To explore the potential factors related to the recurrence of benign and borderline breast phyllodes tumor (PT) and establish a nomogram to predict the recurrence rate of PT.

The clinicopathological and imaging data of 65 patients with benign and borderline PT who were treated in Dongguan People’s Hospital from June 2016 to December 2019 were retrospectively analyzed. The univariate and multivariate logistic regression were used to analyze the independent risk factors for recurrence, and a nomogram was constructed accordingly. The receiver operating characteristics (ROC) curve was drawn and the area under the curve (AUC) was calculated. A calibration curve was established by bootstrap method to evaluate calibration performance. The clinical utility of this predictive model was demonstrated by decision curve analysis (DCA).

Univariate logistic regression analysis showed that the time interval between mass discovery and treatment (TIMDT)>6 months, uneven MRI enhancement pattern, ultrasonic findings(irregular shape, uneven edge, tumor lobulation, uneven internal echo, punctate strong echo, moderate/abundant blood flow signal and cystic structure) were related to PT recurrence (all P<0.050). Multivariate logistic regression analysis showed that TIMDT>6 months (OR=32.230, 95%CI: 2.343-443.367, P=0.009), uneven MRI enhancement pattern (OR=16.786, 95%CI: 1.030-273.431, P=0.048) and ultrasonic tumor lobulation (OR=14.861, 95%CI: 1.155-191.205, P=0.038) were independent risk factors for PT recurrence. A nomogram was constructed based on these three independent risk factors. The AUC of ROC was 0.906 (95%CI: 0.811-1.000), the sensitivity was 83.3% and the specificity was 88.7%. The calibration curve of the model showed a good calibration efficiency. The DCA curve displayed high clinical net benefit from predicting PT recurrence at a threshold of 0.04-0.96.

The benign and borderline breast PT patients with the following features (TIMDT>6 months, uneven MRI enhancement pattern, and ultrasonic tumor lobulation) are in high risk of recurrence. The nomogram based on these three factors shows a strong ability to predict the recurrence of PT, indicating high potential in clinical application.