Establishment Of Prognostic Model In Metachronous Bilateral Breast Cancer

Background and objective:Breast cancer(BC)is the most common cancer that threatens women's health worldwide.About 4%~7% patients with unilateral non-metastatic breast cancer will develop contralateral breast cancer.The time interval between diagnosis of bilateral lesions which is more than 6 months is classified as metachronous bilateral breast cancer(MBBC)in this research.The studies on prognosis analysis of MBBC are limited and usually consult results from unilateral breast cancer.The purpose of this study is to explore the way to optimize the prognostic evaluation system for MBBC by selecting and analyzing the factors affecting breast cancer specific death and establish a prognostic model.Methods:Based on the Surveillance,Epidemiology and End Results(SEER)database,13,304 MBBC patients diagnosed during 1990 to 2015 and followed-up for more than or equal to 3 months were retrospectively analyzed.Each patient had two records for bilateral lesions and the fields studied of each record include age at diagnosis,sex,time interval between diagonosis of bilateral lesions,race,marital status,pathological type,grade(G),tumor size(T),lymph nodes metastasis(N),estrogen receptor(ER)status,progesterone receptor(PR)status,human epidermal growth factor receptor 2(HER2)status,surgery,follow-up time and outcomes which include living at the end of the follow-up,death from cancer,and death from non-cancer causes.The competitive risk model proposed by Fine and Gray was used to establish three competitive risk models by univariate and multivariate analysis,named model for primary BC,model for contralateral BC and model for bilateral BC.The clinicopathologic factors affecting the follow-up outcomes independently was selected and subdistribution hazard ratios(SHR)with 95% confidence interval(CI)were calculated in models.SHR of each factor was used to indicated the absolute risk for cancer specific death.Receiver operating characteristic curve(ROC)and areas under the ROC curves(AUC)were used to evaluate the three prognostic models.R3.6.1 software was used to constructed the nomogram for the best of the three competitive risk models.Results:1.The median time interval between diagnosis of MBBC was 7 years,the median age at diagnosis of the first lesion was 58 years,and the median age at diagnosis of the contralateral lesion was 66 years.2.In the model for primary BC and for contralateral BC,the age at diagnosis,time interval of diagnosis,T,N and G were common independent factors affecting breast cancer specific death.The pathological type from the model for primary BC was also an independent factor,while the ER status of the contralateral lesion was an independent prognostic factor in the model for contralateral BC.3.The results from multivariate analysis in the model for bilateral BC showed the cancer specific death risk was reduced by 8% when time interval of diagnosis increased one year with 0.92 SHR(95% CI 0.89-0.95).The other variables were divided into subgroups as “primary BC/contralateral BC” form.Compared with T stages were T1 of the both lesions,the risk of cancer-specific death was increased among other subgroups.The SHR of T2/T3-T4 was the highest,5.19(95% CI2.99-8.99).SHRs of T1/T3-T4,T3-T4/T3-T4 were 4.70(95% CI 3.01-7.35)and 4.25(95% CI 2.38-7.57).SHR of T2/T1 was the lowest,1.65(95% CI 1.24-2.20).Compared with non-lymph node metastasis(N0)on both lesions,N2-N3/N2-N3 had the worst impact on survival,with 6.14 SHR(95% CI 3.82-9.88).The risk of cancer specific death in N1/N2-N3,N2-N3/N1,N0/N2-N3,N1/N1 and N2-N3/N0 decreased successively.Compared with the grade of bilateral lesions were both I-II,grade III-IV/III-IV and grade I-II/III-IV increased the cancer specific death risk to 1.63times(95% CI 1.23-2.16)and 1.34 times(95% CI 1.02-1.76),respectively,but grade III-IV/I-II didn't increased the risk.Compared with invasive ductal carcinoma of both lesions,the risk of death was increased by other pathological types of the primary BC and invasive ductal carcinoma of contralateral BC,of which the SHR is 1.40(95% CI1.08-1.81).Compared with breast conserving surgery for both lesions,simple mastectomy on both was an independent protective factor for prognosis,and SHR was0.48(95% CI 0.27-0.87).Compared with ER positive on both,negative status turned into positive reduced the risk of death by 36%(95% CI 0.46-0.88).4.The three models were evaluated by the ROC and AUC to explore a better way for predicting the risk of breast specific death.When follow-up time was 15 years,the AUC were 50.46,59.15 and 61.39 respectively.5.There were 4 situations of ER status on bilateral cancer which were positive/positive,positive/negative,negative/positive and negative/positive.The risk of cancer specific death in the 4 situations was calculated under different N stages and different grade respectively.Compared with ER positive/positive,the risk was highest in negative/negative and lowest in negative/positive.6.The independent prognostic factors from the model for bilateral BC were involved to establish the nomogram to predict the 5-year,the 10-year and 15-year MBBC specific survival.Conclusion:1.The longer time interval of diagnosis,the lower T stage,the lower N stage,the lower grade,the invasive ductal carcinoma of both lesions,the simple mastectomy performed on both,and the positive/negative ER status were independent protective factors to reduce the risk of MBBC specific death.2.When establishing a competitive risk model for predicting the risk of cancer specific death,it was more valuable to take clinical characteristics of both lesions into consideration than primary or contralateral barely.3.T stage,N stage,grade and ER status in contralateral BC overweighed those in primary BC when referring to the impact on prognosis of MBBC.4.The nomogram may help clinicians predict MBBC specific survival simply and effectively.