(23.7%) Grade 1/2 31(22.6%) 3 25(24.0%) Multicentric tumours Yes 11(21.2%) No 45(23.8%) LVIa Yes 24(21.8%) No 32(24.4%) b BMI 18e24.9 17(47.2%) 25e29.9 16(25.0%) 30e39.9 18(18.9%)  40 2(11.1%) Clinical stage I 3(100%) II 14(26.4%) IIIA 28(26.4%) IIIB 11(13.9%) ER/PR status Negative 18(23.1%) Positive 38(23.3%) Allred score 2-6 14(23.3%) 7-8 22(22.4%) HER2 Negative 29(18.3%) Positive 27(32.5%) Chemotherapy type Anthracycline 37(22.8%) and taxane Anthracycline 19(24.1%) or taxane No. of cycles <6 11(19.6%) 6e8 45(24.3%) Breast density Low 25(30.5%) High 31(19.5%) a b

p

LVI, lymphovascular invasion. BMI, body mass index (BMI data is available for 213 patients only).

Recent reports have linked high BD with features of more aggressive disease like large, high grade, ER-negative tumours with positive lymph nodes, which point to the possible role of BD in tumour aggressiveness [23,24]. In addition, dense breast was associated with increased risk of all subtypes of breast cancer particularly ER-negative disease, which raises the need to include breast density in risk assessment of different tumour subtypes [23,24]. Furthermore, the interplay between the two components of radio-dense tissue, the glandular tissue and the stroma, was found to be involved in malignant initiation and progression [6]. In our opinion, BD can be viewed as the outcome of the underlying biological interactions involved in neoplastic process that reflects tumour behaviour and aggressiveness. In view of this concept, high BD may be considered as an adverse phenotype representing “a biologically bad disease”. Matching with previous reports [23,24], we could not detect any difference in the distribution of HER2-status between low and high

Fig. 2. Receiver operator characteristics (ROC) curve of factors predicting pathological complete response (pCR).

BD patients. However, ER-positive disease and those with high Allred score were more common among high BD patients which may refer to different tumour biology in our cohort. Noteworthy, BD is related to serum oestrogen levels [25] and increases with factors associated with augmented hormonal exposure such as early menarche and use of hormonal replacement therapy while it decreases with increasing age and menopause [26e28]. In addition, genes involved in hormone metabolism are suggested to play a role in BD determination [29]. The quantitative visual assessment of mammographic BD utilized in our study has some limitations. It lacks the accurate estimation of BD provided by computer-aided assessment in addition to the hazard of inter-observer variation in BD evaluation. However, this method has some advantages. It is a simple method that can be easily implemented in the routine practice without requiring special software and it classifies patients into easily interpretable groups. In our cohort, higher BMI was associated with worse response to neoadjuvant chemotherapy. This refers to the importance of keeping normal body weight to improve the outcome of therapy of breast cancer patients. This poor outcome is consistent with data reported by Dawood et al. [30] who displayed worse survival outcome in obese and over-weight patients with locally advanced breast cancer compared to normal/underweight counterparts. Similarly, in a recent meta-analysis including 213,075 patients from 82 studies, obesity was associated with poor overall and breast cancer survival in pre- and post-menopausal breast cancer patients [31]. The biological basis which links obesity with worse breast cancer survival is yet to be established. Several explanations have been proposed such as higher level of oestradiol produced in postmenopausal women through aromatisation of androgens in the adipose tissues [32]. Chemotherapy under-dosing in obese women, suboptimal treatment and obesity-related complications may be involved as well [33]. However, ROC curve analysis showed that BMI and probably BD may have fair predictive accuracy for pCR. In a heterogeneous disease like breast cancer, it seems that no single parameter will

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007

6

S. Elsamany et al. / The Breast xxx (2015) 1e6

provide accurate prediction of response to chemotherapy. Integration of different variables in predictive models is needed for proper estimation of chemotherapy response. Interestingly, in our cohort, pCR rate and pathologic stage did not differ according to the type of neoadjuvant chemotherapy or the number of cycles given. Compared to low BD patients, high BD patients received a higher number of anthracycline/taxan-based chemotherapy cycles. This data points to the importance to tumour biology rather than the type or intensity of chemotherapy as the main determinant of response to chemotherapy. In conclusion, our data suggests that low BD assessed at the time of diagnosis, early clinical stage and low BMI are associated with higher probability of achieving both pCR and lower pathological stage with neoadjuvant chemotherapy. To be validated, the predictive value of BD in this setting should be investigated in a larger cohort of patients. Disclosure None. Ethical approval The study has been approved by IRBs of the contributing institutions. Conflict of interest statement None declared. References [1] Cleator S, Parton M, Dowsett M. The biology of neoadjuvant chemotherapy for breast cancer. Endocr Relat Cancer 2002;9:183e95. [2] Aapro M. Neoadjuvant therapy in breast cancer: can we define its role? Oncologist 2011;6(Suppl. 3):36e9. [3] Tewari M, Krishnamurthy A, Shukla HS. Predictive markers of response to neoadjuvant chemotherapy in breast cancer. Surg Oncol 2008;17(4):301e11. [4] McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2006;15:1159e69. [5] Haars G, van Noord P, Gils C, Grobbee DE, Peeters P. Measurements of breast density: no ratio for a ratio. Cancer Epidemiol Biomarkers Prev 2005;14: 2634e40. [6] Boyd NF, Martin LJ, Yaffe MJ, Minkin S. Mammographic density and breast cancer risk: current understanding and future prospects. Breast Cancer Res 2011;13:223. [7] Maskarinec G, Pagano IS, Little MA, Conroy SM, Park SY, Kolonel LN. Mammographic density as a predictor of breast cancer survival: the Multiethnic Cohort. Breast Cancer Res 2013;15:R7. [8] Alowami S, Troup S, Al-Haddad S, Kirkpatrick L, Watson PH. Mammographic density is related to stroma and stromal proteoglycan expression. Breast Cancer Res 2003;5:R129e35. [9] Kass L, Erler JT, Dembo M, Weaver VM. Mammary epithelial cell: influence of extracellular matrix composition and organization during development and tumorigensis. Int J Biochem Cell Biol 2007;39:1987e94. [10] Butcher DT, Alliston T, Weaver VM. A tense situation: forcing tumour progression. Nat Rev Cancer 2009;9:108e22. [11] Wolfe JN. Risk for breast cancer development determined by mammographic parenchymal pattern. Cancer 1976;37:2486e92. [12] Mauri D, Pavlidis N, Ioannidis JP. Neoadjuvant versus adjuvant systemic treatment in breast cancer: a meta-analysis. J Natl Cancer Inst 2005;97(3): 188e94.

[13] Cortazar P, Zhang L, Untch M, Mehta K, Costantino JP, Wolmark N, et al. Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014;384(9938):164e72. [14] Kim MM, Allen P, Gonzalez-Angulo AM, Woodward WA, Meric-Bernstam F, Buzdar AU, et al. Pathologic complete response to neoadjuvant chemotherapy with trastuzumab predicts for improved survival in women with HER2overexpressing breast cancer. Ann Oncol 2013;24:1999e2004. [15] Elsamany S, Alzahrani A, Abd Elkhalik S, Elemam O, Rawah E, Farooq MU, et al. Prognostic value of mammographic breast density in patients with metastatic breast cancer. Med Oncol 2014;31(8):96. [16] Cil T, Fishell E, Hanna W, Sun P, Rawlinson E, Narod SA, et al. Mammographic density and the risk of breast cancer recurrence after breast-conserving surgery. Cancer 2009;115:5780e7. ~ eda C. Mamographic [17] Rojas KI, Flores R, Flores CJ, Pinto JA, Gomez HL, Castan density and disease-free survival in [HRþ, HER2] locally advanced breast cancer treated with neoadjuvant chemotherapy. J Clin Oncol 2014;32. Suppl. abstr e11536. [18] Castaneda CA, Flores R, Rojas K, Flores C, Castillo M, Milla E. Association between mammographic features and response to neoadjuvant chemotherapy in locally advanced breast carcinoma. Hematol Oncol Stem Cell Ther 2014;7(4):149e56. [19] Chiu SY, Duffy S, Yen AM, Tab ar L, Smith RA, Chen HH. Effect of baseline breast density on breast cancer incidence, stage, mortality, and screening parameters: 25-year follow-up of a Swedish mammographic screening. Cancer Epidemiol Biomarkers Prev 2010;19:1219e28. [20] Porter GJ, Evans AJ, Cornford EJ, Burrell HC, James JJ, Lee AH, et al. Influence of mammographic parenchymal pattern in screening-detected and interval invasive breast cancers on pathologic features, mammographic features, and patient survival. AJR Am J Roentgenol 2007;188:676e83. [21] Gierach GL, Ichikawa L, Kerlikowske K, Brinton LA, Farhat GN, Vacek PM, et al. Relationship between mammographic density and breast cancer death in the breast cancer surveillance consortium. J Natl Cancer Inst 2012;104:1218e27. [22] Boyd NF, Martin LJ, Yaffe MJ, Minkin S. Mammographic density: a hormonally responsive risk factor for breast cancer. J Br Menopause Soc 2006;12(4): 186e93. [23] Yaghjyan L, Colditz GA, Collins LC, Schnitt SJ, Rosner B, Vachon C, et al. Mammographic breast density and subsequent risk of breast cancer in postmenopausal women according to tumor characteristics. J Natl Cancer Inst 2011;103(15):1179e89. [24] Bertrand KA, Tamimi RM, Scott CG, Jensen MR, Pankratz V, Visscher D, et al. Mammographic density and risk of breast cancer by age and tumor characteristics. Breast Cancer Res 2013;15:R104. [25] Ursin G, Qureshi SA. Mammographic density e a useful biomarker for breast cancer risk in epidemiologic studies. Norsk Epidemiologi 2009;19(1):59e68. [26] El-Bastawissi AY, White E, Mandelson MT, Taplin SH. Reproductive and hormonal factors associated with mammographic breast density by age (United States). Cancer Causes Control 2000;11:955e63. [27] Titus-Ernstoff L, Tosteson AN, Kasales C, Weiss J, Goodrich M, Hatch EE, et al. Breast cancer risk factors in relation to breast density (United States). Cancer Causes Control 2006;17:1281e90. [28] Warren R. Hormones and mammographic breast density. Maturitas 2004;49: 67e78. [29] Diorio C, Brisson J, Berube S, Pollak M. Genetic polymorphisms involved in insulin-like growth factor (IGF) pathway in relation to mammographic breast density and IGF levels. Cancer Epidemiol Biomarkers Prev 2008;17:880e8. [30] Dawood S, Broglio K, Gonzalez-Angulo AM, Kau SW, Islam R, Hortobagyi GN, et al. Prognostic value of body mass index in locally advanced breast cancer. Clin Cancer Res 2008;14:1718e25. [31] Chan DSM, Vieira AR, Aune D, Bandera EV, Greenwood DC, McTiernan A, et al. Body mass index and survival in women with breast cancer-systematic literature review and meta-analysis of 82 follow-up studies. Ann Oncol 2014;25:1901e14. [32] Goodwin PJ, Ennis M, Bahl M, Fantus IG, Pritchard KI, Trudeau ME, et al. High insulin levels in newly diagnosed breast cancer patients reflect underlying insulin resistance and are associated with components of the insulin resistance syndrome. Breast Cancer Res Treat 2009;114:517e25. [33] Greenman CG, Jagielski CH, Griggs JJ. Breast cancer adjuvant chemotherapy dosing in obese patients: dissemination of information from clinical trials to clinical practice. Cancer 2008;112:2159e65.

Please cite this article in press as: Elsamany S, et al., Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients, The Breast (2015), http://dx.doi.org/10.1016/j.breast.2015.05.007