| Background and objective:As the most common female malignancy worldwide,breast cancer is a highly heterogeneous disease.Although the choice of treatment based on breast cancer subtypes has reached an international consensus,precise treatment requires more detailed prognostic stratification and better measures for disease monitoring,management,and treatment tailoring.National Comprehensive Cancer Network(NCCN)guidelines recommend 21-gene recurrence score(RS;Oncotype DX)and other multi-gene testing to guide the selection of treatment strategies.However,there are still few gene expression studies based on the Chinese population,and there is a lack of corresponding standards and consensus for the precise treatment of Chinese breast cancer patients.Therefore,simple but accurate prognostic models are urgently needed to prevent the undertreatment of high-risk patients and minimize the overtreatment of low-risk patients.Predictive and prognostic inflammatory biomarkers obtained from the peripheral blood would certainly be helpful for faster,cheaper,and more individualized decision making.However,as of today,much of the relationship between the peripheral immune system and breast cancer is not uncovered.Moreover,the role of the immune inflammatory system in the invasion,progression and metastasis of breast cancer varies with different breast cancer subtypes.The identification of peripheral inflammatory biomarkers in different breast cancer subtypes will not only helps to establish strategies with Chinese characteristics for assessing the risk of recurrence of breast cancer patients,but also helps to reveal the molecular mechanism of the interaction between immune cell populations and tumor cells,and to explore new immunosuppressive targets for breast cancer subgroups.This study aims to explore the relationship between neutrophil(NEUT),lymphocyte(LYM),monocyte(MNC),platelet(PLT),C-reactive protein(CRP),albumin(Alb),neutrophil-to-lymphocyte ratio(NLR),lymphocyte-to-monocyte ratio(LMR),platelet-to-lymphocyte ratio(PLR),C-reactive protein to albumin ratio(CAR),the systemic immune inflammation index(SII),and the prognostic nutritional index(PNI)in preoperative peripheral blood and the prognosis of breast cancer patients with different molecular types,and to evaluate the clinical value of the prognostic scoring model based on inflammatory biomarkers in breast cancer subgroups.By trying to find out whether different breast cancer subtypes have different peripheral immune status,and whether peripheral inflammatory biomarkers can be used to predict the prognosis of breast cancer patients with different subtypes,this study will identify groups with different recurrence risks,and help achieve accurate diagnosis and treatment of breast cancer patients.Methods:The clinicopathological data of 561 female breast cancer patients who underwent surgical treatment and were pathologically confirmed as invasive ductal carcinoma after the operation at the Department of Oncology or General Surgery of Sir Run Run Shaw Hospital of Zhejiang University School of Medicine during the period of April 2013 to December 2015 were collected.The optimal cutoff levels for 12 systemic inflammatory markers,including NEUT,LYM,MNC,PLT,CRP,Alb,NLR,LMR,PLR,CAR,SII,and PNI,were defined by receiver operating characteristic(ROC)curve analysis.The value of different biomarkers for diagnosing breast cancer was compared by the area under the curve(AUC).The associations of systemic inflammatory markers and clinicopathological variables were evaluated using Chi-square tests or Fisher’s exact tests.Comparison of systemic inflammatory markers with disease-free survival(DFS)were then performed using Kaplan-Meier(K-M)survival curves.And the differences between the survival curves were compared by using Log rank test.The multivariate Cox regression analysis was performed on the factors that were shown to be significant on univariate analysis.The prognostic scoring model based on all independent prognostic factors selected by the multivariable Cox regression analysis was used to distinguish different prognostic groups.The concordance index(C‐index),AUC and the calibration curve were used to evaluate the discrimination and calibration of the model.And the performance between different models was compared using the C‐index and the decision curve analysis(DCA).Results:1.Among 561 breast cancer patients,241 patients were younger than 50 years old and320 patients were more than 50 years old.In the term of female physiology,menopausal and non-menopausal patients were 296 and 265 respectively.And there were 299 patients with tumor size ≤20 mm,252 patients with tumor size ≥20 mm and ≤50 mm,10 patients with tumor size ≥50 mm in the aspect of tumor size.Meanwhile,there were 145 ER-negative and 416 ER-positive patients,204PR-negative and 357 PR-positive patients,415 HER2-negative and 146HER2-positive patients.With Ki-67,180 patients were <15%,202 patients were15-30% and 179 patients were >30%.From the perspective of the number of lymph node metastasis,312 patients were with 0,187 with 1-3,40 with 4-9 and 22 with more than 10.Furthermore,the number of patients with WHO histological classification 1,2 and 3 were respectively 76,234,and 251.The clinical staging results were as follows: 317 patients were in stage I,183 in stage II,61 in stage III.In the aspect of molecular subtypes,71 patients were classified into HER2 positive(HR positive)type,75 in HER2 positive(HR negative)type,70 in triple negative type,158 in Luminal A type and 187 in Luminal B type.2.Among all 561 breast cancer patients,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 8.900,1.150,0.450,205.500,0.250,44.450,0.245,3.310,1.715,152.520,54.150,and 690.920,respectively.Chi-square tests or Fisher’s exact tests revealed that PLT,CRP,Alb,NLR,PLR,PNI,and SII were respectively related to some clinicopathological characteristics.The K-M DFS curves of the high-score group and the low-score group of 4 inflammation biomarkers,including Alb,NLR,PLR,and PNI(P<0.05),were significantly different.Univariate analysis showed that tumor size(P<0.05),lymph node metastasis(P<0.01),clinical stage(P<0.01),Alb(P<0.05),NLR(P<0.05),and PNI(P<0.05)were significantly associated with DFS.Using multivariable Cox regression,lymph node metastasis and low NLR values were independent prognostic factors for DFS.The risk of recurrence in breast cancer patients with lymph node metastasis was 1.5 times that of patients without lymph node metastasis.And the risk of recurrence of breast cancer patients with preoperative NLR>1.715 was 0.58 times that of patients with NLR≤1.715.3.Among 71 HER2 positive(HR positive)breast cancer patients,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 4.350,1.550,0.450,188.500,0.750,43.550,0.015,5.875,1.685,79.205,55.500 and 372.465,respectively.Chi-square test or Fisher’s exact test found that CAR and SII were related to some clinicopathological characteristics.The K-M DFS curves showed that there were significant differences between the high-value group and the low-value group of four inflammation biomarkers,including MNC,NLR,LMR,and PLR(P<0.05).Univariate analysis found that menopause(P<0.05),LMR(P<0.05),NLR(P<0.05),PLR(P<0.05),and PNI(P<0.05)were significantly related to DFS of HER2 positive(HR positive)breast cancer patients.However,using multivariable Cox regression,they were not independent prognostic factors.4.Among 75 patients with HER2 positive(HR negative)breast cancer,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 3.450,1.650,0.450,188.000,0.750,44.750,0.015,4.375,1.280,111.075,51.300 and 297.905,respectively.Chi-square test or Fisher’s exact test found that NEUT,CRP,CAR,and SII were respectively related to some clinicopathological characteristics.The K-M DFS curves showed that there were significant differences between high-Alb group and low-Alb group(P<0.05).Univariate analysis found that lymph node metastasis(P<0.01),clinical stage(P<0.05),and Alb(P<0.05)were significantly related to DFS of patients with HER2positive(HR negative)breast cancer.However,this association didn’t retain its significance in multivariable analysis.5.Among 70 patients with triple-negative breast cancer,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 2.650,1.750,0.350,165.500,2.850,43.300,0.065,5.875,4.115,97.165,52.750 and 217.480,respectively.Chi-square test or Fisher’s exact test found that MNC,LYM,and Alb were respectively related to some clinicopathological characteristics.The K-M DFS curves found that there were significant differences between the high-value group and the low-value group of 4 inflammation biomarkers,including PLT,NLR,LMR and SII(P<0.05).Univariate analysis showed that PLT(P<0.05),NLR(P<0.05),and SII(P<0.05)were significantly related to DFS.Using multivariable Cox regression,high NLR values was an independent prognostic factor(P<0.05).The risk of recurrence in triple-negative breast cancer patients with preoperative NLR>4.115 was 8.3 times that of patients with NLR≤4.115.6.Among 158 patients with Luminal A breast cancer,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 4.050,2.050,0.350,182.500,0.350,43.750,0.005,4.290,1.880,137.980,55.450,and 417.040,respectively.Chi-square test or Fisher’s exact test found that LYM,PLT,CRP,CAR,and LMR were respectively related to some clinicopathological characteristics.Survival analysis found that there were significant differences in DFS between the high-value group and low-value group of 5inflammation markers,including LYM,CRP,LMR,PLR,and PNI(P<0.05).Univariate analysis showed that the LYM(P<0.05),CRP(P<0.05),LMR(P<0.05),and PNI(P<0.05)were significantly related to patient’s DFS.However,they were not specifically associated with elevated risks of recurrence using multivariable Cox regression.7.Among 187 patients with Luminal B breast cancer,the best cut-off values of NEUT,LYM,MNC,PLT,CRP,Alb,CAR,LMR,NLR,PLR,PNI,and SII obtained by ROC curve analysis were: 2.450,2.150,0.350,204.500,0.650,44.650,0.015,6.420,1.690,88.895,54.150 and 410.475,respectively.Chi-square test or Fisher’s exact test found that MNC,PLT,LMR,CRP,CAR,PLR,and SII were respectively related to some clinicopathological characteristics.Survival analysis found that there were significant differences in DFS between the high-value group and the low-value group of 4inflammatory biomarkers,including Alb,CRP,CAR and PLR(P<0.05).Univariate Cox regression analysis showed that lymph node metastasis(P<0.05),Alb(P<0.05),CRP(P<0.05),CAR(P<0.05),and PLR(P<0.05)were significantly related to DFS in this study.Using multivariable Cox regression,high Alb values,low CAR values and high NLR values were independent prognostic factors(P<0.05).In Luminal B breast cancer,the risk of recurrence of patients with Alb>44.650 was 2.32 times that of patients with Alb≤44.650.The risk of recurrence of patients with CAR>0.015 was0.38 times that of patients with CAR≤0.015.And the recurrence risk of patients with PLR>88.985 was 6.75 times that of patients with PLR≤88.985.8.A prognostic scoring model based on CAR,Alb,and PLR was established in Luminal B breast cancer patients.The model divided patients into three different prognostic groups,and the DFS was significantly different between them(P<0.05).The AUC of the ROC curve was 0.763,and the C index was 0.76.The calibration curve showed that the actual 3-year and 5-year disease-free survival rates were basically consistent with the predicted disease-free survival rates,indicating that the model has good discrimination and calibration.Compared with the TNM staging model and the clinical staging model,the prognostic scoring model had a better C index and DCA,indicating that it has better prognostic prediction ability and clinical net benefit.Conclusions:1.Preoperative NLR is an independent prognostic marker for DFS in breast cancer patients.For all breast cancer patients,low preoperative NLR(NLR≤1.715)indicates a poor prognosis.For patients with triple-negative breast cancer,high preoperative NLR(NLR>4.115)indicates a poor prognosis.2.Preoperative Alb,CAR and PLR are independent prognostic markers for DFS in patients with Luminal B breast cancer.In patients with Luminal B breast cancer,high Alb(Alb>44.650),low CAR(CAR≤0.015)and high PLR(PLR>88.985)indicate a worse prognosis.3.The prognosis scoring model based on CAR,Alb,and PLR can predict the prognosis of Luminal B breast cancer patients,which has a potential to identify patients with different risks of recurrence,and help achieve individualized and precise treatments. |
PDF Full Text Request