| Popis: |
Background We analyzed overall survival (OS) following the surgical resection of primary tumor (PT) in patients with stage IV breast cancer (S4BC). We generated a predictive model to test our hypothesis that surgery will only benefit a subset of S4BC patients, depending on properties of PT, and patient responsiveness to other treatments, as well as PT resection. Methods For our analysis, we selected 5297 S4BC patients from the SEER database. We screened for potential prognostic indicators related to patients who received PT resection (PTR) surgery, using multivariate logistic regression analysis, in our attempt to identify an optimal patient population for PT surgery. Furthermore, we assessed the performance and practical value of our nomogram using calibration curves, receiver operating characteristic curves (ROC), and decision curve analysis (DCA) of our internal validation and external validation cohorts (VC). Results Overall, we selected 5297 eligible patients with S4BC. Among them, 2276 (42.97%) underwent PTR surgery. Based on our observation, the PTR surgery was an independent correlate of prolonged median OS duration in patients with S4BC, according to multivariate Cox analysis (49 months vs. 29 months, p:0.001). S4BC patients, who underwent PTR, were then arbitrarily separated into either a training (TC) or validation cohort (VC). Next, logistic regression analysis including age, gender, TN stage, histological type, tumor location, degree of differentiation, distant metastases occurrence, and treatment was used to construct a predictive model, which was then displayed as a clinical nomogram. Conclusion We established a nomogram that identified the optimal candidates for surgical PTR among patients with S4BC. |
