Purpose: Uveal melanoma undergoes haematogenous dissemination, particularly to the liver. It is routine practice to screen for metastatic disease. The purpose of this study was to evaluate such screening tests and to determine their value in detecting metastatic disease at presentation.
Methods: In this study of a series of 245 patients with uveal melanoma of whom 55 (22%) died, the sensitivity, specificity, positive and negative predictive powers and likelihood ratio of the blood tests, chest radiographs and liver ultrasound investigations performed at diagnosis were determined.
Results: All the screening investigations were found to have a low yield, but the most sensitive were those for gamma-glutamyl transpeptidase (gamma GT; sensitivity 21%) and alkaline phosphatase (sensitivity 25%), with gamma GT being more specific (specificity 92%). No blood test had a positive predictive power greater than 50%. The chest radiograph and liver ultrasonography were both 100% specific but had sensitivities of 2% and 14% respectively.
Conclusions: Only the chest radiograph and liver ultrasonography had positive predictive powers greater than 50% and these tests had very poor sensitivity. We conclude that there is no good test available for screening for metastatic uveal melanoma.