Background: There are a limited number of therapeutic options available for metastatic melanoma patients progressing on the checkpoint inhibitor(s) and BRAF-targeting drugs. PARP inhibitors have been approved in the setting of BRCA1/2 germline mutations, and have demonstrated activity in the setting of HR deficiency in other cancers. Documentation of PARP inhibitors in melanoma would have clear clinical impact. We have previously shown that 21-34% of metastatic melanomas harbor at least 1 molecular aberration in the HR pathway, considered pathogenic, and likely leading to HR deficiency. Previously, we demonstrated dramatic regression of tumors harboring HR mutation(s) with PARP inhibitor treatment in a PDX model. These findings provide a strong rationale to evaluate the clinical efficacy of a PARP inhibitor in patients with advanced melanoma with HR deficiency. Methods: This is a single arm, open-label, multi-center phase II trial. The primary objective of the study is: To evaluate the clinical efficacy of niraparib in advanced melanoma patients with HR deficiency. The secondary objectives is to evaluate PFS, OS, and safety profile of niraparib. The key eligibility criteria: Advanced melanoma patients with mutation in any of the following: ARID1A/B, ARID2, ATM, ATR, ATRX, BARD1, BRCA1/2, BAP1, BRIP1, CHEK2, FANCD2, MRN11A, RAD50, RAD51, RAD54B, PALB2 (requires a genetic analysis result performed at any CLIA-certified laboratory) disease progression on PD1-antibody therapy and if BRAF mutant, BRAF/MEK inhibitors ≥ 18 years of age; ECOG PS ≤ 1; prior systemic cytotoxic therapy ≤1 regimen; no limit on number of prior immunotherapy or targeted therapy regimens. No symptomatic brain metastasis; active brain lesions ≥6 mm size, or requiring steroid treatment Patients will be treated with oral niraparib once daily (28-day cycle); the dose will be 300 mg (if weight ≥77 kg and platelet counts of ≥150,000 µL); otherwise 200 mg RECIST 1.1 will be used to evaluate clinical response every 8 weeks. Adverse events will be recorded according to CTCAE version 4.03. The primary study endpoint is overall response rate. Simon’s optimal 2-stage design is proposed. With significance level of5% and 80% power, our null hypothesis that the true response rate is 0.10 will be tested against a one-sided alternative: H0: p ≤ 10%, a response rate of no real interest VS H1: p ≥ 30%, a response rate that would be of considerable interest 10 patients will be analyzed for a response evaluation in the first stage of the study, and 19 in the second stage, for a total of 29 patients. If the response rate is 10% (1/10) or less at the end of the first stage, further accrual will be terminated. With 2 or more responses in the first stage, accrual will continue to a total of 29 patients for this cohort. The null hypothesis will be rejected if ≥6 responses are observed in a cohort of 29 subjects. Clinical trial information: NCT03925350.