Background: Patients with advanced stage of melanoma may develop leptomeningeal disease (LMD), a rare form metastasis in the meninges. Unfortunately, there are currently no rationally designed treatments for melanoma-associated LMD (M-LMD) and the prognosis is dismal; with survival measured in weeks. A major barrier to developing effective therapies for this disease is the absence of model systems, such as patient-derived cerebral spinal fluid-circulating tumor cell (PD-CSF-CTC) models to identify and assess novel therapeutics. Added to this challenge is the difficulty of collecting and analyzing M-LMD specimens in the clinic and at autopsy. To address these hurdles, we have made considerable progress recently in overcoming these barriers via CSF and tissue collection (in clinic and rapid autopsy) and successfully propagating CSF-CTCs from M-LMD patients in vitro and in xenograft mouse models. These valuable resources have allowed us to perform proteomic and transcriptomic analyses of CSF and PD-CSF-CTCs, respectively. By comparing M-LMD -omics data against those from normal brain and extracranial disease, we have found unique LMD-specific biological pathways. In this study, we identified clinical compounds that could target these biological pathways and have efficacy against PD-CSF-CTCs in vitro and in vivo. Methods: PD-CSF-CTCs were propagated from individual M-LMD patients. Next, we have developed a 384-well high throughput cell-based assay and performed a pilot screening of more than 1,400 FDA-approved small molecule compound library using Echo to identify pharmaceutics that inhibit cell proliferation. Clinical compounds with the highest sensitivity were selected for validation of efficacy in vivo via intrathecal (IT) delivery of patient-derived cell lines to establish M-LMD xenografts. Results: Of the 1,436 FDA-approved small molecule compounds, 57 (~3.9%) exerted 95% proliferation inhibition and 20 (~1.4%) had 100% killing effect in PD-CSF-CTCs and murine melanoma cell lines. The compounds with the highest sensitivity include ponatinib (EC50: 1.85 – 4.06e^-06), sorafenib (EC50: 9.57 – 9.77e^-06), ceritinib (EC50: 1.84 – 2.05e^-06) and homoharringtonine (HHT) (3.63 – 4.11e^-08). In a randomized murine M-LMD preclinical trial, we selected HHT, since it is a cephalotaxine ester compound that can penetrate the brain. Our results show that HHT was well-tolerated in vivogiven systemically or IT via a murine Ommaya. We found M-LMD mice that received 14.5ng HHT (g.d. IT) showed significant prolonged median survival (control vs. HHT, P value: 0.0006; Mantel Cox test). Conclusions: This is the first demonstration of an approach that allows for the rational development of therapeutics in M-LMD. Results will provide crucial new insights into biology and will identify drug candidates that may be suitable for future clinical trials.