Abstract
Background/Aim: To develop an accurate method to predict nodal pathological complete response (ypN0) in patients after neoadjuvant chemotherapy (NAC) for clinically node-positive breast cancer. Patients and Methods: We included 128 patients with clinically node-positive primary breast cancer who underwent axillary lymph node dissection after NAC. Results: Breast primary tumor clinical complete response (ycT0) was observed in 29.7% and nodal clinical complete response (ycN0) in 44.5% of cases. When ycN0 was predicted as ypN0, the negative predictive value was 77.2%, and the false-negative rate was 19.7%. Estrogen receptor status, ycT0, and ycN0 were independent predictive factors for ypN0 after NAC in patients with clinically node-positive breast cancer. These factors were used to develop a nomogram for ypN0 prediction. The following points were added: 82 in case of estrogen receptor-negative, 56 in case of ycT0, and 100 in case of ycN0. Score summation was used to prognosticate the manifestation of ypN0. Our nomogram predicted ypN0 with a negative predictive value of 92.9% and false-negative rate of 4.5%, demonstrating an approximate 15% improvement over ypN0 prediction using ycN0 alone. Conclusion: Estrogen receptor-negativity, ycT0, and ycN0 are independent predictive factors for ypN0 after NAC in clinically node-positive breast cancer. The nomogram may improve individualized axillary treatment.
- Breast cancer
- clinically node-positive
- neoadjuvant chemotherapy
- pathological complete response
- prediction nomogram
Neoadjuvant chemotherapy (NAC) has transitioned from being a primary treatment for patients with inoperable breast cancer to becoming a viable strategy for those with operable breast cancer (1). The principal aims of contemporary NAC include prediction of patient prognosis based on tumor response, formulation of postoperative therapeutic plans, and mitigation of surgical invasiveness through tumor size reduction (2). Based on these benefits, patients with both clinically node-negative (cN0) and clinically node-positive (cN+) breast cancer are increasingly receiving NAC (3-5). Nevertheless, the optimal strategy for axillary surgery after NAC remains controversial, primarily due to the typically higher rates of pathological complete response (pCR) observed in axillary lymph node metastases compared with those observed in primary breast tumors (6-8)
Axillary lymph node dissection (ALND) has traditionally served as the standard surgical intervention for breast cancer (9). However, it may cause complications, including lymphedema, restricted arm movement, and sensory impairments, which have adverse effects on quality of life (QOL). Consequently, there is growing interest in assessing the viability of sentinel lymph node biopsy (SLNB) as an alternative approach (10, 11). The National Comprehensive Cancer Network Breast Cancer guidelines (12) recommend SLNB for patients with cN+ who become cN0 after NAC. However, the false-negative rate (FNR) of SLNB is >10%. Presently, although the remission of axillary node metastases following NAC treatment for cN+ breast cancer is anticipated, no approach for avoiding ALND has been confirmed. In contrast, the axillary pCR rate to NAC for cN+ breast cancer has been reported to be approximately 40% (13). Therefore, it is necessary to develop a more accurate method for predicting nodal pCR (ypN0) to identify patients who could be exempted from ALND after NAC for cN+ breast cancer.
The aim of this retrospective study was to identify predictive factors for ypN0 after NAC for cN+ breast cancer and to develop a prognostic model to more accurately predict ypN0 after NAC for cN+ breast cancer.
Patients and Methods
Patients. This was a single-institution, retrospective clinical study that included 128 patients with cN+ primary breast cancer who underwent ALND after NAC between January 2014 and July 2021 at Saitama Medical Center. The study was approved by the Institutional Review Board of Saitama Medical Center (approval number: 19-23) and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The requirement for written informed consent was waived owing to the retrospective nature of the study.
Treatment plan. Patients were scheduled to receive four cycles of taxane (docetaxel or nab-paclitaxel) followed by four cycles of epirubicin + cyclophosphamide (14, 15). Trastuzumab and pertuzumab were administered concomitantly on Day 1 as a 3-week cycle for four cycles during taxane administration in patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer (16). The patients underwent surgery 4-8 weeks after completion of NAC.
Assessment. Diagnostic modalities, such as ultrasonography, contrast-enhanced magnetic resonance imaging (MRI), contrast-enhanced computed tomography (CT), or positron emission tomography (PET)-CT, along with cytology, were employed for the diagnosis of clinical axillary lymph node metastasis.
Based on the Response Evaluation Criteria in Solid Tumors (version 1.1) (17), clinical response was evaluated within 4 weeks prior to surgery using CT and MRI. Clinical complete response (cCR) was defined as the disappearance of all tumor foci after the completion of chemotherapy.
The pCR was defined as no microscopic evidence of residual invasive and non-invasive carcinoma in the breast or axillary lymph nodes (18). Both clinical and pathological responses were evaluated independently in the primary tumor and axillary lymph nodes. The cCR and pCR of the primary tumor were defined as ycT0 and ypT0, respectively. The cCR and pCR of the axillary lymph nodes were defined as nodal clinical complete response (ycN0) and ypN0, respectively.
The statuses of the estrogen receptor (ER) and progesterone receptor (PgR) were determined by immunohistochemistry (IHC) utilizing the Allred score (19). HER2 status was assessed using IHC and fluorescence in situ hybridization (FISH). Tumors with an IHC score of 3+ based on IHC, or a gene amplification score of ≥2 based on FISH, were classified as HER2-positive.
Tumors were classified as luminal (ER− or PgR-positive and HER2-negative), HER2-positive (ER− and PgR-positive/negative and HER2-positive), and triple-negative (TN) (ER− and PgR-negative and HER2-negative) subtypes.
Statistical analysis. Statistical analysis was performed using SPSS version 29.0 (SPSS Inc., Chicago, IL, USA). The clinicopathological variables were compared using Fisher’s exact test, a Chi-square test, and logistic regression models, where appropriate. A p<0.05 was considered statistically significant. Multiple comparisons were adjusted using the Bonferroni correction. Receiver operating characteristic (ROC) curves were generated using statistical software, and Youden’s index was employed to identify the optimal cut-off points, defined as the maximum value of (sensitivity + specificity − 1), thereby directly assessing the highest overall diagnostic accuracy (20).
Results
Patient characteristics. The median age at diagnosis was 52 (range=31-79) years; 48.4% (62/128) of the patients were premenopausal women. Of the 128 patients, 38.3% (49/128) had T3 or higher, 22.7% (29/128) had cN2 or higher, and 46.0% (49/128) had stage III disease. NAC included sequential taxanes and anthracyclines in 47.7% of cases. Anti-HER2 therapy was used in all cases of HER2-positive breast cancer (Table I).
Study patient characteristics.
Clinical and pathological response in cN+ breast cancer patients treated with NAC. ycT0, ypT0, ycN0, and ypN0 were detected in 29.7%, 28.9%, 44.5%, and 48.4% of patients, respectively. The ypN0 rate was 19.5% higher across all cases (p<0.001), 16.6% higher in the luminal subtype (p=0.009), 4.5% higher in the TN subtype (p=0.002), and 32.5% higher in the HER2-positive subtype (p=0.027) than the respective ypT0 rates. Additionally, the ypN0 rate was 3.9% higher across all cases, 7.6% lower in the luminal subtype (p<0.001), 13.6% higher in the TN subtype (p<0.001), and 17.5% higher in the HER2-positive subtype (p=0.006) than the respective axillary ycN0 rates (Figure 1).
Differences in cCR and pCR rates between the primary tumor and axillary lymph nodes according to subtype. Data were compared using Fisher’s exact test. A p<0.05 was considered statistically significant. Multiple comparisons were adjusted using the Bonferroni correction. Asterisks were added to the pairs that showed significant differences. cCR, Clinical complete response; pCR, pathological complete response; HER2, human epidermal growth factor receptor type 2; n.s., not significant.
When predicting ypN0 based solely on the disappearance of clinical axillary lymph node metastasis observed during imaging diagnosis, the sensitivity was 80.3%, the specificity was 71.0%, positive predictive value (PPV) was 74.6%, and the negative predictive value (NPV) was 77.2% (Table II).
Correlation between predictive ycN0 and ypN0 in clinically node-positive breast cancer treated with NAC.
Univariate and multivariate analyses of the relationship between preoperative clinicopathological factors and ypN0 in cN+ breast cancer. In the analysis of the relationship between ypN0 status and preoperative clinicopathological factors in cN+ breast cancer patients treated with NAC, univariate analysis revealed significant differences in ER negativity [hazard ratio (HR)=0.186, 95% confidence interval (CI)=0.086-0.401, p<0.001], PgR negativity (HR=0.360, 95% CI=0.176-0.736, p=0.005), HER2 positivity (HER2+) (HR=4.394, 95% CI=1.941-9.949, p<0.001), ycT0 (HR=10.667, 95% CI=4.020-28.304, p<0.001), and ycN0 (HR=9.966, 95% CI=4.399-22.576, p<0.001). Multivariate analysis identified ER status, ycT0, and ycN0 as independent predictive factors for ypN0 following NAC in patients with cN+ breast cancer (Table III).
Univariate and multivariate analyses of the relationship between preoperative clinicopathological factors and pathologically axillary node-negativity after NAC in clinically axillary node-positive breast cancer.
Nomogram predicting ypN0 in cN+ breast cancer patients treated with NAC. The independent factors identified in the multivariate analysis were used to develop a nomogram for predicting ypN0. The following points were added: 82 for ER-negative patients, 56 for ycT0 patients, and 100 for ycN0 patients. The total score was used to predict whether the patient had ypN0 (Figure 2A). When the ROC curve was used to determine the cut-off value of the total points, ≥119 points predicted ypN0 with an NPV of 92.9% and an FNR of 4.5%, demonstrating an approximate 15% improvement over ypN0 prediction using ycN0 alone (Figure 2B-C, Table IV).
Development of the nomogram to improve the prediction of ypN0 in cN+ breast cancer treated with NAC. (A) Nomogram to predict the probability of ypN0. (B) ROC curve. AUC=0.873 (95% CI=0.811-0.935). (C) Improvement of ypN0 prediction using the nomogram. AUC, Area under the curve; cCR, clinical complete response; CI, confidence interval; ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; FNR, false-negative rate; NAC, neoadjuvant chemotherapy; NPV, negative predictive value; pCR, pathological complete response; ROC, receiver-operating characteristic; ypN0, nodal pathological complete response.
Correlation between nomogram and ypN0 prediction in cN+ breast cancer treated with NAC.
Discussion
Individualized surgical treatment and minimally invasive surgery have emerged as focal points in contemporary breast cancer care. Presently, efforts are underway to ameliorate the FNR of SLNB following NAC in patients with cN+ breast cancer. However, this endeavor relies on the assumption that the disappearance of axillary lymph node metastases due to NAC can be predicted. Therefore, the development of an accurate method to forecast the clearance of axillary lymph node metastasis after NAC in patients with cN+ breast cancer holds significance, facilitating the safer omission of ALND. This is particularly crucial because it promises to enhance the QOL of affected individuals. The present study endeavored to identify predictive factors for achieving ypN0 status after NAC in cN+ breast cancer and to develop a novel nomogram to enhance the accuracy of ypN0 prediction.
In this study, the overall ypT0 rate was 28.9%, whereas that of ypN0 was 48.4%. Notably, the ypN0 rate exceeded that of the primary tumor by approximately 20%. Specifically, in TN and HER2-positive breast cancer, the ypN0 rates were higher (TN: 54.5%; HER2: 72.5%) than those of ycT0 (TN: 50.0%; HER2: 40.0%). Previous studies (21-25) have consistently indicated higher ypN0 rates in TN and HER2-positive breast cancer subtypes compared with those in hormone receptor-positive HER2-negative subtypes. Recently, Chen et al. (26) reported the disparity between ypT0 and ypN0 rates in 968 patients with cN+ breast cancer who underwent NAC. ycT0 combined with non-ypN0 constituted 3.2% (31/968) of the total cohort, whereas ypN0 combined with non-ypT0 accounted for 25.3% (245/968) of the total cohort. Notably, non-ypT0 combined with ypN0 exceeded 20% across all breast cancer subtypes, which was particularly evident in TN and HER2-positive subtypes, consistent with our findings. Consequently, these findings underscore the necessity for a predictive model specifically tailored to ypN0 to enable more precise prediction of the feasibility of ALND omission following NAC in patients with cN+ breast cancer.
In this study, the sensitivity for predicting ypN0 based only on the presence or absence of ycN0 after NAC during imaging diagnosis was 80.3%, with a specificity of 71.0%. Additionally, a systematic review and meta-analysis encompassing the diagnostic efficacy of non-invasive imaging in evaluating ypN0 after NAC in cN+ breast cancer has been documented. The sensitivities and specificities of ypN0 on axillary ultrasound, breast MRI, and whole-body PET-CT were 77%, 78%, and 78% and 50%, 58%, and 49%, respectively (27). This meta-analysis underscores the limitations of relying solely on imaging modalities to accurately predict ypN0 following NAC in patients diagnosed with cN+ breast cancer.
Studies investigating clinicopathological factors associated with achieving ypN0 status in cN+ breast cancer have been documented. Recently, Choi et al. (28) conducted a multivariate analysis and reported that clinical T stage, PgR status, HER2 status, and clinical nodal response were independent predictors of ypN0 status. In this study, ER negativity, ycT0, and ycN0 were independent predictive factors for ypN0 after NAC in patients with cN+ breast cancer. Notably, our findings, along with those of previous studies, consistently identified hormone receptor negativity as a significant independent factor for ypN0 status. Interestingly, in our study, HER2 status was not an independent predictor of ypN0 status. This may be attributed to the higher ypN0 rates in both TN and HER2-positive subtypes compared with those in the luminal subtype. Consequently, it can be inferred that HER2 status does not independently predict ypN0 status, possibly owing to the high ypN0 rates observed in both TN and HER2-positive subtypes. Prior reports (21-25) on ypN0 rates following NAC for cN+ breast cancer have consistently shown superior ypN0 rates in HER2-positive (49-82%) and TN (47-73%) subtypes compared with those in the luminal subtype (0-46%). Furthermore, the non-luminal subtype has been found to be significantly associated with a poor prognosis and a high pCR rate after NAC (29, 30). Therefore, the above observations suggest that at least hormone receptor negativity is strongly associated with ypN0.
The nomogram reported by Choi et al. (28) exhibited a sensitivity of 91.2%, an NPV of 96.7%, an FNR of 8.8%, and an accuracy of 72.2%. Conversely, our nomogram demonstrated a sensitivity of 95.5%, an NPV of 92.9%, an FNR of 4.5%, and an accuracy of 79.7%, which were almost comparable to those of the nomogram in the Choi et al. (28) study. Notably, our nomogram exhibited slightly better sensitivity and overall accuracy. Several factors may account for the discrepancy in nomogram prediction accuracy between their study and ours. First, their study conducted a multivariate analysis using both ypT0 and ypN0 as dependent variables. In their study, the ypN0 rate was 47.3% and the ypT0 rate was 24.3%. Consequently, a proportion of patients with non-ycT0 but ypN0 may have contributed to the diminished sensitivity of ypN0 prediction by the nomogram. Second, their study included clinical T stage as a factor in the nomogram. Typically, the size of the clinical primary tumor in the setting of NAC is more strongly associated with tumor disappearance in the primary tumor rather than in the axillary lymph nodes. Therefore, the inclusion of clinical T stage may have indirectly influenced the ypN0 prediction accuracy of the nomogram. Third, approximately 90% of the patients with HER2-positive breast cancer in their study did not receive anti-HER2 therapy. The ypT0ypN0 rate in cN+ and HER2-positive breast cancer in their study was 29.5%. However, the current well-established notion is that the combination of anti-HER2 therapy and chemotherapy significantly increases the ypN0 rate of HER2-positive breast cancer compared with chemotherapy alone (31, 32). Furthermore, a systematic review and meta-analysis (33) reported ypN0 rates of 45-59% in cN+ and HER2-positive breast cancer. Therefore, considering that the current standard treatment for operable HER2-positive breast cancer entails a combination of chemotherapy and anti-HER2 therapy, our nomogram could potentially be more useful in accurately predicting ypN0. Most significantly, both their study and ours show that predicting ypN0 using a nomogram is more accurate than using imaging findings alone. It is expected that this nomogram may contribute to the improved prediction accuracy of ypN0 in cN+ breast cancer.
This study has several limitations, including its retrospective design, small sample size, and variability in the statuses of lymph node metastases, stages, and NAC regimens. Nevertheless, the nomogram developed in this study is valuable because it allows a more accurate prediction of ypN0 than when assessing ycN0 from imaging findings alone. Studies to predict ypN0 after NAC in cN+ breast cancer are anticipated in the future.
Conclusion
The TN and HER2-positive subtypes of cN+ breast cancer are associated with higher rates of ypN0 than ycN0 after NAC, compared to the luminal subtype. We identified ER-negative, ycT0, and ycN0 as independent predictive factors for ypN0 in cN+ breast cancer patients treated with NAC. The nomogram incorporating these factors to predict ypN0 may contribute to improved individualized axillary treatment.
Footnotes
Authors’ Contributions
Drafting and writing of the manuscript: H. Seki; Literature search and analysis: H. Seki and Y. Ishiguro; Data extraction: Y. Ishiguro, Manuscript editing: A. Makino and K. Yamaguchi; Supervision: S. Imoto.
Conflicts of Interest
The Authors have no conflicts of interest to declare.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
- Received September 7, 2024.
- Revision received September 27, 2024.
- Accepted October 11, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
