Background: BM is an independent adverse prognostic factor that can lead to treatment complications and failure in pts with mGCT. We aimed to establish an effective and practical model for prediction of BM in mGCT. Methods: 2,256 consecutive pts with mGCT treated at Indiana University between January 1990 and September 2017 were identified. Pts were divided into 2 categories: BM present (N = 144) and BM absent (N = 2112). Kaplan-Meier methods were used to analyze progression free survival (PFS) and overall survival (OS). Logistic regression was used to determine a predictive model for whether BM was present. The data was separated 50/50 into training and validation datasets with equal numbers of events in each. Results: Baseline characteristics for 2 groups are listed in Table. 2-yr PFS and OS for pts with vs without BM: 17% vs 66% (p < 0.001) and 62% vs 91% (p < 0.001) respectively. Among the 144 pts with BM, 64 (44%) had radiation only (whole-brain radiotherapy or gamma knife), 21 (15%) had BM-surgery only, 14 (10%) had both radiation and BM-surgery. 45 pts (31%) did not receive local therapy for BM. A stepwise selection was used to determine the best model with p < 0.15 as the entry and staying criteria. The model with the largest ROC AUC was used moving forward. The model was tested in the validation dataset. A model was generated including age at diagnosis≥40 (1 point), presence of pulmonary metastases (3 points), bone metastasis (1 point), pre-chemotherapy hCG≥5000 (1 point), and choriocarcinoma predominant histology (1 point). Patients with 0 points had a 0.4% probability of BM, 1 point: 1%, 2 points: 2.6%, 3 points: 7%, 4 points: 16%, 5 points: 32%, 6 points: 56%, and 7 points: 77%. Details regarding analysis in training and validation datasets will be presented. Conclusions: The prediction model developed in this study demonstrated discrimination capability of predicting BM occurrence and can be used by clinicians to identify high-risk pts.