Background: Tandem HDC with carbo/etoposide (CE) is curative for a portion of relapsed GCT pts. However, outcomes of refractory or high-risk relapsed pts remain poor. We tested a new HDC regimen of GemDMC, based on DNA damage repair inhibition. We combined BEV with HDC given their potential synergy and the high vascularity of GCT metastases. Methods: Eligibility: Intermediate (int)/high-risk (Beyer Model), creatinine ≤1.8 mg/dL and adequate organ function. HDC included BEV (5 mg/kg) preceding GemDMC (HDC #1) and ifosfamide/CE (ICE) (HDC #2). Following accrual of 42 pts, we amended the trial omitting BEV. The trial was powered to distinguish a target 50% 2-yr RFS from an expected 25% in this population. Results: We enrolled 69 male pts in cohorts 1 (BEV, N=42) and 2 (no BEV, N=27) (Table). Pts were heavily pretreated and most had refractory tumors. Main AE: mucositis and renal (4 HDC-related deaths in cohort 1, 1 in cohort 2). Tumor markers normalized in 90% pts with active tumors at HDC. After HDC, 19 pts were in CR and 28 in PRm- (of these, 22 had residual lesions resected with no viable tumor found in 20/22, 2 xRT, 4 monitored). Median f/u = 39 (2-105) mo. The 2-yr RFS rates in cohorts 1 and 2 = 52% and 78%, respectively. Their respective 2-yr OS rates = 55% and 81%. Conclusions: Sequential HDC with GemDMC–ICE shows encouraging outcomes in heavily pretreated and refractory GCT, exceeding the anticipated results. Addition of BEV increases toxicity but not tumor control. Clinical trial information: NCT00936936