Clinical characteristics, treatment patterns, and outcomes for patients with germ-cell tumor and brain metastases at diagnosis vs. relapse.

e17014 Background: Brain metastases (BM) occur in 2-3% of all patients (pts) with GCT. Data is limited for clinical characteristics, treatment patterns, and outcomes of pts with BM at diagnosis vs. at relapse. Methods: We identified pts with GCT and BM at initial diagnosis vs. relapse from the prospectively maintained Indiana University testicular cancer database. Pts were stratified by timing of BM (diagnosis vs relapse). Treatment patterns were grouped into chemotherapy (chemo) only vs. chemo plus local therapy (radiation or surgery). Start dates for progression free survival (PFS) and overall survival (OS) were date of BM diagnosis. PFS and OS were estimated using the Kaplan–Meier method and compared using the log-rank test. Results: 205 pts were included. Median age was 29.0 (16-51.5). Primary site was testis in 183 (89.3%), mediastinal in 15 (7.3%), and retroperitoneum in 7 (3.4%). 191 pts (93.2%) had nonseminomatous disease. Predominant histology was choriocarcinoma (38%), embryonal (22%), mixed (16.6%), yolk sac tumor (8.3%), teratoma (7.3%), seminoma (6.3%). IGCCCG risk at diagnosis was poor in 85.9%, intermediate in 5.4%, and good in 7.8%. 127 pts (62%) had BM at initial diagnosis; 78 pts (38%) at relapse. For those with BM at relapse, 52 (66.7%) were at 1 st relapse; 26 (33.3%) at ≥2 nd relapse. In pts with BM at relapse, 67.9% were originally IGCCCG poor risk, with 19.2% originally good risk (p < 0.001). Choriocarcinoma was the predominant histology for 45.7% of pts with BM at diagnosis vs. 25.6% of pts with BM at relapse. In pts with BM at diagnosis, median AFP was 6.9 (1-27,000) vs. 73 (1-52,207) in pts with BM at relapse (p = 0.02). Median hCG for pts with BM at diagnosis was 140,000 (1-2,000,000) vs. 36,200 (0.5-2,576,520) for pts with BM at relapse (p = 0.0018). In those with BM at diagnosis, 66 pts (52%) received chemo plus local therapy and 61 pts (48%) received chemo alone. In those with BM at 1 st relapse, 100% received chemo with local therapy. In those with BM at ≥2 relapse, 76.9% received chemo with local therapy. Median follow-up was 4.7 yrs (0.07, 30.2). Overall, 2yr PFS was 27.5%% (21.1-34.1). 2yr OS was 62.4% (54.6-69.3). For those with BM at diagnosis, 2yr PFS was 23% (15.8-31) vs. 35.7% (24.2-47.2) for those at relapse (p = 0.046). 2yr OS was 62.7% (52.8-71.1) for those with BM at diagnosis vs. 63% (49.6-73.7) for those with BM at relapse (p = 0.72). Conclusions: Pts with BM at diagnosis were more likely to have choriocarcinoma predominant histology and high levels of hCG as opposed to pts with BM at relapse. Here, there was no OS difference from time of BM diagnosis for pts with BM at diagnosis vs. relapse.