Response rate to neoadjuvant chemotherapy measured on imaging predicts early recurrence and death in breast cancer patients with lymph node involvements

The use of preoperative neoadjuvant chemotherapy for breast cancer is effective as postoperative adjuvant therapy, permits more lumpectomies, and can be used to study breast cancer biology (1). Many reports indicate that the histological response to neoadjuvant chemotherapy is the single most important prognostic factor in patients with advanced breast cancer (2, 3). The outcome for patients who achieve a pathological complete response continues to be superior with a 50% reduction in the risk of death compared to patients with residual invasive cancer on pathologic examination (4).

Lymph node metastasis is also one of the most important prognostic factors in breast cancer patients (5). The status of the axilla after neoadjuvant chemotherapy has been shown to be of prognostic importance. Patients with no residual axillary disease after neoadjuvant chemotherapy have increased overall survival and disease-free survival compared with those with residual axillary disease (6–8). Although pathological response is the strongest prognostic factor, response rates vary according to various parameters such as dissociation between breast and axillary node responses, and pathological complete response is not achieved for all patients (9). The present study attempted to obtain further clinical information in addition to diagnosis using radiological examination, to clarify the correlation of response rates between the breast tumor and the metastasized lymph node by neoadjuvant chemotherapy, and to identify the clinical significance of the response rates among breast cancer patients with axillary lymph node involvements.

Material and Methods

The material consisted of 98 patients diagnosed with node-positive breast cancer who received neoadjuvant chemotherapy before surgery at our institution. All were women and had no previous history of malignant disease. The study was approved by our institutional review board, and all patients provided written informed consent. A core biopsy of the tumor was performed before upstart of neoadjuvant chemotherapy in order to obtain adequate tumor tissue for histological diagnosis and determination of histological grade, hormone receptor status and human epidermal growth factor receptor type-2 (HER2) over-expression assessed by immunohistochemical staining. Histological grade was assigned according to the Nottingham grading system (10). Axillary ultrasound was performed to investigate the presence of metastasized lymph nodes. Roundish or irregularly shaped lymph nodes were classified as positive nodes. Perinodal infiltration, thickened cortices and presence of central necrosis were also considered to be associated with node-positivity. Fine-needle aspiration biopsy (FNA) under ultrasound guidance was performed on any lymph nodes with suspicious findings, and lymph node metastases were diagnosed cytologically. Patients with cytologically proven metastasized lymph nodes were recruited in this study, regardless of their tumor size.

CT examination was performed before and after neoadjuvant chemotherapy with a multidetector row helical CT (MDCT) scanner (Aquillion ONE or 64; Toshiba Medical Systems, Tochigi, Japan). Each patient was placed in the supine position with the arms positioned upward. The parameters were 120 kV and 150 mA using a detector collimation of 64 × 0.5 mm (pitch factor 0.7). Scans were started 60 s after intravenous injection of contrast media (Omnipaque, Daiichi Sankyo Co., Tokyo, Japan) at a rate of 3 mL/s via an intravenous catheter inserted in an antecubital vein on the contralateral side. A single breathhold was requested during scanning in every case. The axial images were reconstructed at 2-mm slice thickness.

The intended treatment for all patients was four cycles of an anthracycline containing regimen followed by four cycles of taxan every three weeks. The response to therapy was evaluated by the change in the largest dimensions of the breast mass and regional lymphadenopathy measured on CT axial images before and after neoadjuvant chemotherapy. The percent reduction rate was calculated as a response rate according to the guidelines of Response Evaluation Criteria in Solid Tumors (RECIST) group (11). Definitive surgical treatment was achieved after completion of neoadjuvant chemotherapy from 2004 to 2008. The excised breast tumors were fixed in 10% formalin, cut into multiple serial sections approximately 5 mm thick, paraffin embedded as tissue blocks, stained using hematoxylin and eosin (HE) and evaluated by a pathologist. All surgically excised lymph nodes were also bisected and evaluated using HE staining. In the cases of indefinite findings concerning metastasis, immunohistochemical examinations were performed using the avidin-biotinylated peroxidase complex technique with mouse monoclonal antibody against cytokeratin (Dako Cytomation Co., Kyoto, Japan). A pathological complete response in this study was defined as disappearance of invasive lesions, including absence of lymph node involvement (12). All patients treated with partial resection received whole breast irradiation after surgery. Hormonal treatment was recommended to cases positive for estrogen receptor, and consisted of tamoxifen for premenopausal patients and anastrozole for postmenopausal patients. Trastuzmab was used for the cases with HER2 overexpression. Following these treatments, tri-monthly physical evaluations were performed, and diagnosis of recurrence was confirmed by radiographic and/or ultrasonographic investigation.

The correlation between the response rate by of neoadjuvant chemotherapy and patient outcome was analyzed retrospectively. Statistical differences were determined using the Student's t-test for continuous variables and the chi-square test for categorical variables. Comparison between the response rates on breast tumor and axillary nodes was made using linear regression analysis. Kaplan-Meier curves were generated for presenting DFS, and evaluated by the Logrank test. P values of less than 0.05 were considered to indicate statistically significant differences.

Results

Patients' ages ranged from 31–74 years with a mean of 54 years. All the patients had invasive histology (93 invasive ductal, three invasive lobular, one apocrine, and one squamous carcinoma) with nodal involvements confirmed cytologically by FNA. The pre-chemotherapeutic mean maximum breast tumor dimension was 3.2 cm (range 0.7–11.9 cm). Metastasized lymph node was observed with mean size of 1.6 cm (range 0.5–4.7 cm). Sixty cases (61.2%) had positive estrogen receptor, 27 cases (27.6%) presented over-expression of HER2, and 20 cases (20.4%) were classified as high histological grade. Definitive breast surgery was performed for all patients, and consisted of modified radical mastectomy in 49 cases (50%) and breast conservative surgery with sequential axillary lymph node dissection in the remaining 49 cases. The response rates of the breast tumors and lymph nodes in all patients were 58.2% and 40.8%, respectively (Figs. 1 and 2). No residual cancer cells in the lymph nodes were observed in 45 patients (45.9%). A pathological complete response was noted in 13 (13.3%) patients.

OPEN IN VIEWER

Fig. 1

CT axial images of breast tumor (a) before and (b) after chemotherapy (arrow)

OPEN IN VIEWER

Fig. 2

CT axial images of axillary lymph node (a) before and (b) after chemotherapy (arrow)

The breast tumor response rates were positively and statistically significantly correlated with those of the lymph nodes (R = 0.553, y = 22.8 + 0.87x, p < 0.001, Fig. 3). Twenty-five recurrences and five deaths were observed in the median 48 (16–75) months follow-up. Eleven cases initially recurred in their locoregional sites and the other cases developed distant metastases in sites including the lung, brain, liver and bone. There were no significant differences between relapse-free and recurrence cases with regard to estrogen receptor status and surgical procedure (Table 1). Relapse-free cases showed a 63.1 ± 31.5% (mean ± standard deviation) tumor response rate and recurrence cases showed a 44.0 ± 43.7% response rate (p = 0.021, Fig. 4). The nodal response rate in relapse-free cases was significantly higher than in recurrence cases (46.6 ± 20.9%, 24.1 ± 19.9%, respectively, p < 0.001). Eighteen out of 38 estrogen receptor-negative patients (47.7%) developed recurrences compared to only seven of 60 (11.7%) estrogen receptor-positive patients (p < 0.001). However, among estrogen receptor-negative patients, recurrence cases also had lower response rates than relapse-free cases (24.7± 18.4% and 57.6± 18.2%, respectively, p < 0.001). Cancer-associated deaths were observed more frequently in cases with lower breast response rates (16.5 ± 65.3%, p = 0.007) and lower nodal response rates (18.1 ± 15.2%, p = 0.021), compared with the breast and nodal response rates in surviving patients (60.5± 32.6%, 42.0± 22.5%, respectively, Fig. 5). When 40% of the mean response rates in metastasized lymph nodes was chosen as the cut-off value, a significant difference was found between the high and low response cases (p = 0.001, Fig. 6).

OPEN IN VIEWER

Fig. 3

Correlation of response rates between breast tumors and metastasized lymph nodes

OPEN IN VIEWER

Fig. 4

Recurrence and response rates of (a) breast tumors and (b) lymph nodes

OPEN IN VIEWER

Fig. 5

Cancer associated death and response rates of (a) breast tumors and (b) lymph nodes

OPEN IN VIEWER

Fig. 6

Disease-free survival curves according to nodal response

Discussion

Several modalities are used for breast cancer management. Magnetic resonance imaging (MRI) is a sensitive method for breast cancer detection and has recently emerged as a promising tool in breast cancer management. Ultrasound (US) has also long been known to be a reliable modality for the diagnosis of breast lesions (13). Recent advances in US equipment and the refinement of breast imaging techniques have improved both detection and characterization of breast lesions. However, according to the RECIST guideline (11), US is not useful for assessment of lesion size and should not be used as a method of measurement. US examinations cannot be reproduced, because they are operator-dependent and it cannot be guaranteed that the same technique and measurements will be taken from one assessment to the next. If new lesions are identified by US, confirmation by CT or MRI is recommended. We have previously reported a significant correlation between the tumor reduction after neoadjuvant chemotherapy and enhancing MRI pattern of breast lumps as well as patients' outcome (14). These radiological examinations are considered to bring further information, not only the use for ordinary clinical diagnosis. When precisely estimated, these imaging methods seem to bring the same prognostic information. In this series, the breast lumps and lymph nodes had to be evaluated at the same time. CT is able to measure both lesions simultaneously on sequential slices obtained by same protocol. Therefore, CT was selected as the modality for the assessment of neoadjuvant chemotherapy effect in the present study.

A MDCT examination covering a large body area can be completed within a few seconds, and good image quality can be easily achieved without using complicated technical parameters. Concerning breast cancer management, MDCT is a feasible imaging modality: it has high spatial resolution and can offer accurate diagnoses of distant metastasis, intraductal extension, and axillary nodal status (15). The positive predictive value for diagnosing nodal status on the basis of CT findings has been reported to range from 72% to 95% (16). In the present series, the addition of axillary ultrasound with FNA was used to document nodal involvement by a less invasive approach.

The presence or absence of residual cancer after neoadjuvant chemotherapy has been proposed as a critical prognostic factor for prolonged disease-free survival and overall survival (17, 18). The National Surgical Adjuvant Breast and Bowel Project B-18 trial (4) demonstrated that the outcome for patients who achieve a pathological complete response continues to be superior to that of patients with residual invasive cancer on pathologic examination or to the outcome of patients who do not achieve a pathological complete response. At nine years follow-up, overall survival for patients achieving a pathological complete response was 85% and 73% for patients with residual invasive cancer. For prolonged disease-free survival, the respective rates were 75% and 58%. Moreover, a number of trials in node-positive breast cancer patients have shown that addition of a taxan to anthracycline-based adjuvant chemotherapy, either sequentially or concurrently, increases disease-free survival and overall survival (19, 20). These neoadjuvant chemotherapy protocols have increased the pathological complete response rate from 13% to 26%, increased the proportion of patients with negative axillary nodes at the time of surgery and it has been widely anticipated that increased survival will also result (21, 22).

Several studies have reported the relationship of clinical and pathologic nodal response to neoadjuvant chemotherapy and outcome of locally advanced breast cancer. Fifty percent of the pathological complete response patients had positive axillary lymph nodes compared to 76% of patients who did not have a pathological complete response (23). Rouzier et al. (8) analyzed the pathological complete response of axillary lymph nodes after neoadjuvant doxorubicin plus cyclophosphamide (AC) chemotherapy in 152 patients with cytologically proven axillary metastases. They first showed that pathological complete response of lymph nodes was a stronger prognostic predictor than that of the main tumor in terms of disease-free survival. Hennessy et al. (24) also analyzed 403 patients who were treated with neoadjuvant AC, partly with paclitaxel, and showed that nodal pathological complete response was a better independent predictor for overall survival than the pathological complete response of the main tumor. Pathologic lymph node status is considered as a strong predictor of outcome in both preoperative and postoperative patients.

Patient outcome versus measured microresidual metastases among lymph nodes after neoadjuvant chemotherapy was previously evaluated using multislice sectioning (25). Disease-free survival and overall survival in patients with pathological complete response or ≤0.2 mm nodal metastases was significantly better than that with >0.2 mm. Kilbride et al. (26) classified patients based upon nodal response after neoadjuvant chemotherapy into node-negative, downstaged and persistently positive groups. The downstaged group had lower rates of recurrence than those with residual axillary disease, with a significantly lower risk of distant failure. The status of nodal involvement at the time of presentation and after delivery of neoadjuvant chemotherapy in patients with breast cancer is prognostic of treatment failure.

In the present study, the correlation between response rate (including nodal response) and occurrence of early events was evaluated in node-positive breast cancer patients. The response rates of the breast tumors were statistically correlated with those of lymph nodes. Disease-free cases had a greater response rate than recurrence cases, regardless of tumor size, histological grade and HER2 amplification. Cancer associated death was observed more frequently in cases with lower response rates compared to surviving patients. Moreover, a prognostic difference was found between responders and non-responders, especially concerning nodal response rates. These clinical responses could be used as a surrogate marker for evaluation of the efficacy of neoadjuvant chemotherapy before assessment of the pathological response. Four factors were associated with pathological complete response: higher tumor grade, lack of estrogen receptor expression, HER2 amplification, and negative lymph node status (27). Estrogen receptor-negative disease has been noted by numerous authors to correlate with improved response rate to neoadjuvant chemotherapy (28, 29). Despite a possible association between estrogen receptor-negative disease and chemotherapy response, estrogen receptor-negative disease strongly correlated with a poorer survival (30). The patients without estrogen receptor expression more frequently developed recurrences than those with estrogen receptor-positive disease in the present series, however, the recurrence cases also had a lower response rate than disease-free cases among estrogen receptor-negative cases.

In conclusion, the present series evaluated the therapeutic effect on breast tumors and metastasized lymph nodes following neoadjuvant chemotherapy, and demonstrated a significant correlation between the response rate to neoadjuvant chemotherapy and outcome in early phase among patients, especially with an anthracycline containing regimen followed by taxan administration. Further clinical information, in addition to diagnosis using radiological examination, are obtained by evaluating the response rate. If a suitable borderline value can be established, evaluating the response rate measured by imaging could be used as a surrogate marker for prognosis before assessment of the pathological response which is ordinarily obtained after surgery, and could play a more important role in planning individualized treatment.

Conflict of interest: None.