Aim To review all cases of suspected vitreous seeding of treated or untreated uveal melanoma at our clinic and to compare clinical, cytological and histological findings with patients’ survival.

Methods Retrospective non-randomised study of 23 patients with consecutive uveal melanoma who underwent diagnostic vitrectomy in our clinic between January 2000 and November 2013. Reason for vitrectomy was suspected dissemination of tumour cells inside the eye. Treated as well as treatment-naïve primary uveal melanomas were included in this study. Follow-up data of all patients were collected.

Results The study included 23 patients with a mean age of 66 years. Four patients presented pigmented vitreous debris at initial presentation prior to treatment of the uveal melanoma. All but one of these four patients has been enucleated as a consequence of cytology-proven vitreous spreading of vital melanoma cells. The remaining 19 patients presented pigmented vitreous debris at a mean of 60 months following local tumour treatment. Thirteen of these patients had been treated with a ruthenium plaque (mean scleral dose 1295 Gy, mean apex dose 152 Gy), three with binuclid plaque (mean scleral dose 1005 Gy, mean apex dose 70 Gy) and three with proton beam radiation. Of the 19 patients, 10 showed only melanophages in the vitreous specimen, while the remaining 9 patients had vital tumour cells in vitreous cytology. Four out of these nine patients have been enucleated in the course of follow-up. During follow-up of our cohort of 23 patients, 4 patients died, but only 1 of them due to metastatic disease.

Conclusion The outcome of this small cohort study shows that obtaining a vitreous specimen helps to distinguish melanophages from vital tumour cells. We could not observe an increased risk of metastasis in patients who showed melanoma cell dissemination inside the eye, compared with those patients only showing melanophages. We therefore suggest to carefully re-evaluate the necessity of enucleation in every patient.