We retrospectively reviewed the medical records of patients with breast cancer treated at PUMCH (an academic teaching hospital with 2000 beds and over 3 million outpatients each year) from January 2005 to August 2018. The inclusion criteria were as follows: 1) age ≥18 years, 2) pathologically diagnosed with stage I or II breast cancer, and 3) presented with recurrent or metastatic breast cancer during the follow-up period after systemic therapy. The exclusion criteria were as follows: 1) male gender, 2) bilateral breast cancer, and 3) no record of the pathological diagnosis for either the primary or recurrent/metastatic lesions.

A total of 75 patients were included in the study. Forty patients (53.4%) received mastectomy + axillary lymph node dissection and eight patients (10.6%) received mastectomy + sentinel lymph node biopsy. Thirteen patients (17.3%) received lumpectomy + axillary lymph node dissection and 14 patients (18.6%) received lumpectomy + sentinel lymph node biopsy. Three patients received reoperations for the primary cancer due to positive margins. All patients were operable at diagnosis, and only one patient received neoadjuvant chemotherapy. The puncture result for this patient was consistent with the surgical results but a pathologic complete response (pCR) was not achieved. Most patients received adjuvant chemotherapy. The interval between surgery and systemic therapy was 2–4 weeks for most patients.

In the present study, 56 patients (56/75, 74.7%) opted for chemotherapy. The regimens consisted of doxorubicin/epirubicin + cyclophosphamide, docetaxel + cyclophosphamide, paclitaxel/docetaxel + doxorubicin, and docetaxel + doxorubicin/epirubicin + cyclophosphamide for HER2-negative patients. Docetaxel + cyclophosphamide + trastuzumab and doxorubicin + cyclophosphamide followed by docetaxel + trastuzumab were used for HER2-positive patients. The drug doses were doxorubicin/epirubicin 50/75 mg/m², cyclophosphamide 500 mg/m2, paclitaxel/docetaxel 175/75 mg/m², and trastuzumab 8 mg/kg IV on day 1 followed by 6 mg/kg every 3 weeks.

Twenty-six patients (26/75, 34.7%) received radiotherapy. For postmastectomy therapy, patients were usually treated in the supine position with a customized immobilization device. The target volume included the chest wall and supraclavicular fossa, and the patients were treated with 45–50 Gy at 2 Gy/fx with 6 Mv-X rays from a linear accelerator. The planning target covered the entire mastectomy scar, flaps, surgical clips, and drain sites included in the treatment field, avoiding junctional overdose with the supraclavicular field. A 5 mm bolus was used for half the duration of radiotherapy. An electron boost could be used to bring the total scar dose to 60–66 Gy in high-risk patients. For breast conserving therapy, patients were usually treated in the supine position with a customized immobilization device. Whole-breast irradiation was delivered first with tangential opposed fields using 2-Gy daily fractions up to 50 Gy with 6 Mv-X rays from a linear accelerator. The planning target volume (PTV) covered the superior border at the manubriosternal joint and the inferior border at 1 cm below the inframammary line. The medial border was usually the midline of the sternum, and the lateral border was the midaxillary line, excluding the outermost 5 mm from the superficial skin surface. For regional nodal irradiation, the lower part of the ipsilateral axillary lymph node (LN) area was included in all cases. The upper part of the ipsilateral axillary LN area was also included when metastases to the LNs were present.

Fifty-one patients (51/75, 68.0%) with hormone receptor-positive breast cancer received endocrine therapy with tamoxifen, aromatase inhibitors, and goserelin for ovarian function suppression. The duration of adjuvant endocrine therapy planned for 34 patients was 5 years and that for seven patients with a high risk of recurrence was 5–7 years. The seven patients with a high risk of recurrence had recurrence/metastasis within 5 years after surgery, and their endocrine therapy was terminated afterward. These seven patients had two or three lymph nodes involved and the Ki-67 index was positive. Histopathologic parameters were available for two patients. The ER status for one patient changed from positive to negative, the PR status for two patients changed from positive to negative, the HER2 status for one patient changed from negative to positive, and the Ki-67 status changed from positive to negative. All primary tumors were Luminal B tumors, six were still Luminal B after recurrence/metastasis, and the subtype for one tumor was uncertain after recurrence/metastasis. Tamoxifen was administered at doses of 10 mg twice a day or 20 mg once a day. Aromatase inhibitors included letrozole, anastrozole, and exemestane. Letrozole was used at a dose of 2.5 mg once a day, anastrozole at 1 mg once a day, and exemestane at 25 mg once a day. Goserelin acetate sustained-release depot was injected at a dose of 3.6 mg once a month.

Histopathology was used as the gold standard for the diagnosis. Patients were classified as having anatomic stage I or II breast cancer in accordance with the 8^th American Joint Committee on Cancer (AJCC) cancer staging manual [16, 19]. Recurrent/metastatic lesions were defined as any local or regional recurrence and/or any distant metastases. All patients had histopathological records for both the primary and recurrent/metastatic lesions.

ER and PR staining were evaluated according to the American Society of Clinical Oncology/College of American Pathologists Guidelines [20]. Positive nuclear staining with any intensity observed in at least 1% of tumor cells was defined as ER/PR positivity. HER2 expression was scored as 0, 1+, 2+, and 3 + according to the American Society of Clinical Oncology/College of American Pathologists Guidelines [21]. HER2 status was considered positive if the immunohistochemical staining score was 3+ and negative if the score was 1 + or 0. If the immunohistochemical score was 2+, the HER2 gene amplification status was determined using chromogenic in-situ hybridization [22]. The Ki-67 index was quantified based on the spot with the highest intensity in the high-power field (400x) and 1,200–1,500 cells were counted. The Ki-67 antibody (Clone SP6; Thermo Scientific, Japan) was used to define the Ki-67 status. A positive status for Ki-67 was defined as more than 14% of cells with Ki-67 antigen-antibody reactions counted under high-power magnification (×400) [23, 24]. The discordance rate for each biomolecular parameter was calculated as the frequency of changes between the primary tumors and the recurrent/metastatic lesions. The tumor subtypes were classified as Luminal A (ER and PR positive, HER2 negative, “low” Ki-67, and a “low” recurrence risk based on multi-gene-expression assay results if available), Luminal B (“Luminal B-like (HER2 negative)”: ER positive, HER2 negative, and at least one of the following: “high” Ki-67, “negative or low” PR, or “high” recurrence risk based on multi-gene-expression assay if available. “Luminal B-like (HER2 positive)”: ER positive, HER2 over-expressed or amplified with any Ki-67, and any PR) [25], HER2+ (Hormone receptor-negative and HER2-positive), and triple negative (TN) (Negative ER, PR, and HER2) according to St. Gallen’s Guide in 2013 [25]. There were differences in the reagent use and standardization process in different periods. These differences were minimized by repeated reviews, re-staining, and interpretation of questionable pathological results based on the principle of being faithful to the original results at the time.

The survival outcome was progression-free survival (PFS), which was defined as the time from the start of treatment (neoadjuvant therapies or surgery) to tumor progression or death [26]. The clinical characteristics of the patients were collected. The details of systemic treatment, recurrent or metastatic disease, and discordance in each biomolecular parameter between primary and recurrent/metastatic lesions as well as PFS were recorded. The patients were examined every six months within three years after the operation and once a year regularly after three years. Imaging examinations included: breast and axillary, abdominal, lymph nodes, and uterine and appendages ultrasound, chest CT, and bone scintigraphy (bone scintigraphy once a year for the first five years). Once the patient exhibited any symptoms, imaging examination was performed as soon as possible if necessary.

The continuous variables were presented as median with range, and categorical variables were presented as count and proportions. Fisher's exact test (or Pearson’s chi-square test) and the t-test were used to compare patient and tumor characteristics. Pearson’s chi-square test or Fisher’s exact and regression analyses were used to determine the relationship between discordance and any categorical variables. Survival curves were plotted with the Kaplan–Meier method, and differences in the survival curves were analyzed with the log-rank test. We set p < 0.05 as the threshold for statistical significance. All statistical analyses were performed using the software package SSPS 20.0 (SPSS Inc., United States).

We reviewed the consecutive pathology findings for 8,691 women with breast cancers diagnosed between January 2005 and August 2018 at our hospital. A total of 133 patients were diagnosed with recurrent or metastatic tumors based on histopathology results. Of these, 75 patients had stage I or II disease, 31 had stage III or IV disease, and 27 patients could not be staged. For these 75 patients with stage I or II disease, the range of BMI was from 18.4 to 35.5. The number of patients who had a family history of breast cancer, ovarian cancer, pancreas cancer, and other cancers was 4, 1, 0, and 5, respectively. The most common comorbidity was hypertension. Other complications included hyperthyroidism, diabetes and coronary heart disease. Fifteen patients were postmenopausal when diagnosed with breast cancer, but none of them were treated with hormone replacement therapy.

The clinical characteristics of the included patients are shown in Table 1. The median age at diagnosis for all 75 patients was 44, ranging from 23 to 88 years old. In addition to ductal carcinoma and lobular carcinoma, there were 12 patients with special types of carcinoma, including mucinous carcinoma, micropapillary carcinoma, cribriform carcinoma, and two types of mixed carcinoma. The median pathological size for the maximum diameter of the included primary lesions was 2.0 cm, ranging from 0.1 to 4.5 cm. The median follow-up was 57.2 months (range 7.2 to 163.0 months). The median time after surgery before recurrence/metastasis was 37.0 months (range 3.1–163.5 months). Based on the immunohistochemical results for recurrent or metastatic tumors, patients were treated accordingly (see details in Table 1).

The positive expression results for ER, PR, HER2, and Ki-67 expression in the primary and recurrent or metastatic tumors can be found in Table 1. There was no significant statistical difference in these parameters between the primary tumors and the recurrent or metastatic tumors (p > 0.05).

The shortest interval between surgery and recurrence was 3.1 months, during which the patient underwent modified radical mastectomy after neoadjuvant chemotherapy and developed chest wall recurrence. Two patients died during the follow-up period; one died of brain metastasis and the other died of multiple-organ metastasis.

Changes in the ER status were observed in seven patients (9.3%), with decreased expression (from a positive to negative status) in four patients and increased expression (from a negative to positive status) in three patients (Table 2). Differences in the PR status were found in 11 patients (14.7%), with decreased PR expression being the most common change (9/11). HER2 alterations were found in 10 patients (10/68, 14.7%), with decreased expression in three patients and increased expression in seven patients. Changes in Ki-67 were observed in 14 patients (14/65, 21.5%), with decreased expression in five patients and increased expression in nine patients. The status of ER, PR, HER2, and Ki-67 expression in the primary tumors and the recurrent or metastatic tumors was identical in only 33 patients.

If 10% was considered the threshold for ER and PR positivity, six results would change from the original positive determination to negative. The discordance rates of ER and PR expression were 10.6% and 18.6%, respectively.

Overall, the subtype of the recurrent/metastatic lesions was different in 11 patients (11/64, 17.2%) (Table 3). The highest discordance rate (66.7%) was observed in patients with Luminal A primary tumors. The recurrent/metastatic lesions changed to the Luminal B subtype in three of these patients and to the TN subtype in one patient. On the contrary, TN primary tumors were the most stable, as no change from the TN subtype to other subtypes was observed. Patients with Luminal B and HER2 primary tumors had discordance rates of 11.8% and 14.3%, respectively.

The relationships between the differences in each biomolecular parameter and the potential predictive factors are shown in Table 4. There were statistical differences between the Luminal A subgroups in primary tumors and discordance (p < 0.05). No significant differences (p > 0.05) in any other biomarkers or tumor subtypes were observed for the potential predictors evaluated (e.g., primary tumor characteristics and therapies).

At the primary diagnosis, three patients in the HER2-positive group did not receive anti-HER2 therapy because the tumor was less than 1 cm with negative lymph nodes and four patients did not receive the therapy because of economic or personal reasons. After recurrence, eight patients did not receive anti-HER2 therapy because of chest wall recurrence and four did not receive the therapy because of economic or personal reasons.

Among 18 patients (24.0%, 18/75) with ER/PR expression changes and 10 patients (13.3%, 10/75) with HER2 expression changes, treatments based on the recurrent receptor status were applied for 25 patients (89.3%, 25/28). The treatment for three out of five patients with ER or PR changes from negative to positive was adjusted to endocrine therapy; three received additional chemotherapy and one received additional radiotherapy. Four out of 13 patients with ER or PR changes from positive to negative received chemotherapy after biopsy, eight discontinued endocrine therapy, and one received additional radiotherapy. One patient with HER2 change from negative to positive received anti-HER2 treatment, three received additional endocrine treatment, and two received additional radiotherapy. The treatment for 69 patients was altered after recurrence or metastasis and 66 of these patients remained recurrence-free after resection of the recurrent lesions.

The Kaplan–Meier survival curves for PFS in the patients can be found in Figures 1–4. The PFS was compared according to the discordance vs. accordance of the receptors, negative to positive (gain) vs. an unchanged positive result for the receptor, and positive to negative (loss) vs. an unchanged negative result for the receptor. A significant difference was found between patients with PR loss and those with an unchanged negative PR result (p = 0.020). There was no statistically significant difference in any of the other parameters evaluated (p > 0.05).

There was a statistical difference in the PFS of patients according to the subtype of the primary breast cancer (p = 0.0001). However, no statistical difference in PFS according to the subtype of the recurrent or metastatic breast cancer was found (p > 0.05) (Figure 5).