Abstract Introduction. Intracranial germ cell tumor is an Asian-centric tumor, showing great variation in geographical distribution, with an incidence 5-8 fold greater in Asia/East Asia, than Western countries. Germinoma is one histological subtype of intracranial germ cell tumors. Current standard of care for germinoma, is chemoradiotherapy. Metastatic/recurrent germinoma have poor prognosis. There is currently no targeted therapy. Purpose: To create patient-derived orthotopic xenograft (PDOX) models of distant metastases in brain germinoma beyond brain boundaries to recapitulate sequential metastases of original patient tumors. These tumor models will serve as clinically accurate metastatic models for preclinical testing, replicating patient's pattern of tumor progression/metastases. There is a lack of such metastatic brain tumor models in the field. c-KIT is the most common mutation in germinoma. We evaluated preclinical efficacy of a c-KIT inhibitor. Experimental Procedures. Most PDOX brain tumor models are created by implanting patient-derived tumor cells in cerebrum/cerebellum. Germinoma originates from pineal gland/suprasellar region. We implanted germinoma cells nearer location of tumor origin. We were successful and achieved consistent tumor implantation into this location (100% post-surgery survival, 0% surgical mortality, 0% post-surgical neurological deficits). Summary. We successfully established two models with sequential metastases. Model 1 was created from metastatic cells from spinal recurrence (Patient 1) post-chemo-radiotherapy. This model faithfully created skull base metastases in 100% of murine replicas (n=9). Both tumor cells from brain and skull base metastases (n=7) were serially transplantable. Our findings were replicated (n=5) and consistent. Model 2 was created from metastatic cells of cerebrospinal fluid dissemination (Patient 2). Brain surface metastases were formed recapitulating patient tumor. Both models recapitulated a similar path of metastases in craniospinal axis, corresponding to respective patients. c-KIT inhibitor, was effective in significantly prolonging animal survival in Model 2 (p=0.0046) but not in Model 1 (p=0.0541). In Model 2, c-kit inhibition in-vivo was efficacious and resulted in 3-fold effect, (1) Significantly improved animal survival, (2) Macroscopically reduced frequency, severity of brain surface metastases (Controls=50%, Treated=14%, lesser in severity), (3) Microscopically reduced tumor burden to minimal residual disease limited to cerebrospinal fluid outline. Control mice brains were largely replaced by tumor bulk, while treated mice had minimal residual tumor burden. Conclusions. Understanding metastatic patterns in such aggressive models is relevant in biologic targeting for preclinical studies and clinical decision to extend radiotherapy field in recurrent disease. Citation Format: Shiying Huang, Sekar Karthik, Jack M Su, Qi Lin, YuChen Du, Ching C Lau, Adesina Adekunle, Angela Major, Kam-Man Hui, M. Tarek Elghetany, Xiaonan Li, Wan-Yee Teo. Targeting metastatic brain germinoma using PDOX tumor models with sequential metastases [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 634.