Clinical verification of the relationship between smoking and the immune microenvironment of breast cancer

This retrospective study evaluated data from 149 patients who underwent preoperative chemotherapy (POC) for resectable TNBC or HER2BC between February 2007 and December 2017 at the Osaka City University Hospital. All patients were questioned regarding their smoking history at the initial visit (cigarettes smoked per day and years of smoking), and the data were used to calculate each patient’s smoking amount based on pack-years (Table 1). The breast cancers were diagnosed pathologically and classified according to subtype based on the immunohistochemical expression of estrogen receptor (ER), progesterone receptor (PgR), human epidermal growth factor receptor 2 (HER2), and K-i67. Cases were defined as either HER2BC (ER−, PgR−, and HER2+) or TNBC (ER−, PgR−, and HER2−).

All patients received a standardized outpatient POC regimen that consisted of four courses of FEC100 (fluorouracil: 500 mg/m², epirubicin: 100 mg/m², and cyclophosphamide: 500 mg/m²) every 3 weeks, which was followed by 12 courses of weekly paclitaxel (80 mg/m²). The patients with HER2BC also received trastuzumab during the paclitaxel treatment as a weekly dose (2 mg/kg) or tri-weekly dose (6 mg/kg) [15‐17]. Staging and therapeutic effect were evaluated using ultrasonography, computed tomography, and bone scintigraphy based on the Response Evaluation Criteria in Solid Tumors [18]. Patients who achieved clinically partial or complete response were categorized as “responders” in the objective response rate (ORR), while patients with clinically stable or progressive disease were defined as “non-responders”. The patients subsequently underwent mastectomy or breast-conserving surgery [19], and the pathological therapeutic effect of the POC was evaluated using the resected specimens. Pathological complete response (pCR) was defined as complete disappearance of the lesion’s invasive components, including the lymph nodes, with or without intraductal components, according to the National Surgical Adjuvant Breast and Bowel Project B-18 protocol [20]. All patients received postoperative radiotherapy delivered to the remnant breast, and the standard postoperative adjuvant therapy was selected based on the cancer subtype. Patients were followed-up after surgery to detect recurrence using physical examinations every 3 months, ultrasonography every 6 months, and computed tomography and bone scintigraphy annually. The DFS interval was calculated from the day of surgery to the first instance of recurrence or death, while OS was calculated from the day of surgery to death.

Specimens that were used to pathologically diagnose breast cancer (obtained via core needle biopsy or vacuum-assisted biopsy) were used to determine the TILs density. In the present study, TILs were defined as lymphocytes infiltrating within the tumor stroma [21]. The TILs density was calculated as the average from five randomly selected fields, and the results were classified as a score of 3 (> 50%), a score of 2 (11–50%), a score of 1 (≤ 10%), or a score of 0 (absent) (Additional file 1: Fig. S1). Based on previous reports [22, 23], we defined a high TILs density as scores of 2–3 (i.e., > 10%) and a low TILs density as scores of 0–1 (≤ 10%).

All analyses were performed using JMP software (version 11; SAS Institute, Cary, NC). Differences in the study variables were evaluated using the Chi square test or Fisher’s exact test, as appropriate. The Kaplan–Meier method was used to estimate the DFS and OS outcomes, which were compared using the log-rank test. A Cox proportional hazards model was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs), and multivariable analysis was performed using a Cox regression model and the backward stepwise selection method. Differences were considered statistically significant at p-values of < 0.05.

This study was conducted at the Osaka City University Graduate School of Medicine (Osaka, Japan) according to the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) guidelines. The study protocol involved a retrospectively written research, pathological evaluation, and statistical analysis plan [24]. 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 study’s retrospective protocol was approved by the ethics committee of Osaka City University (#926).

The clinicopathological features of the 149 women are listed in Table 2. The median age at surgery was 56 years (range 24–75 years old). The median follow-up duration was 1288 days after surgery (range 13–3615 days). The median tumor diameter was 27.6 mm (range 10.2–98.0 mm) and 98 patients (65.8%) were diagnosed with N1–3 lymph node metastasis based on their imaging results. Sixty-two patients had HER2BC (41.6%) and 87 patients had TNBC (58.4%). Ninety-one patients (61.1%) had a high TILs density and 58 patients (38.9%) had a low TILs density at their diagnosis. One hundred and five patients (70.5%) reported never smoking, and 44 patients (29.5%) reported a median smoking amount of 20 pack-years (range 2.5–135 pack-years). Based on the receiver operating characteristic curve analysis, the optimal smoking cut-off value for predicting DFS was defined as 2.5 pack-years, which yielded a distribution of 43 patients (28.9%) in the high smoking group and 106 patients (71.1%) in the low-smoking group (area under the curve: 0.588, sensitivity: 0.325, specificity: 0.846) (Additional file 2: Fig. S2). The ORR was 82.6% and 74 patients (49.7%) achieved a pCR. The therapeutic response was significantly higher among patients with HER2BC than among patients with TNBC (p = 0.023) (Table 3). However, there were no significant differences in the two groups’ pCR rates (p = 0.085), TILs density (p = 0.206), or smoking amount (p = 0.134).

Table 4 shows the results of the associations between smoking and the patients’ clinicopathological features. No significant correlation was found between comparing smokers and never smokers. However, when divided into two groups according to smoking amount, correlation with clinicopathological features was recognized. Relative to the low smoking group, the high smoking group had a significant greater TILs density (p = 0.043) and a significantly better pCR rate (p = 0.042). In the univariate analysis, prolonged DFS was significantly associated with pCR (p < 0.001, HR 0.203, 95% CI 0.068–0.499) and a high TILs density (p = 0.001, HR 0.252, 95% CI 0.107–0.553) (Table 5). In addition, prolonged OS was significantly associated with pCR (p = 0.002, HR 0.183, 95% CI 0.042–0.561) and a high TILs density (p = 0.035, HR 0.357, 95% CI 00.129–0.929) (Table 5). However, there was no significant difference according to smoking amount in DFS (p = 0.114) or OS (p = 0.347) (Fig. 1).