To evaluate the consistency of hepatic ultrasonography (US) with staging computed tomography (CT) and magnetic resonance imaging (MRI), to analyze why US was inconsistent with CT/MRI, and to compare CT/MRI.

Reliability analysis.

Two hundred fifteen patients whose primary uveal melanoma was managed in the Helsinki University Hospital and who were diagnosed with hepatic metastases by US within 60 days of staging CT/MRI from January 1999 to December 2016 were included. Patients attended a real-life follow-up schedule including hepatic US, liver function tests (LFT), and a confirmatory CT/MRI. We evaluated the consistency of US with staging CT/MRI regarding the presence and number of metastases.

The enrolled patients underwent 215 US, 167 CT, and 69 MRI examinations, and 67% of them had biopsy-confirmed metastases. Screening was regular for 98% of the patients, and 66% were asymptomatic. US was fully consistent with CT/MRI in detecting metastases in 113 (53%) patients, in 63 (29%) CT/MRI showed more metastases, and in 16 (7%) CT/MRI showed fewer metastases than US. CT/MRI was inconsistent with US in 23 (11%) patients. The sensitivity of US in detecting metastases was 96% (95% confidence interval, 92-98). US failed to suggest metastases in 10 patients. LFT were abnormal in 6 of them, and a newly detected hepatic lesion was present by US in 4.

Hepatic US is a sensitive screening modality in detecting metastases in patients with primary uveal melanoma, if combined with LFT and, in case of any newly detected lesion, a confirmatory MRI.