| Background:pN stage and breast cancer subtypes(BCS)are both well-recognized prognostic indicators.Our previous work has highlighted the interaction that exists between pN stage and BCS.The purpose of this retrospective cohort study was to compare the prognostic value of the combined pN stage and BCS with either pN stage or BCS alone.Methods :We conducted a retrospective analysis of 3924 patients with a diagnosis of invasive ductal carcinoma of the breast who underwent a surgery in the Tianjin Medical University Cancer Institute and Hospital(Tianjin,China)from January 2002 to December 2008.We used the combined pN stage and BCS to create a new variable named pNnew stage.And the pNnew stage were divided into four groups: pN0 new,pN1new,pN2 new,and pN3 new.Survival analysis performed by the Kaplan–Meier method and the log-rank test was used for single-factor analysis.For multivariate analysis,Cox proportional hazard test was applied.Cox proportional hazards models were used to estimate disease-free survival(DFS).Discriminatory accuracy of the models was compared using the area under the receiver operating characteristic curve(AUROC)and the Akaike information criterion(AIC)and the Bayesian information criterion(BIC)values.Results:A total of 3924 patients with invasive ductal carcinoma of the breast in our hospital were analyzed.All patients received modified radical mastectomies and no patients received neoadjuvant chemotherapy.After surgery,3375 patients received adjuvant chemotherapy and 664 patients received adjuvant radiotherapy.Most of the patients with ER+/PR+ tumor received hormonal therapy at least 5 years,however,only 76 of 737 patients with HER2 positive breast cancer treated with Trastuzumab because of economic reasons.Adjuvant chemotherapy regimens including cyclophosphamide+methotrexate+ fluorouracil(CMF),cyclophosphamide+adriamycin/epirubicin+fluorouracil(CAF/CEF),docetaxel+ adriamycin/epirubicin +cyclophosphamide(TAC/TEC),and docetaxel+cisplatin(TP).At the last time of follow-up.pN stage and BCS were significantly associated with disease-free survival both in univariate and multivariate survival analysis.We analyzed the interaction between pN stage and BCS,and we found that there is a statistical interaction between pN stage and BCS.The HR value range from 1 to 28.036.Accroding to the HR value,the pNnew stage were divided into four groups: pN0 new,pN1new,pN2 new,and pN3 new.1≤HR<3 represents pN0new(Luminal A N0-1,Luminal B N0,Triple negative N0),3≤HR<8 represents pN1new(Luminal A N2,Luminal B N1-2,Triple negative N1,HER-2 N0),8≤HR<15 represents pN2new(Luminal A N3,Luminal B N3,Triple negative N2,HER-2 N1-2),and HR≥15 represents pN3new(Triple negative N3,HER-2 N3).pNnew stage(P < 0.001)was also significantly associated with disease-free survival in multivariate survival analysis.Three categories of models(pN stage,BCS,and pNnew stage)were significantly associated with the DFS of breast cancer patients.However,the log-rank χ2 value associated with the pNnew stage was the largest among all the three models in the univariate analysis,which indicated that the pNnew stage had a greater statistical significance than pN stage and BCS alone.From the results of the AIC and BIC values in a Cox regression,we determined that the pNnew stage had the smallest AIC and BIC values among all the three models.Receiver operating characteristic curve analysis was also used to compare the discrimination of three models in our present study: the pN stage alone,the BCS alone,and the pNnew stage.We calculated the AUROC of these above three models for prediction of the 3-year DFS,5-year DFS and 8-year DFS respectively.The AUROC of the pN stage alone was 0.8198,0.7821,0.7429 for prediction of the 3-year DFS,5-year DFS and 8-year DFS respectively.And the AUROC of the BCS alone was 0.7434,0.7406,0.7153 for prediction of the 3-year DFS,5-year DFS and 8-year DFS respectively.To our interesting,the pNnew stage had the highest discrimination,with an AUROC of 0.8530,0.8227 and 0.7794 for prediction of the 3-year DFS,5-year DFS and 8-year DFS respectively.Conclusions:The pNnew stage(combined pN stage and BCS)had better discrimination than either pN stage or BCS alone.Our present study showed that it should be considered as the optimal variable for evaluating the prognosis of breast cancer patients in the clinic. |