Establishing a predicted model to evaluate prognosis for initially diagnosed metastatic Her2-positive breast cancer patients and exploring the benefit from local surgery

Abstract

Background For patients initially diagnosed with metastatic Her2-positive breast cancer (MHBC), we intended to construct a nomogram with risk stratification to predict prognosis and to explore the role of local surgery. Methods We retrieved data from the Surveillance, Epidemiology, and End Results (SEER) database. Kaplan–Meier (KM) method and log-rank test were used for the selection of significant variables. Cox regression analysis and Fine-Gray test were utilized to confirm independent prognostic factors of overall survival (OS) and breast cancer-specific survival (BCSS). A nomogram predicting 1-year, 3-year, and 5-year OS was developed and validated. Patients were stratified based on the optimal cut-off values of total personal score. KM method and log-rank test were used to estimate OS prognosis and benefit from local surgery and chemotherapy. Results There were 1680 and 717 patients in the training and validation cohort. Age, race, marriage, T stage, estrogen receptor (ER) status, visceral metastasis (bone, brain, liver and lung) were identified as independent prognostic factors for OS and BCSS, while histology was also corelated with OS. C-indexes in the training and validation cohort were 0.70 and 0.68, respectively. Calibration plots indicated precise predictive ability. The total population was divided into low- (<141 points), intermediate- (142–208 points), and high-risk (>208 points) prognostic groups. Local surgery and chemotherapy brought various degrees of survival benefit for patients with diverse-risk prognosis. Conclusions We constructed a model with accurate prediction and discrimination. It would provide a reference for clinicians' decision-making. Surgery on the primary lesion was recommended for patients with good physical performance status, while further study on optimal surgical opportunity was needed.