Establishment And Validation Of Survival Prediction Model In Patients With Ⅳ Breast Cancer

Objective: Currently,there are no precise guidelines for the treatment of metastatic breast cancer.If the mortality risk of patients with distant metastatic breast cancer can be predicted more accurately,it will be of great help to provide reference for clinicians’ decision-making.This study retrospectively analyzes the clinicopathological characteristics of patients with distant metastatic breast cancer and develop nomogram.Method: This study collected clinicopathological characteristics of 8349 patients with de novo metastatic breast cancer from 2010 to 2015 using SEER database.8349 patients were randomly divided into a development group(n=5566)and validation group(n=2783)at a ratio 3:2 by R language.Firstly,the independent prognostic factors were identified using univariate and multivariate Cox proportional risk regression analysis.Then,the nomogram of OS and BCSS in patients with metastatic breast cancer was constructed using these factors to predict the 1-,3-and 5-year mortality risk.Calibration curves were used to assess the consistency between predictive and actual observations value.In both development group and validation group,the discrimination was measured by the concordance index and the area receiver operating characteristic curve.Results:1.Cox proportional risk regression analysis showed that age,race,marital status,grade,T stage,N stage,tumor subtypes,brain metastasis,liver metastasis,lung metastasis,hemotherapy,Surgery were independent prognostic factors for OS and BCSS in metastatic breast cancer.Radiation therapy were independent prognostic factors for OS.Based on these factors,the nomograms for predicting the OS and BCSS were constructed.2.In the development group,the C-index for OS nomogram was0.737(95%CI: 0.732～0.742),and the AUCs for 1-,3-and 5-year OS were0.795,0.770,0.765,respectively.In the validation group,the C-index for OS nomogram was 0.740(95% CI: 0.733～0.747)and the AUCs for 1-,3-and5-year OS were 0.791,0.781,0.768,respectively.C-index and AUC indicated that OS prediction model had satisfactory discrimination in internal and external validation.3.In the validation group,the C-index for BCSS nomogram was0.735(95%CI: 0.730～0.740),and the AUCs for 1-,3-and 5-year OS were0.791,0.769,0.764,respectively.In the validation group,the C-index for BCSS nomogram was 0.737(95% CI:0.730～0.744)with AUCs for 1-,3-and 5-year BCSS of 0.793,0.775,0.761,respectively.C-index and AUC indicated that BCSS prediction model had satisfactory discrimination in internal and external validation.4.In the development group and validation group,the calibration curves showed that the 1-,3-,and 5-year mortality risk of patients with metastatic breast cancer predicted by the nomogram of OS and BCSS established by us were in high agreement with the actual observed values.Conclusion:1.Older age,black,single status,high pathological grade,tumor subtypes,site of metastatic,chemotherapy and surgery are independent influencing factors for survival and prognosis of patients with metastatic breast cancer.2.The nomogram established by us can accurately predict the mortality risk of the patients with metastatic breast cancer and provide a valuable personalized scoring scale for patients.