Abstract
Background/Aim: Neoadjuvant chemotherapy (NAC) can affect tumors and the tumor microenvironment. As changes in tumor-infiltrating lymphocytes (TILs) after NAC and the resulting clinical significance have not been clearly defined, we evaluated both in triple negative breast cancer (TNBC). Materials and Methods: TIL level was histologically analyzed in pre-NAC biopsy and post-NAC operation specimens from 104 TNBC cases with residual invasive carcinoma after NAC. Results: Forty-three cases (41.3%) showed decreases in TIL level, whereas 29 cases (27.9%) showed increases and no significant changes were found in 32 cases (30.8%). A decrease or increase in TIL levels corresponded to a better disease-free survival (DFS) as compared to unchanged levels. In multivariate analysis, a change in TIL level was an independent prognostic factor for DFS. Conclusion: We identified a prognostic significance of TIL changes after NAC. Assessment of TILs before and after NAC can provide valuable prognostic information for TNBC patients.
Neoadjuvant chemotherapy (NAC) is not only the first line therapy for locally advanced breast cancer, but also the current trend for treating early stage breast cancer, especially triple-negative breast cancer (TNBC). NAC leads to an increased probability for a breast conserving operation (1). Differences pre- and post-NAC in breast cancer have been actively studied for a variety of systems in the last decade. NAC can change the expression levels of biomarkers such as estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) in breast cancer cells, all of which are important for designating breast cancer subtypes with different treatment strategies. Changes in ER, PR, and HER2 after NAC have been reported in a high proportion of cases as follows: ER 12~51%; PR 6~52%; and HER2 5~43% (2). Ki67 labeling can also be altered, and the prognostic significance of pre- or post-NAC alteration in the proliferation index has been studied in various cohorts of breast cancer (3). After NAC, tumors also show discordant mutational profiles and significant gene expression changes (4-7). These changes in cancer cells could be used to develop targeted therapies that can be used downstream of standardized NAC.
NAC can also affect other cell types in the tumor microenvironment, such as tumor-infiltrating lymphocytes (TILs). TILs are lymphocytes present in the stroma of a tumor or inside tumor cell nests. TILs are assessed through an eyeball estimation by pathologists or by quantification using digital image analysis in tumor tissue sections with hematoxylin and eosin (H&E) or immunohistochemical staining (8, 9). As adoptive TIL therapy and other immunotherapies dealing with immune cells and related molecules are actively being explored as innovative treatment modalities for various tumor types, a better understanding of the immune tumor microenvironment could enhance the success rate of treatments for breast cancer. However, changes in TILs after NAC and the resulting clinical significance have not been clearly defined. We evaluated these changes and assessed the prognostic value of TILs after NAC in TNBC.
Materials and Methods
Patients. TIL level was retrospectively analyzed in a total of 157 TNBC cases using pre-NAC biopsy and post-NAC operation specimens from the Asan Medical Center between 2010 and 2012. NAC regimens were anthracycline/taxane-based. TNBC was defined as ER, PR, and HER-2 negative as confirmed by immunohistochemical staining. In cases of equivocal HER-2 immunohistochemical staining, silver in situ hybridization was performed according to guidelines from the American Society of Clinical Oncology/College of American Pathologists (10). Pathologic complete response (pCR) is defined as no residual invasive carcinoma in the breast and lymph node after NAC.
Clinicopathologic information and survival data were obtained from medical records and surgical pathology reports. Exemption from informed consent after de-identification of information was approved by the Institutional Review Board of Asan Medical Center.
Histologic assessment of TIL. Histological TIL level was estimated for entire cases according to previously published methods from the International Immuno-Oncology Biomarker Working Group on Breast Cancer (11). In brief, TILs are determined by the percentage of the area occupied by mononuclear inflammatory cells, including plasma cells, over the total stromal area within the borders of the invasive carcinoma. As the distribution of TILs is usually not even throughout the tumor tissue, it is recommended that average TILs should be assessed, and not to focus on hot spots. Residual cancer burden (RCB) and Miller-Payne grade were also assessed (12, 13).
Statistical analysis. Wilcoxon signed rank analysis, Spearman's correlation, Jonckheere-Terpstra trend test, and survival analysis were used for analyzing and comparing variables using SPSS 20.0.0.
Results
Clinicopathologic characteristics. Median age at diagnosis was 41 years. Clinical stage distributions were 7.6% (12 cases), 63.7% (100 cases), 20.4% (32 cases), and 8.3% (13 cases), for cT1, cT2, cT3, and cT4, respectively (Table I). Lymph node metastasis was confirmed in 77.1% (121 cases) by biopsy. Post-NAC RCB classes were 29.3% (46 cases), 4.5% (7 cases), 48.4% (76 cases), and 17.8% (28 cases) for stages 0 (pCR), I, II, and III, respectively (Table II). After NAC, three cases shifted in molecular type. Two cases showed conversion of TNBC to HER-2 positive (ER and PR negative), and one case to ER and PR positive (HER-2 negative). Seven cases showed residual lymph node metastasis without residual invasive carcinoma in the breast. The median follow-up period was 72.3 months.
TILs in pre-NAC. The TIL level in pre-NAC biopsy specimens was 23.3±20.8%. TILs in pre-NAC biopsy specimens were negatively correlated with clinical T stage (Jonckheere-Terpstra trend test, p=0.003; Figure 1A). The TIL level of the pre-NAC biopsy specimens with a high histologic grade was significantly higher than that in those with a low histologic grade (grade 3 group, 29.1±20.8%; grade 1 or 2 group 16.1±18.4%; p<0.001, Figure 1B). TIL numbers from pre-NAC pCR cases were significantly higher than those in the non-pathologic complete response (non-pCR) group (pCR group, 30.8±24.9%; non-pCR group, 20.2±18.0%, p=0.017; Figure 1C). Pre-NAC TILs in groups subdivided by RCB class were 30.8±24.9%, 30.7±17.9%, 21.3±18.6%, and 14.7±15.2%, for RCB class 0 (pCR), I, II, and III, respectively. Thus, pre-NAC TILs tended to decrease with increasing RCB class (Jonckheere-Terpstra trend test, p<0.001; Figure 1D).
Clinicopathologic characteristics of pre-neoadjuvant chemotherapy (NAC) status.
Clinicopathologic characteristics of post-neoadjuvant chemotherapy (NAC) status.
TILs in post-NAC. TILs in the 104 non-pCR cases averaged 17.1±24.2%. TIL level in post-NAC operation specimens was negatively correlated with post-NAC pathologic T stage (Jonckheere-Terpstra trend test, p=0.002; Figure 2A), but was not significantly correlated with post-NAC pathologic N stage (Spearman's rho=−0.102, p=0.304), RCB class (rho=−0.148, p=0.135), or Miller-Payne grade (rho=0.172, p=0.08). Post-NAC TILs in groups subdivided by RCB class were 21.3±16.4%, 19.7±3.1%, and 10.0±3.1%, for RCB class I, II, and III respectively. Post-NAC TILs exhibited no correlation with RCB class (Jonckheere-Terpstra trend test, p=0.136; Figure 2B).
Correlation between the level of tumor-infiltrating lymphocytes (TILs) in pre-neoadjuvant chemotherapy (NAC) specimens and other clinicopathologic factors. (A) clinical T stage, (B) histologic grade, (C) pathologic complete response (pCR), (D) residual cancer burden (RCB) class.
Changes in TILs. In a pair-wise analysis of TILs from 104 non-pCR cases, levels were lower in post-NAC operation specimens than in pre-NAC biopsies (biopsies, 19.8±18.0%; operations, 17.1±24.2%), but the difference was not statistically significant (p=0.074, Figure 3).
Changes in TILs were calculated by subtracting the TIL level of pre-NAC biopsy specimens from the TIL level of post-NAC operation specimens in the non-pCR group. The mean value for change in TILs was −2.69, with a 95% confidence interval (CI) of −7.65 and 2.26. Changes in TILs showed a significantly positive correlation with Miller-Payne grade (Jonckheere-Terpstra trend test, p=0.035). However, changes in TILs from post-NAC operation specimens were not significantly correlated with post-NAC pathologic T stage (rho=−0.175, p=0.076), pathologic N stage (rho=0.079, p=0.424), or RCB class (rho=−0.007, p=0.941).
Correlation between the level of tumor-infiltrating lymphocytes (TILs) in post-neoadjuvant chemotherapy (NAC) specimens and pathologic T stage (A) or residual cancer burden (RCB) class (B).
We classified tumors into three groups using the 95%CI for mean value of change in TILs [decreased (below the 95%CI=−7.65), not significantly changed (between 95%CI), and increased (over the 95%CI=2.26)]. Forty-three cases (41.3%) displayed a decrease in TIL level, whereas 29 cases (27.9%) showed an increase and no significant changes were found in 32 cases (30.8%).
Survival analysis. In a univariate analysis of all patients, RCB class [hazard ratio (HR)=2.255, 95%CI=1.718-2.959, p<0.001], TIL level pre-NAC (HR=0.977, 95%CI=0.959-0.995, p=0.013), TIL level post-NAC (HR=0.970, 95%CI=0.960-0.990, p=0.004) were associated with DFS, but age (HR=0.977, 95%CI=0.949-1.006, p=0.117) was not. As the TIL levels pre- and post-NAC were not independent factors, two multivariate analyses were performed. In multivariate analysis including the level of TILs pre-NAC, both RCB class (HR=2.640, 95%CI=1.799-3.874, p<0.001) and level of TILs pre-NAC (HR=0.982, 95%CI=0.964-0.999, p=0.043) were independent prognostic factors for DFS. In multivariate analysis including the level of TILs post-NAC, RCB class (HR=2.278, 95%CI=1.552-4.793, p<0.001) and the level of TILs post-NAC (HR=0.977, 95%CI=0.960-0.995, p=0.013) were also independent prognostic factors for DFS.
As changes in TILs can be evaluated only in non-pCR cases, we analyzed the prognostic significance of post-NAC TILs and changes in TILs in a subgroup of patients with non-pCR. In univariate analysis of non-pCR cases, RCB class (HR=3.202, 95%CI=1.823-5.623, p<0.001), pre-NAC TIL level(HR=0.976, 95%CI=0.958-0.995, p=0.015), and post-NAC TIL level (HR=0.973, 95%CI=0.956-0.992, p=0.004) were significant prognostic factors for DFS, but changes in TILs (HR=0.993, 95%CI=0.983-1.04, p=0.221) were not associated with DFS. Multivariate analysis including the level of pre-NAC TILs revealed that the level of pre-NAC TILs (HR=0.977, 95%CI=0.957-0.996, p=0.019), age (HR=0.949, 95%CI=0.918-0.981, p=0.002), and RCB class (HR=3.690, 95%CI=2.020-6.738, p<0.001) were significant prognostic factors for DFS. In multivariate analysis including the level of TILs post-NAC, the level of post-NAC TILs (HR=0.979, 95%CI=0.961-0.997, p=0.023), age (HR=0.964, 95%CI=0.934-0.994, p=0.019), and RCB class (HR=3.316, 95%CI=1.834-5.998, p<0.001) were also significant independent prognostic factors for DFS.
Multivariate analysis for disease-free survival in non-pathologic complete response cases with change in tumor-infiltrating lymphocytes (TILs) level subgroups.
Kaplan–Meier curves demonstrated that DFS was stratified by RCB class in all cases (p<0.001; Figure 4A). In the non-pCR group, altered TIL level subgroups (decreased, not significantly changed, and increased) correlated with different DFS (p<0.001; Figure 4B). Groups with a change in TIL level (decrease or increase) exhibited better DFS than that of the unchanged group. Although the increased TIL group showed better survival outcomes than the decreased TIL group, the difference was not statistically significant (p=0.338).
Changes in tumor-infiltrating lymphocytes (TILs) after neoadjuvant chemotherapy (NAC) in non-pathologic complete response (pCR) cases.
In multivariate analysis of non-pCR cases, RCB class (p<0.001), age (p=0.008), and change in TIL level (p=0.012) were independent prognostic factors for DFS (Table III).
Discussion
We found a prognostic significance for altered levels of TILs in pre- and post-NAC specimens. The relationship between TILs and outcome in breast cancer was first reported in 1992 (14). Since then, the prognostic value of TILs in breast cancer has been extensively studied (15, 16), with large scale clinical trials for adjuvant chemotherapy revealing that increased levels of TILs (for each 10% increment of stromal TILs) in TNBC were associated with improved clinical outcome (HR≤0.87 for DFS, p≤0.032) (16). In TNBC patients without adjuvant systemic therapy, low levels of TILs (≤10%) were significantly correlated with distant metastasis (p=0.013) and inferior 10 year breast cancer specific survival rate (88.3% for pathologic N0 TNBC patients with low level of TILs vs. 100% for those with high level of TILs, p=0.036, n=64) (17).
TILs in pre-NAC pCR cases were significantly more numerous than those in the non-pCR group. High pre-NAC levels of TILs are a known predictive marker for NAC response in all breast cancer subtypes (16, 18, 19). In addition, pre-NAC TIL levels are reported to be a strong prognostic factor for overall survival (20) and distant recurrence (21). We also identified pre-NAC TILs as an independent prognostic factor for DFS in all patients, and a subgroup of patients with non-pCR.
Kaplan–Meier survival curves for disease-free survival (DFS). (A) Kaplan–Meier curves with residual cancer burden (RCB) class for all cases, (B) Kaplan–Meier curves with changes in TIL in non-pathologic complete response (pCR) cases.
Recently, the International Immuno-Oncology Biomarker Working Group published guidelines for assessing TILs in specimens from patients after NAC (22). The principle of assessing TILs is fundamentally similar for primary breast cancer without NAC. It is recommended that TILs be assessed in the stroma of the residual tumor bed of invasive carcinoma. Further, TILs in tumor necrosis and crush artefacts should be excluded, while TILs around tumor cells embedded within histiocytic aggregations should be assessed. In addition to pre-NAC TILs, post-NAC TIL levels in residual tissues are significantly associated with prognosis in TNBC (23-25). In this study, the level of post-NAC TILs was an independent prognostic factor for DFS. TILs post-NAC added significant prognostic value to a multivariate analysis including RCB class for DFS (p<0.001) in an analysis of 375 TNBC cases (23). Post-NAC CD4+ TILs were also associated with survival in RCB-II and RCB-III classes (p=0.05 and p=0.05, respectively)(26).
The prognostic significance of changes in TILs after NAC has largely not been studied. Liu et al. reported that increased TILs after NAC were associated with poor prognosis in 30 cases of breast cancer including all subtypes (27). On the other hand, Miyashita et al. reported that increased CD8+ TILs after NAC were associated with better prognosis than the decreased CD8+ TIL group in 78 cases of TNBC (28). In our cohort, patients with increased levels of TILs after NAC also exhibited improved survival outcomes compared to the survival outcomes in those with decreased levels of TILs, but the difference was not statistically significant. However, patients with significant changes in TILs after NAC, both an increase and decrease, showed better prognosis than did the unchanged group. To clarify the prognostic significance of changes in TILs after NAC, further studies with a larger sample size are necessary.
In pair-wise analysis, the levels of TILs were lower in post-NAC operation specimens than in pre-NAC biopsies, but not to a statistically significant degree in this study. Pelekanou et al. compared 43 pairs of pre- and post-NAC breast cancer tissues, and the median TIL counts were 5% (range=0-60%) for pre-NAC and 7.5% (range=0-40%) for post-NAC (29). They found that 5% was the median difference in TIL levels (range=0-50%) between pre- and post-NAC, indicating an overall increase in TILs in residual specimens (p=0.09). However, when the same author compared 59 pairs of pre- and post-NAC breast cancer tissues from the SWOG S0800 phase II NAC trial, a majority of tumors (78%) showed decreases in TIL levels after NAC, and the mean change in TILs was a 15% decrease (30). In our 517 non-pCR cases of all subtypes of breast cancer with NAC, the level of TILs decreased in 64.0% (n=331), was not changed in 18.4% (n=95), and increased in 17.6% (n=91) (unpublished data). Similarly, Hamy et al. reported that TIL levels decreased in 61.6% of tumors (n=441), did not change in 17.7% (n=127), and increased in 20.7% (n=148) after NAC in all subtypes of breast cancer.
Dieci et al. evaluated TILs on 19 TNBC biopsy tissues derived from patients with a high (>60%) post-NAC (anthracycline- or anthracycline-taxane-based regimen) TIL level (24). Fifteen out of 19 patients (79%) had a low (≤60%) pre-NAC TIL level. All but one case showed a higher level of TILs in post-NAC tissues compared to that in pre-NAC biopsies. Likewise, in our 104 non-pCR cases of TNBC with NAC, 12 cases showed high post-NAC TIL levels. Except for one, pre-NAC TIL levels were lower than those in post-NAC, in all cases. Therefore, despite the majority of cases showing decreased or no significant changes in TIL levels, several cases displayed increased TIL influx and a lymphocyte predominant phenotype after NAC in TNBC.
Reductions in immune cells might be a direct effect of chemotherapy. Increases in TILs can be explained by NAC inducing the release of damage-associated molecular patterns (DAMPs) through killing or otherwise damaging cancer cells. DAMPs can activate immune cells, alter the tumor microenvironment, generate increased tumor-specific neoantigens in cancer cells, and recruit immune cells. Immune monitoring of the tumor microenvironment could be an additional prognostic factor and might enhance the successful treatment of breast cancer with NAC in the era of innovative immunotherapies.
TILs can be used as anti-cancer immune cell therapeutics; we previously cultured breast cancer TILs and found that the amount of cultured TILs was less in breast cancers with NAC than in those without NAC (31). However, in TNBC, TILs showing in vitro anti-tumor reactivity toward autologous cancer cells were derived from tumors with NAC (unpublished data). Therefore, TILs after NAC preserve T cells with T cell receptors reactive to autologous tumor cells and have the potential to attack tumors. Immune monitoring of the tumor microenvironment could enhance the successful treatment of breast cancer with NAC. Further studies are required to characterize immune infiltration in residual disease after NAC to identify candidates who could benefit from adjuvant therapy including immune checkpoint inhibitors.
In conclusion, we found a prognostic significance for altered levels of TILs in pre- and post-NAC TNBC specimens. Assessment of TILs before and after NAC can provide valuable prognostic information for TNBC patients.
Acknowledgements
This study was supported by Basic Science Research Programs through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning, Republic of Korea (NRF-2018R1D1A1B07048831).
Footnotes
Authors' Contributions
Conceptualization: HJL; Data curation: ML; Formal analysis: HL, HJL; Investigation: HL, ML, GG, HJL; Methodology: HL, HJL; Project administration: HL; Writing—original draft: HL; Writing—review & editing: ML, JHS, GG, HJL
Conflicts of Interest
No potential conflict of interest relevant to this article was reported.
- Received February 10, 2020.
- Revision received February 19, 2020.
- Accepted February 20, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
