- Email: [email protected]
Evaluation of Discordance in Primary Tumor and Lymph Node Response After Neoadjuvant Therapy in Breast Cancer
Original Study
Evaluation of Discordance in Primary Tumor and Lymph Node Response After Neoadjuvant Therapy in Breast Cancer Christina A. Fleming,1 Karen McCarthy,1 Ciara Ryan,2 Aoife McCarthy,2 Seamus O’Reilly,1 Deirdre O’Mahony,1 Tara Jane Browne,2 Paul Redmond,1 Mark A. Corrigan1 Abstract Primary tumor and lymph node (LN) response discordance may be observed in breast cancer patients receiving neoadjuvant therapy. We quantified the rate of this discordance and identified factors predictive of it. Almost 15% of primary tumors with complete pathologic response had persistently positive LNs. This is concerning, as we may be undertreating a significant portion of patients. Background: Neoadjuvant therapy (NAT) offers a unique opportunity to assess tumor response to systemic agents. However, a discrepancy may exist between the response of the primary tumor and involved nodes. We report on the frequency of response discordance after NAT in breast cancer. Patients and Methods: All consecutive node-positive patients receiving NAT in our department from 2009 to 2014 were identified. Patient demographics, and radiologic and pathologic features were tabulated. Tumor response was estimated by magnetic resonance imaging of the breast. Lymph node (LN) response was estimated from pathologic treatment response measurements. Statistical analysis was performed. Results: A total of 108 node-positive patients treated with NAT were eligible for inclusion. Median age was 51.73 years (range, 20-87 years). All patients underwent axillary clearance, and 62% underwent mastectomy. A 40% mean reduction in tumor size was observed. Statistically, a positive correlation between tumor and LN response after NAT was observed (Spearman correlation coefficient, r ¼ 0.46, P < .001). Complete pathologic response was observed in 17 patients (15.7%). However, 21 patients experienced complete LN response, with only 81% of these patients (n ¼ 17) experiencing a complete response in tumor also. A complete response was observed in tumor in 20 patients, and this predicted complete nodal response in 85% of cases (n ¼ 17). Fifteen percent of primary tumors with complete pathologic response had persistently positive LNs. Conclusion: A significant discordance exists between the primary tumor and LN response, representing a concern for the lack of response of occult regional or systemic metastases due to potential biologic heterogeneity. Clinical Breast Cancer, Vol. -, No. -, --- ª 2017 Elsevier Inc. All rights reserved. Keywords: Axillary surgery, Biogenetics, Breast surgery, Neoadjuvant therapy, Response discordance
Introduction The role of neoadjuvant therapy (NAT) in the management of breast cancer is evolving.1 Use of NAT is dependent on tumor size, receptor status, and nodal status, and it utilizes chemotherapy or hormone-based treatment. Historically, the primary aims of NAT 1
Breast Research Centre Department of Histopathology, Cork University Hospital, Cork, Ireland
2
Submitted: Oct 15, 2017; Revised: Nov 20, 2017; Accepted: Nov 21, 2017 Address for correspondence: Christina A. Fleming, MB, MCh, MRCSI, Department of Academic Surgery, Cork University Hospital, Wilton, Cork, Ireland E-mail contact: christina.fl[email protected]
1526-8209/$ - see frontmatter ª 2017 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.clbc.2017.11.016
were to downsize inoperable tumors, thus allowing surgical resection, and to decrease operable tumor size, ultimately allowing higher rates of breast-conserving surgery (BCS).2,3 However, NAT has also been reported to reduce the risk of distant recurrence by targeting circulating micrometastases.4 Furthermore, NAT offers a unique opportunity to assess tumor response to systemic agents, which may be described as a human model system to explore the efficacy of therapies and response to therapy, and can also identify specific patient cohorts at higher risk of future disease relapse.5,6 Currently, axillary lymph node (LN) status is still considered a significant prognostic factor for overall outcomes in primary breast cancer.7 Additionally, axillary LN disease burden after completion of NAT
Clinical Breast Cancer Month 2017
-1
Discordance in Tumor and LN Response is considered to be a strong predictor of subsequent disease relapse.8 The aim of treatment is therefore to synchronously improve primary tumor and LN tumor cell burden. It has been estimated that up to 40% of node-positive disease reverts to node-negative status after NAT.9 Complete pathologic response (ypCR) rates as high as 67% can be achieved in specific breast cancer groups, most notably human epidermal growth factor receptor 2 (HER2)-positive tumors.10 Some reports have associated overall ypCR with a survival benefit.11,12 It is emerging that a discrepancy may exist between the response of the primary tumor and involved nodes after NAT. A differential response is difficult to explain but perhaps reflects biomolecular heterogeneity in the tumor observed in primary and secondary metastatic sites after lymphatic spread or circulation of micrometastases.13 Performing sentinel LN biopsy after NAT has been suggested to assess LN response to NAT independent of tumor response and thus identify patients who could avoid the concurrent morbidity of axillary lymphadenectomy.14 However, the ACOSOG Z1071 study has concluded that the sensitivity of sentinel LN biopsy in this setting is insufficient to change current practice.15 We sought to assess the overall frequency of discordance in tumor and LN response after NAT in breast cancer, regardless of tumor type. We further aimed to identify tumor factors associated with nonresponse or reduced response in LN when response, even ypCR, is observed in primary tumor.
Tumor response to NAT was calculated on the basis of size by subtraction of the largest dimension in size of tumor bed calculated on histologic analysis of the surgical specimen (ypT) from that calculated on pre-NAT magnetic resonance imaging (MRI) of the breast (cT) and a percentage response calculated, as previously validated.20 Where MRI was unavailable or not performed, ultrasound, mammography, and computed tomography (CT) were used, in this sequence. Tumor size was noted as the maximal dimension when coronal, sagittal, and axial sections were measured. LN response to therapy was calculated by expression of number of LNs collected with evidence of viable tumor and/or treatment effect (fibrosis) or response to therapy compared to the entire number of LNs collected after axillary lymphadenectomy.
Statistical Analysis and Ethical Approval
All patients at the Southern Breast Cancer Centre at Cork University Hospital who had node-positive disease at diagnosis and went on to receive NAT (chemotherapy and/or hormone therapy) over a 5-year inclusive period from 2009 to 2014 were identified using a prospectively maintained database. For inclusion, patients were required to have undergone surgery during this time period that included level II axillary lymphadenectomy to allow for full histologic analysis of LN response. The following parameters were recorded: patient demographics including age and gender, surgical procedure including requirement for reexcision of margins or completion mastectomy, and whether NAT was administered as chemotherapy or as a hormonal agent.
Histopathologic Analysis
Results
Data on tumor and LN status before NAT and postoperatively after completion of NAT and subsequent surgery were recorded. All breast cancer specimens were analyzed by breast-specialist consultant histopathologists using a standardized approach. All cases were processed similarly in a single laboratory with uniform preanalytics as per the American Society of Clinical Oncology/College of American Pathologists guidelines.16 Equivocal HER2 positivity was further analyzed by fluorescence in-situ hybridization (FISH) or bright-field dual-color in-situ hybridization (BDISH), and repeat HER2 testing on grade III cancers was also performed as per current guidelines.16,17 Pre- and postoperative predictive and prognostic parameters including tumor histologic subtype, grade, presence or absence of lymphovascular invasion, and hormone and HER2 receptor status were reported, and tumors were staged according to American Joint Committee on Cancer (AJCC) 7th edition
Demographics and Treatment
Patients and Study Design
-
Calculation of Tumor and LN Response to NAT
A prior sample size calculation indicated that a sample of 34 patients was necessary to detect a medium effect size (Cohen’s f ¼ 0.50) in a paired t test comparing tumor response and LN response with a power of 80% and a level of significance of .05.21 To account for lack of normality in distribution, the sample size was increased by 15%, and a sample of 40 patients was therefore necessary.22 All other statistical analysis was performed by SPSS 22 (IBM SPSS, Chicago, IL). Descriptive analyses were performed on demographic information. Correlation of tumor and LN response to NAT was calculated by Spearman’s test with normality not assumed, and strength of correlation was calculated on the basis of standard methods.23 Discordance was calculated on the basis of percentage subtraction. Univariate analysis was performed to identify factors predictive of response discordance, and factors achieving P < .25 were included for multivariate analyses.24 Multivariate analyses were performed on a logistic regression model to identify tumor features predictive of discordance response in LN disease. Ethical approval was prospectively obtained from the Clinical Research and Ethics Committee of the Cork Teaching Hospitals.
Methods
2
guidelines and the College of American Pathologists.17,18 The MD Anderson Residual Cancer Burden Protocol was used to analyze all residual disease in breast specimens after surgery.19
Clinical Breast Cancer Month 2017
A total of 3275 breast cancer patients (screen detected and symptomatic) were treated at our institute between 2009 and 2014. Of these, 203 patients were treated with NAT; of these, 108 patients proceeded to level II axillary lymphadenectomy for nodepositive disease and were eligible for study inclusion. All patients included were women; median age was 51.73 years (range, 20-87 years). An overall mastectomy rate of 64.8% was observed (62% at first operation and a further 2.8% after initial BCS). Reexcision of margins was performed in 10.2% of cases.
Tumor Characteristics, and Nodal and Metastatic Disease Table 1 lists the tumor and nodal characteristics at diagnosis and the quantification of overall response to NAT using the MD Anderson Residual Cancer Burden (RCB) score. Overall, 86.1% of
Christina A. Fleming et al Table 1 Tumor and Nodal Characteristics at Diagnosis and Quantification of Overall Response to NACT Using RCB Score Characteristic
Value
At Diagnosis Tumor Type IDC
93 (86.1)
ILC
8 (7.4)
Mixed
7 (6.5)
Tumor Grade Low (I þ II)
44 (40.7)
High (III)
64 (59.3)
Tumor Size (mm) Mean
47.9
SD
21.57
Range
4-100
Receptor Status ERþPRþ
60 (55.6)
HER2þ
52 (48.1)
Triple negative
16 (14.8)
LVI 28 (25.9)
Absent
80 (74.1)
Lymph Node Status Positive
106 (98.1)
Negative
2 (1.9)
After NACT RCB Score SD Range
2.68 1.53 0-5.45
RCB Class ypCR
Response to NAT Tumor size on pre-NAT MRI was available in 54.6% of cases (n ¼ 59). Where MRI was unavailable or was not performed, ultrasound (n ¼ 20), mammography (n ¼ 28), and CT (n ¼ 1) were used. Accuracy of these imaging modalities for tumor size estimation has been described previously.25,26 Overall, 80.6% of tumors exhibited some response to NAT (mean SD response, 40.53 56.7% reduction in size) (Table 2). Conversion rate from mastectomy to BCS based on tumor size alone was 16.7% (tumors measuring 5 cm on pre-NAT MRI that reduced in size and proceeded to BCS). ypCR was seen in 15.7% (n ¼ 17) of patients (ie, combined ypCR in tumor and LNs). However, a total of 18.5% of tumors (n ¼ 20) underwent complete response after NAT, while 14.8% of tumors (n ¼ 16) did not respond or were found to be larger than originally estimated. Evidence of nodal disease regression was seen in 60.2% (n ¼ 65) and 19.4% (n ¼ 21) of patients who experienced ypCR (Table 2).
Tumor
Present
Mean
diagnosis (n ¼ 5 bone). Only 2 of these patients had lymphovascular invasion at initial core biopsy. All 5 had lymphovascular invasion at final specimen analysis.
17 (15.7)
I
4 (3.7)
II
39 (36.1)
III
48 (44.5)
Data are presented as n (%) unless otherwise indicated. Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; IDC ¼ invasive ductal carcinoma; ILC ¼ invasive lobular carcinoma; LVI ¼ lymphovascular invasion; NACT ¼ neoadjuvant chemotherapy; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden; ypCR ¼ pathologic complete response.
patients (n ¼ 93) had invasive ductal carcinoma (IDC) at initial biopsy, and 7.4% (n ¼ 8) had invasive lobular carcinoma. At initial biopsy, 59.3% of tumors were deemed grade III. This figure dropped to 35.2% after NAT (n ¼ 4 grade I, n ¼ 46 grade II) and was statistically significant (P < .001). The mean pre-NAT tumor size was 47.9 mm (range, 4-100 mm). Postoperatively, based on histologic assessment, this value reduced to 28.66 mm (range, 0-150 mm), reflecting overall positive tumor response to treatment (P < .001). The increased range after NAT may be explained by either underestimation of tumor size on pre-NAT imaging or by tumor progression, which was observed in 14.8% of cases (n ¼ 16). A mean collection of 22.08 LNs was achieved (range, 4-52 LNs; SD 9.184). Five patients (4.6%) had distant metastases at the time of
LN Response Correlation. An overall positive correlation was observed between tumor and LN response after NAT (Spearman correlation coefficient, r ¼ 0.461, P < .001). The strength of correlation increases (r ¼ 0.594, P < .001) when LNs with a zeroresponse rate are excluded. This may be explained by a tumor progression rate of 23.3% in this group and potential progression of nodal disease that is not fully quantifiable preoperatively. Seventeen patients (15.7%) experienced combined ypCR in tumor and LN disease (correlation coefficient, r ¼ 1). Tumor characteristics associated with greatest discordance in tumor and LN response are outlined in Table 3. Tumors showing the strongest correlation in response included invasive lobular carcinoma measuring > 5 cm with a triple-negative hormone profile (Table 3). LN Response Discordance. A mean discordance in response of tumor and LN disease of 36.9% was observed after NAT (Table 4). Most discordant tumors (> 50% discordance) were triple positive, followed by HER2 positive regardless of estrogen receptor (ER) status and IDC tumors measuring 2 to 5 cm before NAT
Table 2 Independent Tumor Response and Lymph Node Response to Neoadjuvant Therapy Characteristic ypCR Response No response Disease progression
Tumor Responsea 20 87 5 16
(18.5) (80.6) (4.6) (14.8)
Lymph Node Response 21 (19.4) 65 (60.2) 43 (39.8) —b
Abbreviation: ypCR ¼ pathologic complete response. a Overall number of tumors that responded to therapy and those with ypCR. b Progression is nonquantifiable as a result of the inability to perform complete lymph node assessment preoperatively. Progression may be confounded by initial underestimation of tumor burden on imaging thanks to the utility of differing modalities.
Clinical Breast Cancer Month 2017
-3
Discordance in Tumor and LN Response Table 3 Tumor and Lymph Node Percentage Response and Discordance Value by Subgroup With Correlations Performed Tumor Response (%), Mean (SD)
Lymph Node Response, % Mean (SD)
Differential Response, % (TLNR)
Pa
Correlation Coefficient, R (P)
Invasive ductal carcinoma
42.28 (56.86)
37.56 (38.64)
4.72 (0.89)
.43
0.33 (.001)
Invasive lobular carcinoma
30.63 (61.37)
30.25 (38.63)
0.38 (0.99)
.98
0.73 (.41)
Mixed
28.57 (55.03)
42.86 (53.45)
14.29 (0.67)
.22
0.87 (.01)
2-5 cm
38.76 (58.29)
37.47 (39.2)
1.29 (0.97)
.86
0.31 (.01)
>5 cm
43.31 (54.7)
37.19 (40.03)
6.12 (0.86)
.40
0.55 (<.001)
22.44 (57.97)
26.63 (31.56)
4.19 (0.84)
.63
0.30 (.049)
53 (53.23)
44.38 (42.82)
8.62 (0.84)
.20
0.40 (.001)
Characteristic Tumor Type
Size
Grade Low (I þ II) High (III) LVI Present
61.07 (54.5)
44.43 (46.07)
16.64 (0.73)
.10
0.48 (.01)
Absent
33.34 (56.017)
34.89 (36.70)
1.55 (0.96)
.802
0.351 (.001)
ERþPRþ
25.35 (59.96)
31.83 (34.76)
6.48 (0.79)
.38
0.39 (.002)
HER2þ
51.77 (51)
47.25 (41.56)
4.52 (0.91)
.57
0.26 (.07)
HER2
30.09 (60.11)
28.18 (35.08)
1.91 (0.94)
.79
0.48 (<.001)
Triple negative
38.58 (60.76)
25.17 (38.55)
13.41 (0.65)
.39
0.54 (.07)
Receptor Status
RCB Class 100
100 (1)
—
—
I
94.50 (9.71)
56.25 (47.95)
38.25 (0.60)
.251
0.32 (.684)
II
45.92 (51.85)
39.97 (36.27)
5.95 (0.87)
.56
0.27 (.873)
III
8.49 (50.99)
13.85 (19.09)
5.36 (0.61)
.488
0.532 (.094)
ypCR
100
Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor; LVI ¼ lymphovascular invasion; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden; TLNR ¼ tumorelymph node response ratio. a Statistically significant at P < .05.
(35%-40%). Larger (> 5 cm) triple-negative tumors showed the least discordance in response (< 35%). Table 4 lists factors predictive of tumor and LN discordance on univariate analysis. Triple-positive disease (P ¼ .002), IDCs (P ¼ .045), and disease measuring < 5 cm (P ¼ .021) remained significantly associated with differential response on multivariate analysis.
Discussion
4
-
We identified that a discordance in tumor and LN response to NAT in breast cancer exists. This observation is consistent with suggested hypotheses in other studies.27-30 We have further identified particular tumor characteristics that are associated with a differential response (IDCs, < 5 cm size, triple-positive receptor profile). We also identified that even tumors with ypCR will not always have ypCR in nodal disease. Although there is a significant association between ypCR in tumor and nodal disease, an absolute correlation between tumor and LN response was not observed. Differential response in primary tumor and LN disease burden in breast cancer may be due to biomolecular differences in the tumor microenvironment of the primary tumor site and secondary metastatic foci or tumor heterogeneity.13,27 Central to a differential response is an observed discordance in hormone (ER and
Clinical Breast Cancer Month 2017
progesterone receptor) and HER2 receptor status between primary tumor and metastatic sites.27,30 Although this phenomenon has been reported previously, in practice, receptor testing is still reserved for primary tumor tissue assessment only and not metastatic foci (including LNs) before NAT planning. Hormone therapy and HER2-targeted therapy are central to the current neoadjuvant management of breast cancer.31-33 The purpose of NAT is combined reduction in primary tumor size and a reduction of neoplastic disease in metastatic foci.2-4 However, differential receptor status can lead to therapeutic discrepancy. It is suggested that change in hormone receptor status may occur between primary tumor and LN disease sites, which may explain response in primary site and not in nodal disease and vice versa.34,35 Further hypotheses to explain discordance in hormone receptor status between primary tumor and metastatic sites have been provided.36-44 It is unknown if heterogeneity of receptor status is a true biologic finding or if it is due to reporting discrepancies. Tissue fixation and test type (most commonly immunohistochemistry) have been associated with technical discrepancies in studies examining receptor status in primary and metastatic sites.36 Thresholds for defining hormone receptor positivity may also be involved. Currently, international guidelines used in practice require low-level
Table 4 Factors Associated With Discordance in Tumor and Lymph Node Response to Neoadjuvant Therapya
Characteristic Overall
Response Discordance (%), Mean (SD)
Discordance Univariate Analysis Response Discordance P
Discordance Multivariate Analysis
OR
95% CI
P
OR
95% CI
P
1.36
0.33-1.79
.021
1.18
0.15-3.50
.045
1.38
0.26-1.62
.66
1.59
0.47-5.72
.44
36.96 (40.27)
Tumor Size 2-5 cm
39.42 (43.52)
<.001
1.88
0.821-2.01
<.001
>5 cm
32.95 (34.44)
.337
1.19
0.75-1.90
.55
IDC
38.66 (42.16)
<.001
1.58
0.50-4.98
.02
ILC
32.63 (24.37)
.525
1.15
0.46-2.89
.76
Mixed
15.50 (22.78)
.048
1.53
0.463-5.023
.44
Low (I þ II)
38.93 (41.44)
<.001
1.49
0.59-2.84
.06
High (III)
35.75 (40.05)
.622
1.16
0.75-1.80
.32
ERþPRþ
37.30 (42.02)
.250
1.92
0.068-2.1
.08
HER2þ
38.02 (42.33)
.037
1.45
0.46-2.11
.56
Triple positive (n ¼ 10)
51.20 (66.30)
<.001
2.1
1.63-3.35
<.001
1.86
0.17-4.96
.002
Triple negative (n ¼ 12)
34.75 (39.10)
.24.7
1.31
0.57-3.02
.07
1.01
0.40-12.91
.35
I
56.25 (41.46)
.331
1.55
0.79-1.95
.06
1.34
0.428-2.30
.141
II
35.29 (44.65)
.752
1.38
0.83-2.29
.14
1.25
0.43-2.98
.257
III
41.06 (40.41)
.355
1.85
0.85-4.01
.041
1.306
0.923-1.85
.086
Tumor Type
Grade
Receptor Status
Clinical Breast Cancer Month 2017
Abbreviations: CI ¼ confidence interval; ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; IDC ¼ invasive ductal carcinoma; ILC ¼ invasive lobular carcinoma; OR ¼ odds ratio; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden. a Discordance for univariate and multivariate analysis is defined as difference in tumor and lymph node response of > 30%.
Christina A. Fleming et al
RCB Score
-5
Discordance in Tumor and LN Response
6
-
positivity for receptor positivity status on immunohistochemistry; only 1% positive tumor nuclei identified on immunostaining is required for a positive finding for ER and progesterone receptor status, with equivocal HER2 positivity further assessed using FISH or BDISH.15,16 However, it is also accepted that significant intratumor heterogeneity can exist, so receptor positivity may not be representative of the entire tumor burden.37,38 HER2 heterogeneity is reported to occur in 5% to 40% of breast cancers, depending on the definition used.45 This may potentially result in a false-negative HER2 report on core needle biopsy, although this is uncommon.46 Tumor evolution is also driven by the diversity of cancer cell populations and their microenvironment, and changes in hormone receptor status may occur during the metastatic process.39 Loss of hormone receptor positivity after circulation of micrometastases and metastatic disease has been reported. One explanation for this may be clonal selection during the metastatic process.40 As already described, significant tumor heterogeneity and polyclonality may occur in as many as 70% of cases by molecular analysis.37,41 At initiation of the metastatic process, differential clones are balanced toward those with a preferential potential for metastasing.42 Therefore, a small neoplastic focus within the heterogeneous tumor bed may metastasize with genotypic and phenotypic differences from that predominant in the primary tumor.41 Other biomolecular mechanisms that may be involved include genetic miscoding during tumor progression, discriminatory effect of prior chemotherapy or hormone therapy, and selection of resistant tumor clones.38,43,44 With increasing use of NAT in breast cancer, the number of patients experiencing complete response has been more widely observed.11 Accurate diagnosis of ypCR before surgery may allow avoidance of unnecessary surgery as well as avoidance of unnecessary surgical morbidity and mortality. Preoperative MRI after NAT has significant accuracy for diagnosing tumor ypCR.47 Positron emission tomographic scanning has been shown to be superior for diagnosing ypCR in both tumor bed and nodal distribution, though it is not widely utilized.48 Tumor bed size is certainly measurable via preoperative imaging; however, nodal disease is not fully quantifiable. Sentinel LN biopsy after completion of NAT has in the past been discouraged as a result of its association with significantly high false-negative rates, as observed in the SENTINA trial and in the ACOSOG Z1071 study findings.14,49 Although an association between tumor and LN ypCR was identified in our study, an absolute correlation was not observed; thus, identifying ypCR of primary tumor after NAT MRI cannot sufficiently predict ypCR in nodal disease. While clear attention to study design was paid, there were a number of limitations to this study. This is a retrospective account of tumor and LN response. Histopathologic assessment of breast cancer specimens at our institute is standardized and is performed in a uniform manner for each patient, in keeping with AJCC 7th edition guidelines.18 Reporting is also standardized. The patient cohort may be considered heterogeneous as a result of inclusion of patients who received neoadjuvant chemotherapy and/or hormone therapy, and different biologic subtypes were also included. For example, small cohorts of lobular carcinoma patients were included, and only one patient had a grade I tumor. Of note, we do not routinely evaluate the proliferative marker Ki-67 in our pathology
Clinical Breast Cancer Month 2017
laboratory. A low conversion rate from mastectomy to BCS were also noted. However, this was based on an analysis of tumor size alone without consideration of other factors, such as patient preference. It should also be noted that only 38.1% of these whose disease had no response or those who experienced tumor progression underwent pre-NAT MRI, and initial tumor burden was based on ultrasound (n ¼ 4, 19%), mammography (n ¼ 8, 38.1%), and CT (n ¼ 1, 4.8%). Therefore, assessment of progression may be subject to bias in this smaller subgroup. All patients underwent clinical assessment midway through NAT to identify progression of disease.
Conclusion We found that a discordance in response may exist between primary tumor and LN disease after NAT in breast cancer, and that ypCR of primary tumor is not always associated with complete nodal response. This discordance may be explained by intratumor heterogeneity and alterations in phenotypic receptor expression after lymphatic and hematogenous circulation, creating a scenario of reduced therapeutic responsiveness. This theory is supported by a 40% nonresponse in nodal disease. In response discordance, it may be reasonably inferred that sites of metastases are sampled and fully profiled in significantly discordant groups to allow for fully targeted, patient-tailored NAT at both the primary and metastatic tumor level. This is important because nodal response to NAT has prognostic value, independent of tumor response.50 Further investigation of the concept of tumor and LN response discordance can continue to enhance our ongoing clinical knowledge of this disease.
Clinical Practice Points Anecdotally, it has been reported that patients may not experi-
ence the same degree of response in their primary tumor and sites of LN disease after NAT for breast cancer. This may be explained by biomolecular or even genetic heterogeneity in primary tumor and metastatic disease sites. However, to date, this phenomenon has not been adequately quantified and reported. Furthermore, factors that predict this discordance have not been identified. We found that 15% of breast cancer patients who experience ypCR in their breast primary tumor after NAT have persistent LN disease, and 1 in 10 shows no response in nodal disease. We found that IDCs < 5 cm in size with a triple-positive receptor profile had the highest rate of response discordance. This finding is clinically relevant, as nodal response to NAT has prognostic value independent of tumor response. In response to this discordance, it may be reasonable to sample and fully profile LN metastases in markedly discordant groups to allow for improved targeted patient-tailored NAT at both the primary and metastatic tumor levels in this era of precision medicine.
Acknowledgments We acknowledge and thank all staff members at the Southern Breast Cancer Centre, Cork University Hospital, who contributed greatly to the completion of this work.
Christina A. Fleming et al Disclosure The authors have stated that they have no conflict of interest.
References 1. Rastogi P, Anderson SJ, Bear HD, et al. Preoperative chemotherapy: updates of National Surgical Adjuvant Breast and Bowel Project protocols B-18 and B-27. J Clin Oncol 2008; 26:778-85. 2. Peintinger F, Symmans WF, Gonzalez-Angulo AM, et al. The safety of breastconserving surgery in patients who achieve complete pathological response after neoadjuvant chemotherapy. Cancer 2006; 107:1248-54. 3. Wang S, Zhang Y, Yang X, et al. Shrink pattern of breast cancer after neoadjuvant chemotherapy and its correlation with clinical pathological factors. World J Surg Oncol 2013; 11:166. 4. Kasimir-Bauer S, Bittner AK, König L, et al. Does primary neoadjuvant systemic therapy eradicate minimal residual disease? Analysis of disseminated and circulating tumor cells before and after therapy. Breast Cancer Res 2016; 18:20. 5. Schwartz GF, Hortobagyi GN. Proceedings of the consensus conference on neoadjuvant chemotherapy in carcinoma of the breast, April 26-28, 2003, Philadelphia, Pennsylvania. Cancer 2004; 100:2512-32. 6. Yadav BS, Sharma SC, Singh R, Singh G. Patterns of relapse in locally advanced breast cancer treated with neoadjuvant chemotherapy followed by surgery and radiotherapy. J Cancer Res Ther 2007; 3:75-80. 7. Wu SG, Sun JY, Zhou J, et al. Number of negative lymph nodes can predict survival of breast cancer patients with four or more positive lymph nodes after postmastectomy radiotherapy. Radiat Oncol 2014; 9:284. 8. Escobare PF, Patrick RJ, Rybicki LA, Hicks D, Weng DE, Crowe JP. Prognostic significance of residual breast cancer and axillary node involvement for patients who had primary induction chemotherapy for advanced breast cancer. Ann Surg Oncol 2006; 13:783-7. 9. Thompson AM, Moulder-Thompson SL. Neoadjuvant treatment of breast cancer. Ann Oncol 2012; 23(suppl 10):231-6. 10. Buzdar AU, Ibrahim NK, Francis D, et al. Significantly higher pathological complete remission rate after neoadjuvant therapy with trastuzumab, paclitaxel and epirubicin chemotherapy: results of a randomized trial in human epidermal growth factor receptor 2epositive operable breast cancer. J Clin Oncol 2005; 23:3676-85. 11. Wang-Lopez Q, Chalabi N, Abrial C, et al. Can complete pathological response (pCR) be used as a surrogate marker of survival after neoadjuvant therapy for breast cancer? Crit Rev Oncol Hematol 2015; 95:88-104. 12. Chollet P, Amat S, Cure H, et al. Prognostic significance of a complete pathological response after induction chemotherapy in operable breast cancer. Br J Cancer 2002; 86:1041-6. 13. Symmans WF, Liu J, Knowles DM, Inghirami G. Breast cancer heterogeneity: evaluation of clonality in primary and metastatic lesions. Hum Pathol 1995; 26: 210-6. 14. Kuehn T, Bauerfeind I, Fehm T, et al. Sentinel-lymph-node biopsy in patients with breast cancer before and after neoadjuvant chemotherapy (SENTINA): a prospective, multicenter cohort study. Lancet Oncol 2013; 14:609. 15. Boughey JC, Suman VJ, Mittendorf EA, et al. Sentinel lymph node surgery after neoadjuvant chemotherapy in patients with node-positive breast cancer: the ACOSOG Z1071 (Alliance) clinical trial. JAMA 2013; 310:1455-61. 16. Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007; 25:118-45. 17. Nitta H, Hauss-Wegrzyniak B, Lehrkamp M, et al. Development of automated bright field double in situ hybridization (BDISH) application for HER2 gene and chromosome 17 centromere (CEN 17) for breast carcinomas and an assay performance comparison to manual dual color HER2 fluorescence in situ hybridization (FISH). Diagn Pathol 2008; 3:41-53. 18. Edge S, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. American Joint Committee on Cancer (AJCC) cancer staging manual. 7th ed. New York: Springer-Verlag; 2011. 19. Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007; 25: 4414-22. 20. Lee HJ, Song IH, Seo AN, et al. Correlations between molecular subtypes and pathologic response patterns of breast cancers after neoadjuvant chemotherapy. Ann Surg Oncol 2015; 22:392-400. 21. Cohen J. A power primer. Psychol Bull 1992; 112:155-9. 22. Erdfelder E, Faul F, Buchner A. GPOWER: a general power analysis program. Behav Res Methods 1996; 28:1-11.
23. Croux C, Dehon C. Influence functions of the Spearman and Kendall correlation measures. Stat Methods Appl 2012; 19:497-515. 24. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159-74. 25. Fischer U, Kopka L, Grabbe E. Breast carcinoma: effect of preoperative contrastenhanced MR imaging on the therapeutic approach. Radiology 1999; 213:881. 26. Yeh E, Slanetz P, Kopans DB, et al. Prospective comparison of mammography, sonography, and MRI in patients undergoing neoadjuvant chemotherapy for palpable breast cancer. AJR Am J Roentgenol 2005; 184:868. 27. Arslan C, Sari E, Aksoy S, Altundag K. Variation in hormone receptor and HER-2 status between primary and metastatic breast cancer: review of the literature. Expert Opin Ther Targets 2011; 15:21-30. 28. Polyak K. Heterogeneity in breast cancer. J Clin Invest 2015; 121:3786-8. 29. Holdaway IM, Bowditch JV. Variation in receptor status between in primary and metastatic breast cancer. Cancer 1983; 5:479-85. 30. Sari E, Guler G, Hayran M, Gullu I, Altundag K, Ozisik Y. Comparative study of the immunohistochemical detection of hormone receptor status and HER-2 expression in primary and paired recurrent/metastatic lesions of patients with breast cancer. Med Oncol 2011; 28:57-63. 31. Cortazar P, Zhang L, Untch M, et al. Pathological complete response and longterm clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014; 384:164. 32. Carey LA, Winer EP. Defining success in neoadjuvant breast cancer trials. Lancet 2014; 384:115. 33. Mieog JS, van der Hage JA, van de Velde CJ. Preoperative chemotherapy for women with operable breast cancer. Cochrane Database Syst Rev 2007:CD005002. 34. Lower EE, Glass EL, Bradley DA, et al. Impact of metastatic estrogen receptor and progesterone receptor status on survival. Breast Cancer Res Treat 2005; 90:65-70. 35. Lower EE, Glass E, Blau R, Harman S. HER-2/neu expression in primary and metastatic breast cancer. Breast Cancer Res Treat 2009; 113:301-6. 36. Gown AM. Current issues in ER and HER2 testing by HC in breast cancer. Mod Pathol 2008; 21(suppl 2):S8-15. 37. Teixeira MR, Pandis N, Bardi G, et al. Karyotypic comparisons of multiple tumorous and macroscopically normal surrounding tissue samples from patients with breast cancer. Cancer Res 1996; 56:855-9. 38. Rasbridge SA, Gillett CE, Seymour AM, et al. The effects of chemotherapy on morphology, cellular proliferation, apoptosis and oncoprotein expression in primary breast carcinoma. Br J Cancer 1994; 70:335-41. 39. Wagenblast E, Soto M, Gutierrez-Angel S, et al. A model of breast cancer heterogeneity reveals vascular mimicry as a driver of metastasis. Nature 2015; 520: 358-62. 40. Kamby C, Rasmussen BB, Kristensen B. Estrogen receptor status of primary breast carcinomas and their metastases. Relation to pattern of spread and survival after recurrence. Br J Cancer 1989; 60:252-7. 41. Teixeira MR, Pandis N, Bardi G, et al. Clonal heterogeneity in breast cancer: karyotypic comparisons of multiple intra- and extra-tumorous samples from 3 patients. J Cancer 1995; 63:63-8. 42. Kerbel RS. Growth dominance of the metastatic cancer cell: cellular and molecular aspects. Adv Cancer Res 1990; 55:87-132. 43. Bieche I, Lidereau R. Genetic alterations in breast cancer. Genes Chromosomes Cancer 1995; 14:227-51. 44. Pertschuk LP, Axiotis CA, Feldman JG, et al. Marked intratumoral heterogeneity of the proto-oncogene Her-2/neu determined by three different detection systems. Breast J 1999; 5:369-74. 45. Hanna WM, Ruschoff J, Bilous M. HER2 in situ hybridization in breast cancer: clinical implications of polysomy 17 and genetic heterogeneity. Mod Pathol 2014; 27:4-18. 46. Prendeville S, Feeley L, Bennett MW, O’Connell F, Browne TJ. Reflex repeat HER2 testing of grade 3 breast carcinoma at excision using immunohistochemistry and in situ analysis: frequency of HER2 discordance and utility of core needle biopsy parameters to refine case selection. Am J Clin Pathol 2016; 145:75-80. 47. Dialani V, Chadashvili T, Slanetz PJ. Role of imaging in neoadjuvant therapy for breast cancer. Ann Surg Oncol 2015; 22:1416-24. 48. Pahk K, Kim S, Choe JG. Early prediction of pathological complete response in luminal B type neoadjuvant chemotherapy-treated breast cancer patients: comparison between interim 18F-FDG PET/CT and MRI. Nucl Med Commun 2015; 36:887-91. 49. Boileau JF, Poirier B, Basik M, et al. Sentinel node biopsy after neoadjuvant chemotherapy in biopsy-proven node-positive breast cancer: the SN FNAC study. J Clin Oncol 2015; 33:258-64. 50. Zhang GC, Zhang YF, Xu FP, et al. Axillary lymph node status, adjusted for pathologic complete response in breast and axilla after neoadjuvant chemotherapy, predicts differential disease-free survival in breast cancer. Curr Oncol 2013; 20: e180-92.
Clinical Breast Cancer Month 2017
-7
Evaluation of Discordance in Primary Tumor and Lymph Node Response After Neoadjuvant Therapy in Breast Cancer Christina A. Fleming,1 Karen McCarthy,1 Ciara Ryan,2 Aoife McCarthy,2 Seamus O’Reilly,1 Deirdre O’Mahony,1 Tara Jane Browne,2 Paul Redmond,1 Mark A. Corrigan1 Abstract Primary tumor and lymph node (LN) response discordance may be observed in breast cancer patients receiving neoadjuvant therapy. We quantified the rate of this discordance and identified factors predictive of it. Almost 15% of primary tumors with complete pathologic response had persistently positive LNs. This is concerning, as we may be undertreating a significant portion of patients. Background: Neoadjuvant therapy (NAT) offers a unique opportunity to assess tumor response to systemic agents. However, a discrepancy may exist between the response of the primary tumor and involved nodes. We report on the frequency of response discordance after NAT in breast cancer. Patients and Methods: All consecutive node-positive patients receiving NAT in our department from 2009 to 2014 were identified. Patient demographics, and radiologic and pathologic features were tabulated. Tumor response was estimated by magnetic resonance imaging of the breast. Lymph node (LN) response was estimated from pathologic treatment response measurements. Statistical analysis was performed. Results: A total of 108 node-positive patients treated with NAT were eligible for inclusion. Median age was 51.73 years (range, 20-87 years). All patients underwent axillary clearance, and 62% underwent mastectomy. A 40% mean reduction in tumor size was observed. Statistically, a positive correlation between tumor and LN response after NAT was observed (Spearman correlation coefficient, r ¼ 0.46, P < .001). Complete pathologic response was observed in 17 patients (15.7%). However, 21 patients experienced complete LN response, with only 81% of these patients (n ¼ 17) experiencing a complete response in tumor also. A complete response was observed in tumor in 20 patients, and this predicted complete nodal response in 85% of cases (n ¼ 17). Fifteen percent of primary tumors with complete pathologic response had persistently positive LNs. Conclusion: A significant discordance exists between the primary tumor and LN response, representing a concern for the lack of response of occult regional or systemic metastases due to potential biologic heterogeneity. Clinical Breast Cancer, Vol. -, No. -, --- ª 2017 Elsevier Inc. All rights reserved. Keywords: Axillary surgery, Biogenetics, Breast surgery, Neoadjuvant therapy, Response discordance
Introduction The role of neoadjuvant therapy (NAT) in the management of breast cancer is evolving.1 Use of NAT is dependent on tumor size, receptor status, and nodal status, and it utilizes chemotherapy or hormone-based treatment. Historically, the primary aims of NAT 1
Breast Research Centre Department of Histopathology, Cork University Hospital, Cork, Ireland
2
Submitted: Oct 15, 2017; Revised: Nov 20, 2017; Accepted: Nov 21, 2017 Address for correspondence: Christina A. Fleming, MB, MCh, MRCSI, Department of Academic Surgery, Cork University Hospital, Wilton, Cork, Ireland E-mail contact: christina.fl[email protected]
1526-8209/$ - see frontmatter ª 2017 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.clbc.2017.11.016
were to downsize inoperable tumors, thus allowing surgical resection, and to decrease operable tumor size, ultimately allowing higher rates of breast-conserving surgery (BCS).2,3 However, NAT has also been reported to reduce the risk of distant recurrence by targeting circulating micrometastases.4 Furthermore, NAT offers a unique opportunity to assess tumor response to systemic agents, which may be described as a human model system to explore the efficacy of therapies and response to therapy, and can also identify specific patient cohorts at higher risk of future disease relapse.5,6 Currently, axillary lymph node (LN) status is still considered a significant prognostic factor for overall outcomes in primary breast cancer.7 Additionally, axillary LN disease burden after completion of NAT
Clinical Breast Cancer Month 2017
-1
Discordance in Tumor and LN Response is considered to be a strong predictor of subsequent disease relapse.8 The aim of treatment is therefore to synchronously improve primary tumor and LN tumor cell burden. It has been estimated that up to 40% of node-positive disease reverts to node-negative status after NAT.9 Complete pathologic response (ypCR) rates as high as 67% can be achieved in specific breast cancer groups, most notably human epidermal growth factor receptor 2 (HER2)-positive tumors.10 Some reports have associated overall ypCR with a survival benefit.11,12 It is emerging that a discrepancy may exist between the response of the primary tumor and involved nodes after NAT. A differential response is difficult to explain but perhaps reflects biomolecular heterogeneity in the tumor observed in primary and secondary metastatic sites after lymphatic spread or circulation of micrometastases.13 Performing sentinel LN biopsy after NAT has been suggested to assess LN response to NAT independent of tumor response and thus identify patients who could avoid the concurrent morbidity of axillary lymphadenectomy.14 However, the ACOSOG Z1071 study has concluded that the sensitivity of sentinel LN biopsy in this setting is insufficient to change current practice.15 We sought to assess the overall frequency of discordance in tumor and LN response after NAT in breast cancer, regardless of tumor type. We further aimed to identify tumor factors associated with nonresponse or reduced response in LN when response, even ypCR, is observed in primary tumor.
Tumor response to NAT was calculated on the basis of size by subtraction of the largest dimension in size of tumor bed calculated on histologic analysis of the surgical specimen (ypT) from that calculated on pre-NAT magnetic resonance imaging (MRI) of the breast (cT) and a percentage response calculated, as previously validated.20 Where MRI was unavailable or not performed, ultrasound, mammography, and computed tomography (CT) were used, in this sequence. Tumor size was noted as the maximal dimension when coronal, sagittal, and axial sections were measured. LN response to therapy was calculated by expression of number of LNs collected with evidence of viable tumor and/or treatment effect (fibrosis) or response to therapy compared to the entire number of LNs collected after axillary lymphadenectomy.
Statistical Analysis and Ethical Approval
All patients at the Southern Breast Cancer Centre at Cork University Hospital who had node-positive disease at diagnosis and went on to receive NAT (chemotherapy and/or hormone therapy) over a 5-year inclusive period from 2009 to 2014 were identified using a prospectively maintained database. For inclusion, patients were required to have undergone surgery during this time period that included level II axillary lymphadenectomy to allow for full histologic analysis of LN response. The following parameters were recorded: patient demographics including age and gender, surgical procedure including requirement for reexcision of margins or completion mastectomy, and whether NAT was administered as chemotherapy or as a hormonal agent.
Histopathologic Analysis
Results
Data on tumor and LN status before NAT and postoperatively after completion of NAT and subsequent surgery were recorded. All breast cancer specimens were analyzed by breast-specialist consultant histopathologists using a standardized approach. All cases were processed similarly in a single laboratory with uniform preanalytics as per the American Society of Clinical Oncology/College of American Pathologists guidelines.16 Equivocal HER2 positivity was further analyzed by fluorescence in-situ hybridization (FISH) or bright-field dual-color in-situ hybridization (BDISH), and repeat HER2 testing on grade III cancers was also performed as per current guidelines.16,17 Pre- and postoperative predictive and prognostic parameters including tumor histologic subtype, grade, presence or absence of lymphovascular invasion, and hormone and HER2 receptor status were reported, and tumors were staged according to American Joint Committee on Cancer (AJCC) 7th edition
Demographics and Treatment
Patients and Study Design
-
Calculation of Tumor and LN Response to NAT
A prior sample size calculation indicated that a sample of 34 patients was necessary to detect a medium effect size (Cohen’s f ¼ 0.50) in a paired t test comparing tumor response and LN response with a power of 80% and a level of significance of .05.21 To account for lack of normality in distribution, the sample size was increased by 15%, and a sample of 40 patients was therefore necessary.22 All other statistical analysis was performed by SPSS 22 (IBM SPSS, Chicago, IL). Descriptive analyses were performed on demographic information. Correlation of tumor and LN response to NAT was calculated by Spearman’s test with normality not assumed, and strength of correlation was calculated on the basis of standard methods.23 Discordance was calculated on the basis of percentage subtraction. Univariate analysis was performed to identify factors predictive of response discordance, and factors achieving P < .25 were included for multivariate analyses.24 Multivariate analyses were performed on a logistic regression model to identify tumor features predictive of discordance response in LN disease. Ethical approval was prospectively obtained from the Clinical Research and Ethics Committee of the Cork Teaching Hospitals.
Methods
2
guidelines and the College of American Pathologists.17,18 The MD Anderson Residual Cancer Burden Protocol was used to analyze all residual disease in breast specimens after surgery.19
Clinical Breast Cancer Month 2017
A total of 3275 breast cancer patients (screen detected and symptomatic) were treated at our institute between 2009 and 2014. Of these, 203 patients were treated with NAT; of these, 108 patients proceeded to level II axillary lymphadenectomy for nodepositive disease and were eligible for study inclusion. All patients included were women; median age was 51.73 years (range, 20-87 years). An overall mastectomy rate of 64.8% was observed (62% at first operation and a further 2.8% after initial BCS). Reexcision of margins was performed in 10.2% of cases.
Tumor Characteristics, and Nodal and Metastatic Disease Table 1 lists the tumor and nodal characteristics at diagnosis and the quantification of overall response to NAT using the MD Anderson Residual Cancer Burden (RCB) score. Overall, 86.1% of
Christina A. Fleming et al Table 1 Tumor and Nodal Characteristics at Diagnosis and Quantification of Overall Response to NACT Using RCB Score Characteristic
Value
At Diagnosis Tumor Type IDC
93 (86.1)
ILC
8 (7.4)
Mixed
7 (6.5)
Tumor Grade Low (I þ II)
44 (40.7)
High (III)
64 (59.3)
Tumor Size (mm) Mean
47.9
SD
21.57
Range
4-100
Receptor Status ERþPRþ
60 (55.6)
HER2þ
52 (48.1)
Triple negative
16 (14.8)
LVI 28 (25.9)
Absent
80 (74.1)
Lymph Node Status Positive
106 (98.1)
Negative
2 (1.9)
After NACT RCB Score SD Range
2.68 1.53 0-5.45
RCB Class ypCR
Response to NAT Tumor size on pre-NAT MRI was available in 54.6% of cases (n ¼ 59). Where MRI was unavailable or was not performed, ultrasound (n ¼ 20), mammography (n ¼ 28), and CT (n ¼ 1) were used. Accuracy of these imaging modalities for tumor size estimation has been described previously.25,26 Overall, 80.6% of tumors exhibited some response to NAT (mean SD response, 40.53 56.7% reduction in size) (Table 2). Conversion rate from mastectomy to BCS based on tumor size alone was 16.7% (tumors measuring 5 cm on pre-NAT MRI that reduced in size and proceeded to BCS). ypCR was seen in 15.7% (n ¼ 17) of patients (ie, combined ypCR in tumor and LNs). However, a total of 18.5% of tumors (n ¼ 20) underwent complete response after NAT, while 14.8% of tumors (n ¼ 16) did not respond or were found to be larger than originally estimated. Evidence of nodal disease regression was seen in 60.2% (n ¼ 65) and 19.4% (n ¼ 21) of patients who experienced ypCR (Table 2).
Tumor
Present
Mean
diagnosis (n ¼ 5 bone). Only 2 of these patients had lymphovascular invasion at initial core biopsy. All 5 had lymphovascular invasion at final specimen analysis.
17 (15.7)
I
4 (3.7)
II
39 (36.1)
III
48 (44.5)
Data are presented as n (%) unless otherwise indicated. Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; IDC ¼ invasive ductal carcinoma; ILC ¼ invasive lobular carcinoma; LVI ¼ lymphovascular invasion; NACT ¼ neoadjuvant chemotherapy; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden; ypCR ¼ pathologic complete response.
patients (n ¼ 93) had invasive ductal carcinoma (IDC) at initial biopsy, and 7.4% (n ¼ 8) had invasive lobular carcinoma. At initial biopsy, 59.3% of tumors were deemed grade III. This figure dropped to 35.2% after NAT (n ¼ 4 grade I, n ¼ 46 grade II) and was statistically significant (P < .001). The mean pre-NAT tumor size was 47.9 mm (range, 4-100 mm). Postoperatively, based on histologic assessment, this value reduced to 28.66 mm (range, 0-150 mm), reflecting overall positive tumor response to treatment (P < .001). The increased range after NAT may be explained by either underestimation of tumor size on pre-NAT imaging or by tumor progression, which was observed in 14.8% of cases (n ¼ 16). A mean collection of 22.08 LNs was achieved (range, 4-52 LNs; SD 9.184). Five patients (4.6%) had distant metastases at the time of
LN Response Correlation. An overall positive correlation was observed between tumor and LN response after NAT (Spearman correlation coefficient, r ¼ 0.461, P < .001). The strength of correlation increases (r ¼ 0.594, P < .001) when LNs with a zeroresponse rate are excluded. This may be explained by a tumor progression rate of 23.3% in this group and potential progression of nodal disease that is not fully quantifiable preoperatively. Seventeen patients (15.7%) experienced combined ypCR in tumor and LN disease (correlation coefficient, r ¼ 1). Tumor characteristics associated with greatest discordance in tumor and LN response are outlined in Table 3. Tumors showing the strongest correlation in response included invasive lobular carcinoma measuring > 5 cm with a triple-negative hormone profile (Table 3). LN Response Discordance. A mean discordance in response of tumor and LN disease of 36.9% was observed after NAT (Table 4). Most discordant tumors (> 50% discordance) were triple positive, followed by HER2 positive regardless of estrogen receptor (ER) status and IDC tumors measuring 2 to 5 cm before NAT
Table 2 Independent Tumor Response and Lymph Node Response to Neoadjuvant Therapy Characteristic ypCR Response No response Disease progression
Tumor Responsea 20 87 5 16
(18.5) (80.6) (4.6) (14.8)
Lymph Node Response 21 (19.4) 65 (60.2) 43 (39.8) —b
Abbreviation: ypCR ¼ pathologic complete response. a Overall number of tumors that responded to therapy and those with ypCR. b Progression is nonquantifiable as a result of the inability to perform complete lymph node assessment preoperatively. Progression may be confounded by initial underestimation of tumor burden on imaging thanks to the utility of differing modalities.
Clinical Breast Cancer Month 2017
-3
Discordance in Tumor and LN Response Table 3 Tumor and Lymph Node Percentage Response and Discordance Value by Subgroup With Correlations Performed Tumor Response (%), Mean (SD)
Lymph Node Response, % Mean (SD)
Differential Response, % (TLNR)
Pa
Correlation Coefficient, R (P)
Invasive ductal carcinoma
42.28 (56.86)
37.56 (38.64)
4.72 (0.89)
.43
0.33 (.001)
Invasive lobular carcinoma
30.63 (61.37)
30.25 (38.63)
0.38 (0.99)
.98
0.73 (.41)
Mixed
28.57 (55.03)
42.86 (53.45)
14.29 (0.67)
.22
0.87 (.01)
2-5 cm
38.76 (58.29)
37.47 (39.2)
1.29 (0.97)
.86
0.31 (.01)
>5 cm
43.31 (54.7)
37.19 (40.03)
6.12 (0.86)
.40
0.55 (<.001)
22.44 (57.97)
26.63 (31.56)
4.19 (0.84)
.63
0.30 (.049)
53 (53.23)
44.38 (42.82)
8.62 (0.84)
.20
0.40 (.001)
Characteristic Tumor Type
Size
Grade Low (I þ II) High (III) LVI Present
61.07 (54.5)
44.43 (46.07)
16.64 (0.73)
.10
0.48 (.01)
Absent
33.34 (56.017)
34.89 (36.70)
1.55 (0.96)
.802
0.351 (.001)
ERþPRþ
25.35 (59.96)
31.83 (34.76)
6.48 (0.79)
.38
0.39 (.002)
HER2þ
51.77 (51)
47.25 (41.56)
4.52 (0.91)
.57
0.26 (.07)
HER2
30.09 (60.11)
28.18 (35.08)
1.91 (0.94)
.79
0.48 (<.001)
Triple negative
38.58 (60.76)
25.17 (38.55)
13.41 (0.65)
.39
0.54 (.07)
Receptor Status
RCB Class 100
100 (1)
—
—
I
94.50 (9.71)
56.25 (47.95)
38.25 (0.60)
.251
0.32 (.684)
II
45.92 (51.85)
39.97 (36.27)
5.95 (0.87)
.56
0.27 (.873)
III
8.49 (50.99)
13.85 (19.09)
5.36 (0.61)
.488
0.532 (.094)
ypCR
100
Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor; LVI ¼ lymphovascular invasion; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden; TLNR ¼ tumorelymph node response ratio. a Statistically significant at P < .05.
(35%-40%). Larger (> 5 cm) triple-negative tumors showed the least discordance in response (< 35%). Table 4 lists factors predictive of tumor and LN discordance on univariate analysis. Triple-positive disease (P ¼ .002), IDCs (P ¼ .045), and disease measuring < 5 cm (P ¼ .021) remained significantly associated with differential response on multivariate analysis.
Discussion
4
-
We identified that a discordance in tumor and LN response to NAT in breast cancer exists. This observation is consistent with suggested hypotheses in other studies.27-30 We have further identified particular tumor characteristics that are associated with a differential response (IDCs, < 5 cm size, triple-positive receptor profile). We also identified that even tumors with ypCR will not always have ypCR in nodal disease. Although there is a significant association between ypCR in tumor and nodal disease, an absolute correlation between tumor and LN response was not observed. Differential response in primary tumor and LN disease burden in breast cancer may be due to biomolecular differences in the tumor microenvironment of the primary tumor site and secondary metastatic foci or tumor heterogeneity.13,27 Central to a differential response is an observed discordance in hormone (ER and
Clinical Breast Cancer Month 2017
progesterone receptor) and HER2 receptor status between primary tumor and metastatic sites.27,30 Although this phenomenon has been reported previously, in practice, receptor testing is still reserved for primary tumor tissue assessment only and not metastatic foci (including LNs) before NAT planning. Hormone therapy and HER2-targeted therapy are central to the current neoadjuvant management of breast cancer.31-33 The purpose of NAT is combined reduction in primary tumor size and a reduction of neoplastic disease in metastatic foci.2-4 However, differential receptor status can lead to therapeutic discrepancy. It is suggested that change in hormone receptor status may occur between primary tumor and LN disease sites, which may explain response in primary site and not in nodal disease and vice versa.34,35 Further hypotheses to explain discordance in hormone receptor status between primary tumor and metastatic sites have been provided.36-44 It is unknown if heterogeneity of receptor status is a true biologic finding or if it is due to reporting discrepancies. Tissue fixation and test type (most commonly immunohistochemistry) have been associated with technical discrepancies in studies examining receptor status in primary and metastatic sites.36 Thresholds for defining hormone receptor positivity may also be involved. Currently, international guidelines used in practice require low-level
Table 4 Factors Associated With Discordance in Tumor and Lymph Node Response to Neoadjuvant Therapya
Characteristic Overall
Response Discordance (%), Mean (SD)
Discordance Univariate Analysis Response Discordance P
Discordance Multivariate Analysis
OR
95% CI
P
OR
95% CI
P
1.36
0.33-1.79
.021
1.18
0.15-3.50
.045
1.38
0.26-1.62
.66
1.59
0.47-5.72
.44
36.96 (40.27)
Tumor Size 2-5 cm
39.42 (43.52)
<.001
1.88
0.821-2.01
<.001
>5 cm
32.95 (34.44)
.337
1.19
0.75-1.90
.55
IDC
38.66 (42.16)
<.001
1.58
0.50-4.98
.02
ILC
32.63 (24.37)
.525
1.15
0.46-2.89
.76
Mixed
15.50 (22.78)
.048
1.53
0.463-5.023
.44
Low (I þ II)
38.93 (41.44)
<.001
1.49
0.59-2.84
.06
High (III)
35.75 (40.05)
.622
1.16
0.75-1.80
.32
ERþPRþ
37.30 (42.02)
.250
1.92
0.068-2.1
.08
HER2þ
38.02 (42.33)
.037
1.45
0.46-2.11
.56
Triple positive (n ¼ 10)
51.20 (66.30)
<.001
2.1
1.63-3.35
<.001
1.86
0.17-4.96
.002
Triple negative (n ¼ 12)
34.75 (39.10)
.24.7
1.31
0.57-3.02
.07
1.01
0.40-12.91
.35
I
56.25 (41.46)
.331
1.55
0.79-1.95
.06
1.34
0.428-2.30
.141
II
35.29 (44.65)
.752
1.38
0.83-2.29
.14
1.25
0.43-2.98
.257
III
41.06 (40.41)
.355
1.85
0.85-4.01
.041
1.306
0.923-1.85
.086
Tumor Type
Grade
Receptor Status
Clinical Breast Cancer Month 2017
Abbreviations: CI ¼ confidence interval; ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; IDC ¼ invasive ductal carcinoma; ILC ¼ invasive lobular carcinoma; OR ¼ odds ratio; PR ¼ progesterone receptor; RCB ¼ MD Anderson Residual Cancer Burden. a Discordance for univariate and multivariate analysis is defined as difference in tumor and lymph node response of > 30%.
Christina A. Fleming et al
RCB Score
-5
Discordance in Tumor and LN Response
6
-
positivity for receptor positivity status on immunohistochemistry; only 1% positive tumor nuclei identified on immunostaining is required for a positive finding for ER and progesterone receptor status, with equivocal HER2 positivity further assessed using FISH or BDISH.15,16 However, it is also accepted that significant intratumor heterogeneity can exist, so receptor positivity may not be representative of the entire tumor burden.37,38 HER2 heterogeneity is reported to occur in 5% to 40% of breast cancers, depending on the definition used.45 This may potentially result in a false-negative HER2 report on core needle biopsy, although this is uncommon.46 Tumor evolution is also driven by the diversity of cancer cell populations and their microenvironment, and changes in hormone receptor status may occur during the metastatic process.39 Loss of hormone receptor positivity after circulation of micrometastases and metastatic disease has been reported. One explanation for this may be clonal selection during the metastatic process.40 As already described, significant tumor heterogeneity and polyclonality may occur in as many as 70% of cases by molecular analysis.37,41 At initiation of the metastatic process, differential clones are balanced toward those with a preferential potential for metastasing.42 Therefore, a small neoplastic focus within the heterogeneous tumor bed may metastasize with genotypic and phenotypic differences from that predominant in the primary tumor.41 Other biomolecular mechanisms that may be involved include genetic miscoding during tumor progression, discriminatory effect of prior chemotherapy or hormone therapy, and selection of resistant tumor clones.38,43,44 With increasing use of NAT in breast cancer, the number of patients experiencing complete response has been more widely observed.11 Accurate diagnosis of ypCR before surgery may allow avoidance of unnecessary surgery as well as avoidance of unnecessary surgical morbidity and mortality. Preoperative MRI after NAT has significant accuracy for diagnosing tumor ypCR.47 Positron emission tomographic scanning has been shown to be superior for diagnosing ypCR in both tumor bed and nodal distribution, though it is not widely utilized.48 Tumor bed size is certainly measurable via preoperative imaging; however, nodal disease is not fully quantifiable. Sentinel LN biopsy after completion of NAT has in the past been discouraged as a result of its association with significantly high false-negative rates, as observed in the SENTINA trial and in the ACOSOG Z1071 study findings.14,49 Although an association between tumor and LN ypCR was identified in our study, an absolute correlation was not observed; thus, identifying ypCR of primary tumor after NAT MRI cannot sufficiently predict ypCR in nodal disease. While clear attention to study design was paid, there were a number of limitations to this study. This is a retrospective account of tumor and LN response. Histopathologic assessment of breast cancer specimens at our institute is standardized and is performed in a uniform manner for each patient, in keeping with AJCC 7th edition guidelines.18 Reporting is also standardized. The patient cohort may be considered heterogeneous as a result of inclusion of patients who received neoadjuvant chemotherapy and/or hormone therapy, and different biologic subtypes were also included. For example, small cohorts of lobular carcinoma patients were included, and only one patient had a grade I tumor. Of note, we do not routinely evaluate the proliferative marker Ki-67 in our pathology
Clinical Breast Cancer Month 2017
laboratory. A low conversion rate from mastectomy to BCS were also noted. However, this was based on an analysis of tumor size alone without consideration of other factors, such as patient preference. It should also be noted that only 38.1% of these whose disease had no response or those who experienced tumor progression underwent pre-NAT MRI, and initial tumor burden was based on ultrasound (n ¼ 4, 19%), mammography (n ¼ 8, 38.1%), and CT (n ¼ 1, 4.8%). Therefore, assessment of progression may be subject to bias in this smaller subgroup. All patients underwent clinical assessment midway through NAT to identify progression of disease.
Conclusion We found that a discordance in response may exist between primary tumor and LN disease after NAT in breast cancer, and that ypCR of primary tumor is not always associated with complete nodal response. This discordance may be explained by intratumor heterogeneity and alterations in phenotypic receptor expression after lymphatic and hematogenous circulation, creating a scenario of reduced therapeutic responsiveness. This theory is supported by a 40% nonresponse in nodal disease. In response discordance, it may be reasonably inferred that sites of metastases are sampled and fully profiled in significantly discordant groups to allow for fully targeted, patient-tailored NAT at both the primary and metastatic tumor level. This is important because nodal response to NAT has prognostic value, independent of tumor response.50 Further investigation of the concept of tumor and LN response discordance can continue to enhance our ongoing clinical knowledge of this disease.
Clinical Practice Points Anecdotally, it has been reported that patients may not experi-
ence the same degree of response in their primary tumor and sites of LN disease after NAT for breast cancer. This may be explained by biomolecular or even genetic heterogeneity in primary tumor and metastatic disease sites. However, to date, this phenomenon has not been adequately quantified and reported. Furthermore, factors that predict this discordance have not been identified. We found that 15% of breast cancer patients who experience ypCR in their breast primary tumor after NAT have persistent LN disease, and 1 in 10 shows no response in nodal disease. We found that IDCs < 5 cm in size with a triple-positive receptor profile had the highest rate of response discordance. This finding is clinically relevant, as nodal response to NAT has prognostic value independent of tumor response. In response to this discordance, it may be reasonable to sample and fully profile LN metastases in markedly discordant groups to allow for improved targeted patient-tailored NAT at both the primary and metastatic tumor levels in this era of precision medicine.
Acknowledgments We acknowledge and thank all staff members at the Southern Breast Cancer Centre, Cork University Hospital, who contributed greatly to the completion of this work.
Christina A. Fleming et al Disclosure The authors have stated that they have no conflict of interest.
References 1. Rastogi P, Anderson SJ, Bear HD, et al. Preoperative chemotherapy: updates of National Surgical Adjuvant Breast and Bowel Project protocols B-18 and B-27. J Clin Oncol 2008; 26:778-85. 2. Peintinger F, Symmans WF, Gonzalez-Angulo AM, et al. The safety of breastconserving surgery in patients who achieve complete pathological response after neoadjuvant chemotherapy. Cancer 2006; 107:1248-54. 3. Wang S, Zhang Y, Yang X, et al. Shrink pattern of breast cancer after neoadjuvant chemotherapy and its correlation with clinical pathological factors. World J Surg Oncol 2013; 11:166. 4. Kasimir-Bauer S, Bittner AK, König L, et al. Does primary neoadjuvant systemic therapy eradicate minimal residual disease? Analysis of disseminated and circulating tumor cells before and after therapy. Breast Cancer Res 2016; 18:20. 5. Schwartz GF, Hortobagyi GN. Proceedings of the consensus conference on neoadjuvant chemotherapy in carcinoma of the breast, April 26-28, 2003, Philadelphia, Pennsylvania. Cancer 2004; 100:2512-32. 6. Yadav BS, Sharma SC, Singh R, Singh G. Patterns of relapse in locally advanced breast cancer treated with neoadjuvant chemotherapy followed by surgery and radiotherapy. J Cancer Res Ther 2007; 3:75-80. 7. Wu SG, Sun JY, Zhou J, et al. Number of negative lymph nodes can predict survival of breast cancer patients with four or more positive lymph nodes after postmastectomy radiotherapy. Radiat Oncol 2014; 9:284. 8. Escobare PF, Patrick RJ, Rybicki LA, Hicks D, Weng DE, Crowe JP. Prognostic significance of residual breast cancer and axillary node involvement for patients who had primary induction chemotherapy for advanced breast cancer. Ann Surg Oncol 2006; 13:783-7. 9. Thompson AM, Moulder-Thompson SL. Neoadjuvant treatment of breast cancer. Ann Oncol 2012; 23(suppl 10):231-6. 10. Buzdar AU, Ibrahim NK, Francis D, et al. Significantly higher pathological complete remission rate after neoadjuvant therapy with trastuzumab, paclitaxel and epirubicin chemotherapy: results of a randomized trial in human epidermal growth factor receptor 2epositive operable breast cancer. J Clin Oncol 2005; 23:3676-85. 11. Wang-Lopez Q, Chalabi N, Abrial C, et al. Can complete pathological response (pCR) be used as a surrogate marker of survival after neoadjuvant therapy for breast cancer? Crit Rev Oncol Hematol 2015; 95:88-104. 12. Chollet P, Amat S, Cure H, et al. Prognostic significance of a complete pathological response after induction chemotherapy in operable breast cancer. Br J Cancer 2002; 86:1041-6. 13. Symmans WF, Liu J, Knowles DM, Inghirami G. Breast cancer heterogeneity: evaluation of clonality in primary and metastatic lesions. Hum Pathol 1995; 26: 210-6. 14. Kuehn T, Bauerfeind I, Fehm T, et al. Sentinel-lymph-node biopsy in patients with breast cancer before and after neoadjuvant chemotherapy (SENTINA): a prospective, multicenter cohort study. Lancet Oncol 2013; 14:609. 15. Boughey JC, Suman VJ, Mittendorf EA, et al. Sentinel lymph node surgery after neoadjuvant chemotherapy in patients with node-positive breast cancer: the ACOSOG Z1071 (Alliance) clinical trial. JAMA 2013; 310:1455-61. 16. Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007; 25:118-45. 17. Nitta H, Hauss-Wegrzyniak B, Lehrkamp M, et al. Development of automated bright field double in situ hybridization (BDISH) application for HER2 gene and chromosome 17 centromere (CEN 17) for breast carcinomas and an assay performance comparison to manual dual color HER2 fluorescence in situ hybridization (FISH). Diagn Pathol 2008; 3:41-53. 18. Edge S, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. American Joint Committee on Cancer (AJCC) cancer staging manual. 7th ed. New York: Springer-Verlag; 2011. 19. Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007; 25: 4414-22. 20. Lee HJ, Song IH, Seo AN, et al. Correlations between molecular subtypes and pathologic response patterns of breast cancers after neoadjuvant chemotherapy. Ann Surg Oncol 2015; 22:392-400. 21. Cohen J. A power primer. Psychol Bull 1992; 112:155-9. 22. Erdfelder E, Faul F, Buchner A. GPOWER: a general power analysis program. Behav Res Methods 1996; 28:1-11.
23. Croux C, Dehon C. Influence functions of the Spearman and Kendall correlation measures. Stat Methods Appl 2012; 19:497-515. 24. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159-74. 25. Fischer U, Kopka L, Grabbe E. Breast carcinoma: effect of preoperative contrastenhanced MR imaging on the therapeutic approach. Radiology 1999; 213:881. 26. Yeh E, Slanetz P, Kopans DB, et al. Prospective comparison of mammography, sonography, and MRI in patients undergoing neoadjuvant chemotherapy for palpable breast cancer. AJR Am J Roentgenol 2005; 184:868. 27. Arslan C, Sari E, Aksoy S, Altundag K. Variation in hormone receptor and HER-2 status between primary and metastatic breast cancer: review of the literature. Expert Opin Ther Targets 2011; 15:21-30. 28. Polyak K. Heterogeneity in breast cancer. J Clin Invest 2015; 121:3786-8. 29. Holdaway IM, Bowditch JV. Variation in receptor status between in primary and metastatic breast cancer. Cancer 1983; 5:479-85. 30. Sari E, Guler G, Hayran M, Gullu I, Altundag K, Ozisik Y. Comparative study of the immunohistochemical detection of hormone receptor status and HER-2 expression in primary and paired recurrent/metastatic lesions of patients with breast cancer. Med Oncol 2011; 28:57-63. 31. Cortazar P, Zhang L, Untch M, et al. Pathological complete response and longterm clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014; 384:164. 32. Carey LA, Winer EP. Defining success in neoadjuvant breast cancer trials. Lancet 2014; 384:115. 33. Mieog JS, van der Hage JA, van de Velde CJ. Preoperative chemotherapy for women with operable breast cancer. Cochrane Database Syst Rev 2007:CD005002. 34. Lower EE, Glass EL, Bradley DA, et al. Impact of metastatic estrogen receptor and progesterone receptor status on survival. Breast Cancer Res Treat 2005; 90:65-70. 35. Lower EE, Glass E, Blau R, Harman S. HER-2/neu expression in primary and metastatic breast cancer. Breast Cancer Res Treat 2009; 113:301-6. 36. Gown AM. Current issues in ER and HER2 testing by HC in breast cancer. Mod Pathol 2008; 21(suppl 2):S8-15. 37. Teixeira MR, Pandis N, Bardi G, et al. Karyotypic comparisons of multiple tumorous and macroscopically normal surrounding tissue samples from patients with breast cancer. Cancer Res 1996; 56:855-9. 38. Rasbridge SA, Gillett CE, Seymour AM, et al. The effects of chemotherapy on morphology, cellular proliferation, apoptosis and oncoprotein expression in primary breast carcinoma. Br J Cancer 1994; 70:335-41. 39. Wagenblast E, Soto M, Gutierrez-Angel S, et al. A model of breast cancer heterogeneity reveals vascular mimicry as a driver of metastasis. Nature 2015; 520: 358-62. 40. Kamby C, Rasmussen BB, Kristensen B. Estrogen receptor status of primary breast carcinomas and their metastases. Relation to pattern of spread and survival after recurrence. Br J Cancer 1989; 60:252-7. 41. Teixeira MR, Pandis N, Bardi G, et al. Clonal heterogeneity in breast cancer: karyotypic comparisons of multiple intra- and extra-tumorous samples from 3 patients. J Cancer 1995; 63:63-8. 42. Kerbel RS. Growth dominance of the metastatic cancer cell: cellular and molecular aspects. Adv Cancer Res 1990; 55:87-132. 43. Bieche I, Lidereau R. Genetic alterations in breast cancer. Genes Chromosomes Cancer 1995; 14:227-51. 44. Pertschuk LP, Axiotis CA, Feldman JG, et al. Marked intratumoral heterogeneity of the proto-oncogene Her-2/neu determined by three different detection systems. Breast J 1999; 5:369-74. 45. Hanna WM, Ruschoff J, Bilous M. HER2 in situ hybridization in breast cancer: clinical implications of polysomy 17 and genetic heterogeneity. Mod Pathol 2014; 27:4-18. 46. Prendeville S, Feeley L, Bennett MW, O’Connell F, Browne TJ. Reflex repeat HER2 testing of grade 3 breast carcinoma at excision using immunohistochemistry and in situ analysis: frequency of HER2 discordance and utility of core needle biopsy parameters to refine case selection. Am J Clin Pathol 2016; 145:75-80. 47. Dialani V, Chadashvili T, Slanetz PJ. Role of imaging in neoadjuvant therapy for breast cancer. Ann Surg Oncol 2015; 22:1416-24. 48. Pahk K, Kim S, Choe JG. Early prediction of pathological complete response in luminal B type neoadjuvant chemotherapy-treated breast cancer patients: comparison between interim 18F-FDG PET/CT and MRI. Nucl Med Commun 2015; 36:887-91. 49. Boileau JF, Poirier B, Basik M, et al. Sentinel node biopsy after neoadjuvant chemotherapy in biopsy-proven node-positive breast cancer: the SN FNAC study. J Clin Oncol 2015; 33:258-64. 50. Zhang GC, Zhang YF, Xu FP, et al. Axillary lymph node status, adjusted for pathologic complete response in breast and axilla after neoadjuvant chemotherapy, predicts differential disease-free survival in breast cancer. Curr Oncol 2013; 20: e180-92.
Clinical Breast Cancer Month 2017
-7