Abstract
BACKGROUND: Grading based on histopathologic indicators cannot accurately assess the prognosis of phyllodes tumor (PT) of the breast. This article aimed to investigate the correlation between PT prognosis and clinicopathological features, treatment, and surgical margin. METHODS: The clinicopathological data of patients with pathologically confirmed PT at our institution were retrospectively collected. Univariate and multivariate Cox proportional risk models were employed to test the effects of different variables on the prognosis of PT. A nomogram to predict the 1-, 3-, 5-, and 10-year recurrence-free survival (RFS) of PT was proposed, and its discriminative ability and calibration were tested using the concordance index (C-index), area under the curve (AUC), and calibration plots. All statistical analyses were performed using R. RESULTS: A total of 342 PT patients were included, including 242 benign (70.8%), 75 borderline (21.9%) and 25 malignant (7.3%) cases. The median follow-up period was 64.5 months (range, 3-179 months), 66 PT patients had local recurrence (LR), and four patients had distant metastasis. The 1-, 3-, 5-, and 10-year RFS of the PT patients were 90.8%, 81.8%, 78%, and 76.7%, respectively. Age, fibroadenoma (FA) surgery history, treatment, mitotic activity, and surgical margin were selected as the independent factors for PT prognosis. The nomogram showed good discriminative ability and calibration, as indicated by the C-index [0.78, 95% confidence interval (CI): 0.75-0.11]. CONCLUSIONS: Independent predictors related to PT prognosis were selected to establish a nomogram for predicting the RFS of PT. This nomogram was able to objectively stratify PT patients into prognostic groups and performed well in the internal validation.