Association Between CD8+ T-Cell Infiltration and Breast Cancer Survival
Association Between CD8+ T-Cell Infiltration and Breast Cancer Survival
Details of the participating studies and patient characteristics are provided in supplementary Tables S1–S2 and Figure S3, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. In total, there were 12 439 patients of which 2674 (21%) died of breast cancer within 10 years of diagnosis. The median survival was 9.57 years (range 0.05–20.6 years). Supplementary Figure S4, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 illustrates the distribution of lymphocyte counts by study, tumour morphology and molecular subtype.
In ER-negative tumours, the presence of S- and iT-CD8+ lymphocytes was independently associated with a reduced relative risk of death from breast cancer (Table 1 and Figure 1). Supplementary Tables S5 and S6, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 detail univariate Cox regression analyses for all variables. Table 1 contains the multivariate Cox regression models (supplementary Tables S7–S9, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 detail multivariate models based on complete data and including hormone therapy as a covariate). The adjusted hazard ratio (HR) for iT-CD8+ tumours was 0.72 [95% confidence interval (CI) 0.62–0.84, P = 0.00003] and, for S-CD8+, the HR was 0.79 (95% CI 0.67–0.93, P = 0.004). For women with ER-negative breast cancer, absolute survival estimates (Kaplan–Meier survival function) for tumours positive for both iT-CD8+ and S-CD8+ lymphocytes compared with tumours negative for both were 77% (95% CI 74% to 79%) versus 66% (95% CI 62% to 70%) at 5 years and 71% (95% CI 68% to 74%) versus 58% (95% CI 54% to 63%) at 10 years (Figure 1). The presence of CD8+ lymphocytes was not associated with BCSS in ER-positive breast tumours (supplementary Table S5, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1). The presence of FOXP3+ lymphocytes was not associated with BCSS after adjustment for known prognostic factors (supplementary Tables S10 and S11, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1) irrespective of ER status. Unadjusted Cox regression analyses including an interaction term between cytotoxic and regulatory T-cell variables did not reveal a significant interaction between iT or S T-cell types irrespective of ER status (supplementary Table S12, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1).
(Enlarge Image)
Figure 1.
Kaplan–Meier survival plot of patient groups defined by the presence of intratumoral (iT)-CD8+ and stromal (S)-CD8+ cells in ER-negative disease. Unadjusted survival estimates at 5 and 10 years for double-positive and double-negative tumours are shown. Note: numbers at risk account for delayed entry of patients enrolled in the SEARCH study.
Significant heterogeneity of the prognostic effect of T cells was observed for different patient and tumour subgroups. Figure 2 shows HRs and 95% CIs from univariate Cox regression analyses of iT-CD8+ lymphocytes as separate forest plots for ER-positive and ER-negative disease; equivalent disease. Equivalent plots for S-CD8+, iT-FOXP3+ and S-FOXP3+ status are presented as supplementary Figures S5–S11, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. In particular, the prognostic effect of iT-CD8+ lymphocytes differed by HER2 status in ER-positive breast cancer (Pheterogeneity = 0.006). For ER-positive HER2-negative tumours, the HR was 1.16 (95% CI 1.02–1.32) and, for ER-positive HER2-positive tumours, the HR was 0.76 (95% CI 0.58–1.00). Following adjustment for histological grade, the HR associated with iT-CD8+ status in ER-positive HER2-negative tumours was 1.04 (95% CI 0.91–1.20). Figure 3 shows the absolute differences in survival of iT-CD8+ status within luminal and non-luminal breast tumours. Based on this finding, multivariate analysis of iT-CD8+ lymphocytes was conducted within the ER-positive HER2-positive subgroup as detailed in Table 2 and supplementary Table S13, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. The HR for iT-CD8+ lymphocytes was 0.73 (95% CI 0.56–0.96, P = 0.022) after adjustment for known prognostic factors.
(Enlarge Image)
Figure 2.
Subgroup analyses by patient and tumour characteristics of the prognostic effect of iT-CD8+ cells in ER-positive (left) and ER-negative (right) breast cancer. Definitions of molecular subtypes within ER-positive breast cancer (lunminal 1a, luminal 1b, luminal 2) and ER-negative breast cancer (HER2, CBP = core basal phenotype, 5NP = five-marker-negative phenotype) are defined in supplementary Table S4, available at Annals of Oncology online.
(Enlarge Image)
Figure 3.
Kaplan–Meier survival plots of patient groups defined by the presence of iT-CD8+ cells in luminal (left) and non-luminal (right) tumours. Note: numbers at risk account for delayed entry of patients enrolled in the SEARCH study. Definitions of molecular subtypes within ER-positive breast cancer (lunminal 1a, luminal 1b, luminal 2) and ER-negative breast cancer (HER2, CBP = core basal phenotype, 5NP = five-marker-negative phenotype) are defined in supplementary Table S4, available at Annals of Oncology online.
Supplementary Figures S12 and S13, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 depict the HRs and 95% CIs from Cox regression models adjusted for tumour size, positive lymph nodes and grade according to whether adjuvant chemotherapy was received for different T-cell types. There was no significant heterogeneity of the prognostic effect of T cells according to whether chemotherapy was administered. In order to account for differences in chemotherapeutic regimens between studies, subgroup analyses by study were conducted for iT-CD8+ status (supplementary Figure S14, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1); no significant difference in prognostic effect was observed by whether chemotherapy had been received within each study. supplementary Figures S15 and S16, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 are forest plots of the adjusted HRs and 95% CIs for the benefit of the addition of epirubicin to CMF in the NEAT trial in patient subgroups defined by T-cell status. Although there was no evidence of significant differential benefit of epirubicin in these T-cell-defined subgroups, there was a trend toward increased benefit in ER-positive patients with tumours devoid of S cytotoxic T cells (Pheterogeneity = 0.087) and, in ER-negative patients, with tumours positive for iT-CD8+ lymphocytes (Pheterogeneity = 0.12) using imputed data. Analysis restricted to cases with complete data only showed that the presence of iT-CD8+ cells was significantly associated with increased relative benefit from epirubicin (HR = 0.60, 95% CI 0.37–0.96) compared with cases negative for iT-CD8+ cells (HR = 1.47, 95% CI 0.72–3.02; Pheterogeneity = 0.039).
Further subgroup analyses were conducted to determine whether the prognostic effect of T cells varied according to tumour cell proliferation as suggested by a recent study and according to patient age as a result of the age-related decline of the immune system known as immunosenescence. The results of these analyses are depicted in supplementary Figure S17, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. No significant heterogeneity was observed by tumour proliferation status or age at diagnosis.
Results
Details of the participating studies and patient characteristics are provided in supplementary Tables S1–S2 and Figure S3, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. In total, there were 12 439 patients of which 2674 (21%) died of breast cancer within 10 years of diagnosis. The median survival was 9.57 years (range 0.05–20.6 years). Supplementary Figure S4, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 illustrates the distribution of lymphocyte counts by study, tumour morphology and molecular subtype.
Prognostic Value of T Cells
In ER-negative tumours, the presence of S- and iT-CD8+ lymphocytes was independently associated with a reduced relative risk of death from breast cancer (Table 1 and Figure 1). Supplementary Tables S5 and S6, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 detail univariate Cox regression analyses for all variables. Table 1 contains the multivariate Cox regression models (supplementary Tables S7–S9, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 detail multivariate models based on complete data and including hormone therapy as a covariate). The adjusted hazard ratio (HR) for iT-CD8+ tumours was 0.72 [95% confidence interval (CI) 0.62–0.84, P = 0.00003] and, for S-CD8+, the HR was 0.79 (95% CI 0.67–0.93, P = 0.004). For women with ER-negative breast cancer, absolute survival estimates (Kaplan–Meier survival function) for tumours positive for both iT-CD8+ and S-CD8+ lymphocytes compared with tumours negative for both were 77% (95% CI 74% to 79%) versus 66% (95% CI 62% to 70%) at 5 years and 71% (95% CI 68% to 74%) versus 58% (95% CI 54% to 63%) at 10 years (Figure 1). The presence of CD8+ lymphocytes was not associated with BCSS in ER-positive breast tumours (supplementary Table S5, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1). The presence of FOXP3+ lymphocytes was not associated with BCSS after adjustment for known prognostic factors (supplementary Tables S10 and S11, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1) irrespective of ER status. Unadjusted Cox regression analyses including an interaction term between cytotoxic and regulatory T-cell variables did not reveal a significant interaction between iT or S T-cell types irrespective of ER status (supplementary Table S12, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1).
(Enlarge Image)
Figure 1.
Kaplan–Meier survival plot of patient groups defined by the presence of intratumoral (iT)-CD8+ and stromal (S)-CD8+ cells in ER-negative disease. Unadjusted survival estimates at 5 and 10 years for double-positive and double-negative tumours are shown. Note: numbers at risk account for delayed entry of patients enrolled in the SEARCH study.
Subgroup Analysis and Chemotherapy
Significant heterogeneity of the prognostic effect of T cells was observed for different patient and tumour subgroups. Figure 2 shows HRs and 95% CIs from univariate Cox regression analyses of iT-CD8+ lymphocytes as separate forest plots for ER-positive and ER-negative disease; equivalent disease. Equivalent plots for S-CD8+, iT-FOXP3+ and S-FOXP3+ status are presented as supplementary Figures S5–S11, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. In particular, the prognostic effect of iT-CD8+ lymphocytes differed by HER2 status in ER-positive breast cancer (Pheterogeneity = 0.006). For ER-positive HER2-negative tumours, the HR was 1.16 (95% CI 1.02–1.32) and, for ER-positive HER2-positive tumours, the HR was 0.76 (95% CI 0.58–1.00). Following adjustment for histological grade, the HR associated with iT-CD8+ status in ER-positive HER2-negative tumours was 1.04 (95% CI 0.91–1.20). Figure 3 shows the absolute differences in survival of iT-CD8+ status within luminal and non-luminal breast tumours. Based on this finding, multivariate analysis of iT-CD8+ lymphocytes was conducted within the ER-positive HER2-positive subgroup as detailed in Table 2 and supplementary Table S13, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. The HR for iT-CD8+ lymphocytes was 0.73 (95% CI 0.56–0.96, P = 0.022) after adjustment for known prognostic factors.
(Enlarge Image)
Figure 2.
Subgroup analyses by patient and tumour characteristics of the prognostic effect of iT-CD8+ cells in ER-positive (left) and ER-negative (right) breast cancer. Definitions of molecular subtypes within ER-positive breast cancer (lunminal 1a, luminal 1b, luminal 2) and ER-negative breast cancer (HER2, CBP = core basal phenotype, 5NP = five-marker-negative phenotype) are defined in supplementary Table S4, available at Annals of Oncology online.
(Enlarge Image)
Figure 3.
Kaplan–Meier survival plots of patient groups defined by the presence of iT-CD8+ cells in luminal (left) and non-luminal (right) tumours. Note: numbers at risk account for delayed entry of patients enrolled in the SEARCH study. Definitions of molecular subtypes within ER-positive breast cancer (lunminal 1a, luminal 1b, luminal 2) and ER-negative breast cancer (HER2, CBP = core basal phenotype, 5NP = five-marker-negative phenotype) are defined in supplementary Table S4, available at Annals of Oncology online.
Supplementary Figures S12 and S13, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 depict the HRs and 95% CIs from Cox regression models adjusted for tumour size, positive lymph nodes and grade according to whether adjuvant chemotherapy was received for different T-cell types. There was no significant heterogeneity of the prognostic effect of T cells according to whether chemotherapy was administered. In order to account for differences in chemotherapeutic regimens between studies, subgroup analyses by study were conducted for iT-CD8+ status (supplementary Figure S14, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1); no significant difference in prognostic effect was observed by whether chemotherapy had been received within each study. supplementary Figures S15 and S16, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1 are forest plots of the adjusted HRs and 95% CIs for the benefit of the addition of epirubicin to CMF in the NEAT trial in patient subgroups defined by T-cell status. Although there was no evidence of significant differential benefit of epirubicin in these T-cell-defined subgroups, there was a trend toward increased benefit in ER-positive patients with tumours devoid of S cytotoxic T cells (Pheterogeneity = 0.087) and, in ER-negative patients, with tumours positive for iT-CD8+ lymphocytes (Pheterogeneity = 0.12) using imputed data. Analysis restricted to cases with complete data only showed that the presence of iT-CD8+ cells was significantly associated with increased relative benefit from epirubicin (HR = 0.60, 95% CI 0.37–0.96) compared with cases negative for iT-CD8+ cells (HR = 1.47, 95% CI 0.72–3.02; Pheterogeneity = 0.039).
Further subgroup analyses were conducted to determine whether the prognostic effect of T cells varied according to tumour cell proliferation as suggested by a recent study and according to patient age as a result of the age-related decline of the immune system known as immunosenescence. The results of these analyses are depicted in supplementary Figure S17, available at Annals of Oncology online http://annonc.oxfordjournals.org/content/25/8/1536/suppl/DC1. No significant heterogeneity was observed by tumour proliferation status or age at diagnosis.
Source...