8523 Background: The median overall survival (OS) for metastatic uveal melanoma is less than 6 months with a median progression free survival(PFS) of 3 months. No systemic or regional therapy has shown a survival advantage over best supportive care. Despite pre-clinical evidence suggesting anti-tumour activity for the KIT tyrosine kinase inhibitor imatinib, there are several negative phase II trials in unselected patients. Our primary aim was to test the efficacy of imatinib in patients with prospectively-tested c-kit immunopositive metastatic uveal melanoma. Secondary aims included assessment of toxicity and patient recruitment. Methods: A phase II UK multicentre single-arm, two-stage Gehan design recruited 25 evaluable patients receiving imatinib 400mg OD until progression/unacceptable toxicity. Primary efficacy outcome was PFS at 3 months. Secondary outcomes were OS, overall PFS, disease response (RECIST) and toxicity. Prospective sample collection for putative biomarkers was included. Results: After 16.6 months 37 patients were screened, with 25 were registered and included in final efficacy analyses. The sample included PS 0-1 patients with a median age of 63 yrs and a median 9.3 months from diagnosis of metastatic disease. 82% had high LDH levels (>460IU/L), 65% had no previous treatment for metastatic disease and 8% did not have liver involvement by metastasis. Preliminary final results indicate the estimated proportion of patients progression free at 3 months is 0.24 (95% CI, 0.09 to 0.45). Median PFS and OS were 12.0 weeks (95% CI,11.6 to 14.3) and 29.6 weeks (95% CI, 19.3 to 61.0) respectively. Two patients had confirmed PR with response duration of 93 and 112 weeks, respectively, despite the absence of mutations in exons 11, 13, and 17 of the c-KIT gene. Conclusions: The trial successfully recruited to target in this rare disease area failing to convincingly show improved PFS in a selected cohort of c-kit immunopositive patients. Both patients with PR experienced long periods of disease control. Response was not dependent upon the presence of activating mutations in KIT. Further translational studies are ongoing to determine putative biomarkers of response.