Main

Breast cancer is among the most deadly malignancies in developed countries, with metastatic spread being the major reason behind this fatal outcome.1 According to the European Society for Medical Oncology guidelines, the number of metastasized lymph nodes is a fundamental prognostic factor in breast cancer, constituting a cornerstone for multimodal and combined treatment options, namely chemotherapy, immunotherapy, radiotherapy, and surgery.2, 3, 4 Notably, the biology within the lymph node itself has been completely undervalued regarding clinical decisions. In fact, the molecular characteristics of primary tumors form the basis for the classification into one of four molecular subtypes (Luminal A, Luminal B, epidermal growth factor receptor (HER) 2-overexpressing carcinomas and basal-like carcinomas), which is highly correlated with patient prognosis and set the course for adjuvant treatment options.5, 6, 7, 8, 9, 10 Nevertheless, despite that many countries in the world today depend on this classification, independently of the molecular profile of the primary tumor, ~6–10% of the patients show metastatic disease at diagnosis and systemic recurrence develops in 25–30% of the patients treated with curative intent.11, 12, 13 In recent years, it is becoming clear that having axillary biomarkers with prognostic value would be crucial to support clinical decisions, as there is still a very heterogeneous behavior regarding disease progression in patients within the same molecular subtype or in patients with lymph node involvement, some of which never develop distant metastasis.14, 15, 16 Several studies recognize that metastases exhibit a gene expression profile that differs from the originating tumor, with variations already accumulating in the lymph nodes.17, 18, 19, 20, 21, 22, 23 Gene copy number profiling and gene expression analysis comparing primary tumors and metastases from the same patient have revealed interesting genetic differences, pointing for new putative prognostic biomarkers.19, 20, 21, 22, 23 Understanding this heterogeneity constitutes a potential critical point for predicting patient prognosis. In fact, some immunohistochemical studies have addressed the expression of biomarkers in axillary lymph nodes and in distant metastasis, focusing mainly in the classical markers that define treatment options and allow carcinoma subtype classification, such as ER, progesterone receptor (PR), Ki67, epidermal growth factor receptor 1 (EGFR), and 2 (HER2).4, 24, 25, 26, 27, 28 Importantly, most studies lack prognostic information for individual patients or are limited to a small cohort.

Accumulating evidence indicates that local recurrences and/or distant metastases originate from rare tumor cells that may function like stem cells in their ability to initiate, propagate and hierarchically organize secondary tumors.29 These breast cancer stem cells are shown to exhibit unique growth abilities including self-renewal, differentiation potential and resistance to most anti-cancer agents, including chemo- and/or radiotherapy, which contribute to the overall aggressiveness of recurrent or metastatic lesions.30, 31 Thus, efforts are needed to define sensitive and reliable breast cancer stem cell biomarkers for metastatic lesions.

In pathology studies, CD44 and CD49f are among the most extensively used breast cancer stem cell markers. CD44 is the transmembrane receptor for hyaluronic acid and is highly involved in cell–cell and cell–matrix interactions.32 A subpopulation of tumor cells strongly expressing CD44, but low levels of CD24 (the CD44+CD24−D24b phenotype), was identified for the first time as a population of breast CSCs by Al-Hajj et al.33 and this evidence was subsequently confirmed by several other authors. Specifically, the CD44+ phenotype is correlated positively with colon, breast, prostate, and pancreatic cancer initiating cells.33, 34, 35, 36 Specifically in breast cancer, CD44 has been demonstrated as a marker of cancer cells with a basal-like phenotype, being present in poor prognosis tumors, and associated with more motile and invasive cancer cells.37 Concerning CD49f (also known as α6 integrin, involved in the cell interaction with laminin at the hemidesmosome), it has been extensively used to purify mouse and human mammary stem cells,38, 39, 40 as well as to characterize a population of human cancer cells with stem-like behavior.41, 42 Recently, Ghebeh et al.43 showed that the combination of CD44high/CD24low/CD49f+ might significantly improve and strengthen the measurement of breast cancer stem cells with significantly higher stem/progenitor ability and the expression of both CD44 and CD49f in breast primary carcinomas was associated with worse clinical outcome and tumor aggressive features.41, 44

In addition, our group showed that P-cadherin also behaves as a breast cancer stem cell marker in human breast carcinomas. P-cadherin is a calcium-dependent cell–cell adhesion molecule, which expression was found to be significantly associated with CD49f and CD44 expression,45, 46 therapy resistance,45 and cancer cell invasion.47, 48 Furthermore, P-cadherin mediates stem cell properties, namely by increasing the cell growth in anchorage-independent conditions, by activating α6β4 integrin signaling, and by inducing tumorigenic ability in nude mice.45, 46 In fact, we have identified P-cadherin as an independent indicator of poor prognosis in primary breast tumors,49 being associated with high histological grade tumors and negative hormonal receptors status.49, 50, 51, 52, 53 P-cadherin overexpression is predominantly found in basal-like breast carcinomas49, 54 and is strongly linked to BRCA1 mutations.55 The tumor-promoting properties mentioned for P-cadherin in breast cancer, namely its invasive ability, are found in several other tissue contexts, suggesting a putative role for this molecule in the metastatic process.56

In this study, we addressed, for the first time, the expression of P-cadherin, CD44, and CD49f in primary breast carcinomas and matched axillary loco-regional metastases, in order to evaluate their impact in patients’ survival. P-cadherin arises as an independent indicator of prognosis in the metastatic setting and as a putative useful biomarker for axillary-based breast cancer decisions in clinical practice.

Materials and methods

Patient Selection and Material Characterization

A series of 135 formalin-fixed paraffin-embedded cases of invasive breast carcinomas, including primary tumors and matched lymph node metastases, was obtained from the Pathology Departments of Hospital Xeral-Cíes (Vigo, Spain) and Hospital São João (Porto, Portugal), under patient informed consent and with ethical approval by both hospitals Ethical Commissions. Clinical and pathological features were retrieved for this study.

Patient follow-up information was available for 118 patients, which were diagnosed between 1978 and 1992, with a maximum follow-up of 120 months after diagnosis. These breast carcinoma patients followed the adequate protocols for chemotherapy, radiotherapy, and hormone therapy given at that time. All patients have been treated with adjuvant chemotherapy, which consisted of a protocol of six cycles of cyclophosphamide, methotrexate, and fluorouracil. Patients with ER positive tumors have been treated with hormonal therapy, which was carried out exclusively with tamoxifen. Patients with HER2-overexpressing carcinomas were not treated with specific targeted therapy (trastuzumab). No neo-adjuvant treatment has been used in the patients included in this series.

The disease-free survival interval was defined as the time from the diagnosis to the date of breast cancer-derived relapse, whereas overall survival was considered as the number of months from the diagnosis to the disease-related death. For each primary tumor, the axillary lymph nodes were embedded in paraffin and carefully evaluated by a pathologist, which identified the lymph nodes with metastatic disease. Each breast cancer case was matched with the lymph node with the highest metastatic burden (macrometastasis).

The series included 26.7% of cases from women diagnosed before 50 years old and 73.3% from women with more than 50 years old. Concerning tumor size, 10.9% of the cases were classified as T1 (<2 cm), 72.8% as T2 (2–5 cm), and 16.3% as T3 (>5 cm). All cases were positive for lymph node metastasis, as this was the criterion to select the samples. Regarding histological grade, 8.9% of the cases were Grade I, 25.4% Grade II, and 65.7% Grade III. Breast carcinomas were classified considering European Society for Medical Oncology guidelines into luminal A carcinomas (ER/PR positive, HER2 negative, and Ki67 low), luminal B carcinomas (ER/PR positive, HER2 and/or Ki67 high), HER2-overexpressing carcinomas (ER/PR negative, HER2 positive), and triple-negative carcinomas (ER/PR/HER2 negative), according to the molecular classification used in the clinical setting.57 The Ki67 cutoff value to distinguish low from high proliferative tumors was 13.25%, based on a study from Cheang and colleagues.58 From the 135 breast cancer cases, 57.0% (77/135) were luminal A, 10.4% (14/135) were luminal B, 14.1% (19/135) were HER2-overexpressing, and 18.5% (25/135) were triple negative. Although the series included only lymph node positive patients, patient follow-up information revealed that clinical outcome varied considerably. Specifically, 10 years (120 months) after diagnosis, 48/118 (40.7%) patients had no relapse of the disease and 52/118 (44.1%) patients were alive (Supplementary Table S1).

The present study was conducted under the national regulative law for the usage of biological specimens from tumor banks, where the samples are exclusively available for research purposes in the case of retrospective studies. All analyses were performed according to the REporting recommendations for tumor MARKer prognostic studies (REMARK) recommendations for prognostic and tumor marker studies.

Immunohistochemistry

Immunohistochemistry was performed in 3 μm sections. The following antibodies were used: P-cadherin (BD Transduction Biosciences, USA; clone 56, 1:50), CD44 (Cell Signalling Technology, USA; clone 156-3C11; 1:100), and CD49f (Sigma-Aldrich, USA; HPA012696; 1:50). Epitope exposure was performed for 30 min using high-temperature antigen retrieval (95 °C) with Tris-EDTA buffer pH=9.0 (Novocastra, UK) for P-cadherin and CD49f, or citrate buffer pH=6.0 (Thermo Scientific, USA) for CD44. Primary antibodies were detected using the horseradish peroxidase polymer (Cytomation Envision System HRP; DAKO, USA) (for P-cadherin and CD49f) or the labeled biotin-streptavidin method (DAKO, USA) (for CD44), according to manufacturer’s instructions. Diaminobenzidine was used as chromogen.

Immunohistochemical Evaluation

The expression of P-cadherin, CD44, and CD49f was evaluated according to the grading system already described.45, 59, 60, 61, 62 In brief, staining was scored only when detected at the membrane of tumor cells and the staining extension was considered as follows: 0, 0–10% of positive tumor cells; 1+, 10–25% of positive tumor cells; 2+, 25–50% of positive tumor cells; 3+, >50% of positive tumor cells. The cases classified as (0) were considered negative, whereas (1+), (2+), and (3+) were scored as positive.

Statistical Analysis

Associations between clinicopathological parameters and the expression of the markers evaluated in primary tumors and loco-regional metastases were assessed by Pearson’s χ2- and Fisher’s exact tests. Survival analyses were estimated using the Kaplan–Meier method and compared using the log-rank test. For uni/multivariate survival analysis, Cox regression models were fitted to estimate hazard ratios (HRs) and the corresponding 95% confidence interval. Statistical analyses were carried out using SPSS statistics V.17.0 software (SPSS Inc., Chicago, IL), and a significance level of 5% was considered statistically significant.

Results

P-Cadherin is Associated with CD44 and CD49f Expression in Metastatic Lymph Nodes

A breast carcinoma series comprising 135 invasive carcinomas and matched lymph node metastasis was evaluated for the expression of the breast cancer stem cell markers P-cadherin, CD44, and CD49f by immunohistochemistry (Figure 1). P-cadherin membrane staining was found in 26.7% (36/135) of primary breast carcinomas and 23.0% (31/135) of metastasized lymph nodes. Regarding CD44, 47.4% (63/133) of primary tumors and 48.5% (64/132) of lymph node metastasis had a positive staining pattern. Considering CD49f, 14.3% (19/133) of primary tumors and 31.7% (40/126) of metastatic lymph nodes were scored as positive (Supplementary Table S2). Interestingly, CD49f displayed the major difference in expression comparing primary tumors and metastatic lymph nodes, which was found to be statistically significant (P=0.001).

Figure 1
figure 1

Breast cancer stem cell markers staining patterns. Representative images showing clear positive immunohistochemical staining for P-cadherin, CD44 and CD49f in primary tumors (ac, respectively) and in lymph nodes metastases (df, respectively). Scale=50 μm.

Further, the expression of breast cancer stem cell markers was correlated with classic prognostic factors and with the primary tumor molecular subtype. P-cadherin expression was able to significantly discriminate primary breast carcinomas with high histological grade (P=0.023) (Supplementary Table S3), confirming our previously reported data.51 Triple-negative breast carcinomas were significantly enriched in the expression of CD44 (P=0.009) and CD49f (P=0.007). P-cadherin expression was also enriched within the triple-negative molecular subtype, but the association was not statistically significant (P=0.105) (Supplementary Table S3). Notably, the expression of all three breast cancer stem cell markers in lymph node metastases was significantly associated with the triple-negative phenotype as well (P<0.001 for P-cadherin, P=0.007 for CD44 and P<0.001 for CD49f).

Interestingly, we found that P-cadherin expression in the primary tumor was also significantly associated with an increased number of metastasized lymph nodes: the majority of cases with a low number of metastasized lymph nodes (1–3 involved lymph nodes) were P-cadherin negative (89.4% (42/47)), whereas only 10.6% (5/47) were positive for P-cadherin. In contrast, a significant increase of P-cadherin was found in cases with 4–9 metastasized lymph nodes (36.5% (19/52)), as well as >10 metastasized lymph nodes (33.2% (10/31)) (P=0.009). We did not find this association regarding CD44 or CD49f (Supplementary Table S4).

Examining the association between the three markers in lymph node metastases, we observed that P-cadherin expression was significantly associated with CD44 expression (73.3% of P-cadherin positive cases were CD44 positive, P=0.003) and CD49f expression (59.3% of P-cadherin positive cases were CD49f positive, P=0.001) (Table 1). In contrast, in primary breast carcinomas, no statistically significant association was found between P-cadherin and CD44 expression, or between P-cadherin and CD49f. However, CD44 and CD49f were significantly associated in both primary tumors (P=0.001) and lymph node metastasis (P=0.002) (Table 1).

Table 1 Associations between the expression of P-cadherin, CD44, and CD49f in primary tumors and in metastatic lymph nodes

P-Cadherin Expression in Metastatic Lymph Nodes is Associated with Poor Clinical Outcome

We set out to explore whether evaluating the expression of breast cancer stem cell markers at diagnosis in patients with lymph node metastatic disease could allow the prediction of clinical outcome. Regarding the expression of P-cadherin, CD44, or CD49f exclusively in primary carcinomas, analysis of survival showed no significant association with overall survival or disease-free survival (Supplementary Figure S1). However, P-cadherin expression in tumor-associated lymph nodes impacted in the patient outcome, denoting a significantly worse disease-free survival and overall survival (Figure 2a, overall survival P=0.002, disease-free survival P=0.004).

Figure 2
figure 2

Analysis of breast carcinoma patients according to breast cancer stem cell markers expression in lymph nodes metastasis. Kaplan–Meier survival plots of overall survival and disease-free survival considering the expression of P-cadherin (a), CD44 (b), and CD49f (c) in metastasized lymph nodes in the series of invasive breast carcinomas. (d) Graphical representation of the percentage of cases with gain and loss of P-cadherin, CD44 and CD49f in metastatic lymph nodes vs primary carcinomas within each molecular subtype. For detailed expression, see Supplementary Table S3.

Considering CD44 and CD49f expression in metastatic lymph nodes, no significant impact was observed for the survival functions (Figure 2b and Figure 2c, overall survival P=0.796 and disease-free survival P=0.704 for CD44, and overall survival P=0.763 and disease-free survival P=0.303 for CD49f).

The Gain of P-Cadherin and CD49f Expression in Metastatic Lymph Nodes is Associated with the Triple-Negative Molecular Subtype

Breast cancer stem cell markers presented a clear distinction of expression (gain or loss) between primary tumors and matched lymph nodes metastasis, with 19.2% of the cases showing a discordant result for P-cadherin, 21.5% for CD44 and 26.2% for CD49f (Table 2).

Table 2 Frequency of cases according to the change in expression of breast cancer stem cell markers comparing invasive primary breast carcinomas with axillary lymph node metastasis

Concerning P-cadherin expression, the large number of cases that remain negative in both primary tumor and node metastases were classified as luminal A (61/88, 69.3.%). Interestingly, the majority of the cases with P-cadherin loss were also luminal A, representing 60.0% (9/15) of the cases. Regarding the ones that gained P-cadherin expression in the lymph nodes, these were mainly from the triple-negative molecular subtype, corresponding to 63.6% (7/11) of the cases (Supplementary Table S4) (P<0.001). Accordingly, the triple-negative molecular subtype was significantly enriched in the gain of P-cadherin expression: 38.0% (7/25) of triple-negative carcinomas gained P-cadherin expression, in comparison with only 8.14% (11/135) in the whole series (Figure 2d and Supplementary Table S5).

Regarding CD44, the cases that were negative in both matched neoplastic lesions were frequently classified as luminal A (36/54, 66.7%) (P=0.020) (Supplementary Table S5). We could observe that a high percentage of HER2-overexpressing tumors showed CD44 loss in the lymph nodes (22.2% (4/18)), in comparison with the whole series (10% (13/130)) (Figure 2d and Supplementary Table S5).

Regarding CD49f expression, 71.2% (57/80) of the cases that remained negative were scored as luminal A tumors. The ones where CD49f positivity was maintained were similarly distributed along the four molecular subtypes. As previously noticed, the most frequent alteration was the gain of CD49f expression in metastatic lymph nodes (Table 2 and Figure 2d) and, notably, this phenotype was significantly enriched in the triple-negative molecular subtype: 47.8% (11/23) of triple-negative carcinomas gained CD49f expression, in comparison with only 21.4% (27/126) in the whole series. None of the HER2-overexpressing tumors showed CD49f loss (P<0.001) (Figure 2d and Supplementary Table S5).

The Change in P-Cadherin Expression between Primary Carcinomas and Synchronous Metastatic Lymph Nodes Predicts Patients’ Survival

As the pattern of expression of breast cancer stem cell markers was discordant between primary breast carcinomas and metastatic lymph nodes, we set out to investigate whether the variation in P-cadherin, CD44, and CD49f expression allowed the stratification of patients regarding clinical behavior. Remarkably, we found that breast cancer cases without P-cadherin expression in the primary tumors, but with de novo expression in the corresponding metastatic lymph nodes, were from patients exhibiting the worst overall survival and disease-free survival, in comparison with other combinatorial possibilities. Accordingly, cases scored as P-cadherin positive in the primary tumor, but for which the matched lymph node was classified as negative, showed a better prognosis (Figure 3a, P=0.017 for overall survival and P=0.002 for disease-free survival). Univariate Cox regression analysis showed that P-cadherin gain of expression in the lymph nodes was the most relevant factor in predicting poor patients’ prognosis, with a HR of 4.108 for disease-free survival, when compared with tumors remaining negative in the metastatic axillary site (P=0.001). The same significant result was observed for OS (HR=2.843, P=0.018) (Table 3). Importantly, multivariate analysis showed that this effect was independent of primary tumor size, histological grade and breast cancer molecular subtype (HR=6.711 for disease-free survival, P=0.001; HR=4.091 for overall survival, P=0.033, Table 3).

Figure 3
figure 3

Survival analysis of breast carcinoma patients according to the change of breast cancer stem cell markers expression between primary tumor and lymph nodes metastasis. Kaplan–Meier survival plots of overall survival and disease-free survival of breast cancer patients considering the alterations in the expression of P-cadherin (a), CD44 (b), and CD49f (c) in primary tumors and matched lymph node metastasis.

Table 3 COX regression survival analysis

Concerning CD44 and CD49f expression, the variation found between the primary tumor and the matched metastatic lymph nodes had no impact in patients’ clinical outcome (Figure 3b and Figure 3c).

Discussion

In breast cancer patients, the presence of cancer cells in lymph nodes represents a powerful prognostic factor and the number of metastasized lymph nodes has a profound impact in the recommended adjuvant treatment by clinicians.63 Notably, axillary positive carcinoma patients may present a heterogeneous clinical behavior.14, 15, 16 Still, the biology of metastasized cells is not contemplated in clinical decisions, mainly owing to the assumption that their phenotype is the same as in the primary tumor.

In this study, we focused on studying the expression of breast cancer stem cell markers in primary tumors and matched lymph node metastases, as cancer cells with stem cell properties are considered crucial mediators of tumor heterogeneity, disease aggressiveness and potential seeds of metastasis.29, 31 Thus, we demonstrated that the sole analysis of the expression of the breast cancer stem cell marker P-cadherin exclusively in lymph nodes metastases is enough to predict poor clinical outcome of breast cancer patients (disease-free survival and overall survival). Although CD44 and CD49f were significantly associated with P-cadherin expression in lymph node metastasis, the expression of these markers did not impact patient survival.

As previously demonstrated by other groups, concerning the expression of hormone receptors and biomarkers involved in cell proliferation, differentiation, and apoptosis, we also found a high degree of concordance between breast cancer stem cell markers expression in primary carcinomas and matched node metastasis (80.8% for P-cadherin, 88.5% for CD44 and 73.8% for CD49f).14, 24, 25, 26, 64 However, we highlight that a small, but relevant, difference was found comparing primary breast carcinomas and matched loco-regional disseminated tissue regarding their expression. Actually, we observed that this difference in the expression of P-cadherin showed a statistically significant impact in patient disease-free survival and overall survival (univariate and multivariate survival analysis). Specifically, we demonstrated that negative cases for P-cadherin in primary carcinomas, whereas being positive in lymph node metastases, were significantly associated with the highest probability of patient relapse and death, possibly representing a high-risk group. In contrast, carcinomas that remain negative, or carcinomas that loose the expression of P-cadherin in metastatic lymph nodes showed a better patient prognosis. Thus, the evaluation of P-cadherin in the lymph nodes may explain why some breast carcinoma patients develop rapidly aggressive disease, whereas others never develop distant metastasis.14, 15, 16 This specific result was striking when compared with the classical stem cell markers CD44 or CD49f, for which a comparable variation in expression did not impact the survival of patients. Considering that lymph nodes assessment of stem cell markers is not routinely performed in pathology laboratories, it would be interesting to further validate these results in a larger and independent series. This is particularly relevant, considering that the few existing immunohistochemical studies that report patient outcome comparing primary carcinomas and lymph node metastatic disease are limited to the classical prognostic markers, namely HER2,24 and ER.27, 65

Based on the results obtained, we thought that it would be interesting to cross this data with the information concerning distant dissemination sites for each one of these cases. Unfortunately, we were not able to obtain these records; however, it is reasonable to postulate that the tumors that correlated with earlier relapse and death were the ones that led to distant metastatic dissemination.

Notably, Adamczyk et al.66, 67 reported a dissimilar P-cadherin expression between primary tumor and synchronous lymph node metastases in up to 33.8% of invasive ductal breast carcinomas; however, the impact of this marker in patients’ survival was not clarified in these studies. In that sense, our study brings an important novelty, including a 10-year follow-up of invasive breast carcinoma patients. Although additional studies will be needed to explain the poor prognostic phenotype conferred by the gain of P-cadherin expression within metastatic lymph nodes, these observations correlate with our previous results, which demonstrated that P-cadherin potentiates the invasive and colonizing potential of cancer cells and possibly contributes to the metastatic capacity of breast cancer cells.46, 47, 49 Two independent studies have also analyzed P-cadherin as a putative predictive biomarker of distant metastatic risk in distinct human patient data sets, where its expression in the primary tumor was significantly associated with earlier occurrence of distant metastasis.68, 69 In our study, P-cadherin expression in the primary tumor was unable to stratify patients regarding disease-free survival or overall survival, in contrast to previously published studies.45, 49, 52 However, it is worth mentioning that our series included only patients with lymph node positive disease at diagnosis, suggesting that the assessment of biomarkers solely in the primary lesions is not useful in stratifying these already poor prognosis patients.

In clinical terms, the evaluation of molecular status in residual tumors or in metastatic disease can potentially define better treatment options and estimate the actual patient prognosis. As CD44 or CD49f expression alone was not associated with patient survival in metastatic lymph nodes or primary tumors, our study points that P-cadherin is a stronger candidate biomarker for axillary-based breast cancer decisions in the clinical practice than the classical cancer stem cell markers. Future mechanistic studies should clarify the distinct P-cadherin expression profile found between primary carcinomas and matched lymph node metastasis. This may be due to the selection and expansion of a cell clone with specific advantageous features that originated in the primary tumor or the result of genetic and/or epigenetic changes promoted by the node microenvironment itself.

Finally, it is worth mentioning that patients included this study were not treated with neo-adjuvant therapy and were considered treatment naïve regarding lymph node biomarker assessment. Furthermore, these patients were diagnosed from 1978 up to 1992, being subjected to the same standard adjuvant treatment following the protocols at that time. Patients with HER2+ tumors were not treated with specific targeted therapy (trastuzumab was not used in the clinic practice then) and patients with ER+ tumors were exclusively treated with tamoxifen. Therefore, although the adjuvant treatment could have influenced the results of the study, it is important to mention that an identical therapy protocol was administered to all patients, which means that under these conditions P-cadherin evaluation in lymph nodes metastases after surgery is an important predictor of disease progression.

Conclusion

The immunohistochemical analysis of biomarkers in lymph node malignancy is not routinely performed to ascertain disease outcome. In this report, we show that the breast cancer stem cell markers CD44, CD49f, and P-cadherin show a modest, but meaningful change in expression between primary lesions and paired regional lymph node metastatic disease. P-cadherin emerged as the only independent prognostic indicator. Despite only 8.2% of breast cancer patients gained P-cadherin expression, these cases presented the worst survival of the whole series, denoting a high-risk group of patients. This phenotype was mainly found in the triple-negative breast cancer subtype. Altogether, our study highlights that, although a more expansive study and independent validation will be required, the immunohistochemical detection of P-cadherin in primary carcinomas and matched metastatic nodes, can potentially help clinicians to better stratify patients according to prognosis and to improve the treatment options for breast cancer patients.