Background

Smoking is a risk factor for the development of breast cancer [1]. One proposed explanation for this link between cigarettes and breast cancer carcinogenesis suggests that a tobacco component is delivered directly to the mammary gland tissue via blood circulation, leading to DNA damage in the mammary gland cells [2, 3]. This potential mechanism is supported by tissue culture and animal experiments in which this damage causes changes in breast cells, such as an increased proliferative capacity and malignant transformation [4,5,6].

When determining the course of breast cancer treatment, it is important to evaluate the status of estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2 (HER2) expression. However, changes in the receptor expression patterns over the course of treatment can present clinical challenges. Specifically, these patterns often differ between primary and recurrent tumors, leading to a poor prognosis after recurrence [7, 8]. Therefore, it is necessary to re-evaluate the receptor expression status when a recurrent tumor arises. Given the potential effects of tobacco components on breast cancer cell traits, we hypothesized that smoking may contribute to these changes in receptor expression in recurrent disease. In this study, we aimed to analyze the relationships between changes in each receptor at recurrence, smoking and the subsequent prognosis.

Methods

Patient background

This retrospective study included 989 patients with resectable primary breast cancer who underwent curative resection as the first-line treatment between December 2007 and March 2018 at the Osaka City University Hospital (Osaka, Japan). Patients who received preoperative treatment and those with synchronous or metachronous bilateral breast cancer cases were excluded. At this institution, the patient’s smoking history (cigarettes smoked per day and years of smoking) is routinely recorded at the first visit, which yields the data necessary to calculate the pack-years as the number of cigarettes smoked per day divided by 20, then multiplied by the number of smoking years.

Each breast cancer received a definitive pathological diagnosis and was subjected to immunohistochemistry to determine the expression of ER, PgR, HER2, and Ki-67 (proliferation index). Based on the results, we classified the tumors in accordance with our previous work as hormone receptor-positive breast cancer (HRBC; ER- and/or PgR-positive), HER2-enriched breast cancer (HER2BC; ER−, PgR−, and HER2+), or triple-negative breast cancer (TNBC; ER−, PgR−, and HER2−) [9, 10]. We also applied a Ki-67 cutoff of 14% with reference to a previous report [11]. Tumor stage and resectability were evaluated using ultrasonography (US), computed tomography (CT), and bone scintigraphy.

Patients underwent primary tumor resection via mastectomy or breast-conserving surgery. Sentinel node biopsy or axillary dissection was performed in cases involving axillary nodal surgery; in the former cases, the detection of a sentinel node macrometastasis indicated the need for subsequent axillary dissection. After surgery, the patient was administered postoperative radiotherapy, delivered to the remnant breast, and standard postoperative adjuvant therapy according to the pathological diagnosis of the resected specimen. However, some patients did not receive postoperative treatment because of refusal or a poor general condition. All patients were followed-up via physical examinations, US, CT and bone scintigraphy according to the degree of recurrence risk.

Recurrence occurred in 77 of 989 patients who underwent curative resection. However, 19 of these patients did not undergo biopsy because the recurrent disease involved distant metastasis. This study also included some cases of distant metastasis wherein a biopsy was performed because it was difficult to differentiate the primary cancer of another organ from a distant metastasis of breast cancer. Of the remaining patients with recurrent disease, smoking history were not available for 3 patients. Therefore, we studied the remaining 50 cases (Fig. 1), all of whom underwent biopsy or resection immediately after relapse. No biopsies or resections were performed after the administration of antitumor drug treatment for recurrent disease.

Fig. 1
figure 1

Consort diagram. Recurrence occurred in 77 of 989 patients who underwent curative resection. However, nineteen of them did not undergo biopsy because of distant metastatic recurrence. In this study, such cases are also included. Of the remaining, 3 patients did not know the smoking history, so we studied in the remaining 50 cases

Full size image

Regarding survival outcomes, progression-free survival (PFS) was defined as the time interval from recurrence to deterioration by treatment started after recurrence or death. Post-recurrence survival (PRS) was defined as the time interval from recurrence to death. The 50 patients with recurrent disease were followed for a median of 2128 days (range, 416-–3789 days) postoperatively.

Statistical analysis

Comparisons between the two groups were performed using the Chi square test. The odds ratio (OR) and 95% confidence intervals (CI) were calculated by the logistic analysis. PFS and PRS were estimated using the Kaplan–Meier method and compared between groups using the log-rank test. The hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using the Cox proportional hazards model. A multivariable analysis was performed using the Cox regression model. All statistical analyses were performed using the JMP software package (SAS, Tokyo, Japan), and statistical significance was defined as a p value of < 0.05.

Ethics statement

This study was conducted at the Osaka City University Graduate School of Medicine, according to the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) guidelines. The study protocol involved a retrospectively written plan of research, pathological evaluation, and statistical analysis [12]. The study complied with the provisions of the Declaration of Helsinki, and all patients provided written informed consent for their treatment and data collection. The retrospective protocol was approved by the ethics committee of Osaka City University (approval number #926).

Results

Clinicopathological features

Fifty patients underwent radical surgery without preoperative treatment and a biopsy or resection of a recurrent tumor (Table 1). The median age of these patients was 60 (range, 37–79) years, and the median tumor size at the time of surgery was 21.8 (8.0–45.0). Fourteen patients (28.0%) had a history of smoking before surgery, with a median duration of 30 (1.4–150) pack-years. An evaluation of surgical specimens revealed that seven patients (14.0%) had lymph node metastases, as well as the following distribution of intrinsic subtypes: HRBC, 38 cases (76.0%); HER2BC, 2 (4.0%) cases; and TNBC, 10 (20.0%) cases. All patients with HRBC had a HER2-negative status. Thirteen patients (26.0%) with a pathological diagnosis that suggested a high risk of recurrence received postoperative adjuvant chemotherapy. Eleven patients (22.0%) received postoperative radiotherapy delivered to the remnant breast, and 1 patient (2.0%) received trastuzumab therapy. Moreover, 72% of all patients received endocrine therapy, and this high rate was attributed to the administration of this type of therapy to most patients with HRBC. In contrast, 6 patients (12.0%) did not receive any postoperative treatment.

Table 1 Clinicopathological features of 50 cases biopsied/resected to recurrence
Full size table

The median DFS duration was 792 (99–3300) days, and the median age at recurrence was 62 (range 41–86) years. Most biopsied recurrent tumors involved the local or regional lymph nodes, although biopsies were obtained from distant metastases in 5 cases (10.0%), including the lung in 3 cases (6.0%), brain in 1 case (2.0%), and liver in 1 case (2.0%). In 7 patients (14.0%), recurrences were observed in organs that were not biopsied simultaneously (Additional file 1: Table S1), including 3 patients who underwent biopsy of a local recurrence and also presented with lymph node metastasis, lung metastasis, or bone metastasis and 4 patients who underwent biopsy of a regional lymph node recurrence who presented with lung metastasis. No cases involved simultaneous recurrent lesions in 3 or more organs.

We further explored receptor expression in the recurrent tumors via histopathology. Regarding ER, negative conversion was recognized in 3 patients (6.0%) and positive conversion was recognized in 1 patient (2.0%). Regarding PgR, negative conversion was recognized in 15 patients (30.0%). Regarding HER2, positive conversion was recognized in 6 patients (12.0%). Consequently, 5 patients (10.0%) exhibited a change of intrinsic subtype upon recurrence.

Correlation between changes in receptor expression and clinical factors

The potential correlations of these receptor expression changes with clinical features were explored (Table 2). Patients who exhibited a negative PgR conversion were significantly more likely to have received postoperative endocrine therapy (p = 0.003). Patients who experienced a change in ER expression were less likely to have received postoperative radiation therapy, although this correlation was not significant (p = 0.052). No significant correlations were observed between a change in HER2 expression and any clinical factors. We then examined the potential correlations between smoking and clinical factors (Table 3). However, a significant correlation was only observed between a positive conversion of HER2 in recurrence and a history of smoking (p = 0.024). In addition, a univariate analysis with HER2 changes by pack-years showed that the odds ratio increased as pack-years increased (Table 4).

Table 2 Correlation between changes in receptor expression and clinical factors
Full size table
Table 3 Correlation between smoking and clinical factors
Full size table
Table 4 Univariate analysis with positive conversion of HER2 in recurrence for smoking
Full size table

Prognostic analysis based on a change in receptor expression and smoking history

The analysis revealed no significant difference in PFS between smokers and non-smokers (p = 0.102, log-rank; Fig. 2a). A univariate analysis identified significant correlations of chemotherapy after surgery and a change in intrinsic subtype in the recurrent tumor with a poor PFS (p = 0.015, HR = 3.734, 95% CI 1.316–10.115 and p = 0.039, HR = 3.889, 95% CI 1.083–11.236, respectively) (Table 5). However, no factors independently associated with PFS were identified in a multivariate analysis.

Fig. 2
figure 2

Regarding progression-free survival, there was no difference between smoker and non-smoker (p = 0.102, log-rank) (a). There was no difference in post-recurrence survival between smoker and non-smoker (p = 0.140, log-rank) (b)

Full size image
Table 5 Univariate and multivariate analysis with progression-free survival
Full size table

Similarly, no significant difference in PRS was identified between smokers and non-smokers (p = 0.140, log-rank; Fig. 2b). Although a univariate analysis identified a biopsied distant metastasis as associated significantly with a poor PRS (p = 0.041, HR = 8.527, 95% CI 1.114–52.010), no significant independent factors were identified in a multivariate analysis (Table 6). In summary, our results do not suggest an association between smoking and the prognosis after relapse.

Table 6 Univariate and multivariate analysis with post-recurrence survival
Full size table

Discussion

Decisions regarding the selection of breast cancer therapies require an accurate determination of the ER, PgR, and HER2 expression status of the tumor, which are usually determined via biopsy to achieve a definitive diagnosis. However, a biopsy specimen represents only part of lesion and often differs from the surgical specimen. Previous reports describe ER expression concordance rates between biopsy specimens and surgical specimens as high as 92–98% and similarly high PgR concordance rates of 85–97% [13, 14]. However, the reported HER2 concordance rates are slightly lower, at 80–90% [13, 15, 16]. Moreover, anticancer therapy affects the expression of these receptors. In a meta-analysis of patients who underwent neoadjuvant chemotherapy for breast cancer, ER and PgR discordance rates of 2.5–17% and 5.9–51.7%, respectively, were reported [17]. There are reports that it turns out to be often positive, while others report that it often turns negative. Regarding HER2, studies reported discordance rates between biopsy specimens and surgical specimens of 1.3–20% in patients who received neo-adjuvant chemotherapy (NAC) without trastuzumab and of 12–43% in whose who received NAC with trastuzumab. These data suggest that trastuzumab therapy induces a negative HER2 conversion. In our study, therefore, we targeted cases that were preoperative treatment-naïve to address the potential differences between biopsy and surgical specimens and changes due to NAC.

Some reports have described differences in the patterns of receptor expression between surgical specimens and recurrent tumor specimens [7, 8, 18,19,20]. A change in the ER status is observed in approximately 15% of cases, and the numbers of cases with increasing and decreasing expression are roughly equivalent. In contrast, a change in the PgR status is observed in approximately 25–40% of cases, and usually involves decreased expression. Changes in HER2 are observed in approximately 10% of cases, and more frequently tend to involve decreased expression. Consequently, some reports describe a change in breast cancer subtype to TNBC in recurrence, and these cases tend to have a worse prognosis than those with primary TNBC [7, 8]. In our study, we also compared the receptor expression patterns between surgical specimen and corresponding biopsies of recurrent tumors, which involved the local or regional lymph nodes in 90% of cases. The primary tumor type was HRBC in 76.0% of cases, and the frequencies of change in the ER, PgR, and HER2 statuses between the surgical and recurrent specimens were similar to those in previous reports.

In vitro experiments have demonstrated the ability of tobacco components to increase the proliferative capacity and induce malignant transformation in breast cancer cells [4,5,6], and various reports have described an association between ER expression and smoking in clinical practice [21,22,23,24,25,26]. However, few reports have explored the potential relationship between HER2 expression and smoking in breast cancer. Notably, we observed a significant correlation between smoking and a positive conversion of HER2 in our study. Although smoking is a known etiologic factor in lung cancer, an interesting potential correlation between HER2 mutation and lung cancer in never-smokers has attracted clinical attention [27, 28]. However, in vitro experiments have demonstrated the ability of tobacco components to induce HER2 [29] and amplify the expression of EGFR and HER3 [29, 30]. Crosstalk has been identified within the HER family, and potentially the amplification of another HER family member may enhance the expression of HER2 [31, 32]. In the future, it is necessary to examine the biological changes caused by tobacco components in breast cancer cells using immunohistochemical staining, genetic analysis, and protein quantification in vitro.

The choice of treatment after recurrence varied among the cases in our study, as some patients underwent excision of the recurrent lesions and others began anticancer therapy. Consequently, an evaluation of prognosis was challenging. However, we found that a negative hormone receptor conversion, positive HER2 conversion, and change of the intrinsic breast cancer subtype appears to reduce the DFS. However, smokers in our study appeared to have a better DFS and OS, possibly because the switch from ER+/HER2− to ER+/HER2+ breast cancer in most smokers enabled the administration of more effective drug treatment.

This study had a few limitations of note. Particularly, we only obtained data about the smoking history up to surgery, and the use of an interview to collect these data may have introduced bias. Although we agree that the postoperative smoking status is important, some reports suggest that the total smoking history is more important than the current smoking status with respect to carcinogenesis and recurrence [33, 34]. Moreover, we were not able to reach clear conclusions about receptor expression patterns on distant metastases, as most recurrences occurred in local or regional lymph nodes. However, the identification of a correlation between smoking and the positive conversion of HER2 at recurrence suggests that appropriate treatment may not have been administered to patients with distant metastases. We must therefore consider the possible link between smoking and HER2 amplification when evaluating cases in which a biopsy of a distant metastasis cannot be performed.

Conclusions

In conclusion, our results emphasize that biological changes during breast cancer recurrence should receive careful clinical consideration because of the potential effects on treatment after recurrence. However, smoking only appeared to have an effect on HER2 expression patterns after recurrence, but not on survival prognosis.