How long should hepatic surveillance continue after initial treatment for uveal melanoma? A large retrospective cohort study from Sheffield, UK, suggests the answer may be shorter than many current protocols assume. Reviewing 1086 patients monitored at a UK tertiary center between 2006 and 2022, the investigators found that while 29% developed liver metastases overall, the vast majority were detected within five years of treatment – raising important questions about the efficiency of prolonged routine imaging.
The rationale for surveillance is well established. Uveal melanoma metastasizes predominantly to the liver, and detecting disease before symptoms emerge can expand options for liver-directed treatment, trial enrollment, and earlier palliative input. At Sheffield, patients currently undergo hepatic ultrasound every six months for five years (a surveillance protocol that is similar to the current Melanoma Focus guidelines), and then annually up to ten years, with MRI being used to further characterize suspicious lesions. But as the study authors note, this creates a significant imaging burden in a resource-constrained system, particularly if MRI use expands.
What makes this dataset valuable is its scale. Of the 1086 included patients, 317 developed liver metastases (29% of patients), including 39 detected at the initial surveillance scan. Among those who metastasized, the crude median time from diagnosis to liver metastasis was one year and 11 months, and 294 of 317 metastatic cases – 93% – were detected within the first five years of surveillance. Only 23 additional cases were identified in years six to 11. The cumulative metastasis curve described in the study shows a steep early rise followed by a clear flattening after year five, visually reinforcing the diminishing yield of later surveillance. In practical terms, the study suggests that surveillance becomes more than five times less efficient beyond year five.
The study also highlights which patients may warrant more tailored follow-up. Increasing T stage was strongly associated with metastatic risk, with hazard ratios rising from 2.8 for T2 disease to 11.47 for T4 compared with T1 tumors. Tumor location also mattered: ciliary body involvement carried a significantly higher metastatic risk, while choroidal and iris primaries were more favorable.
Interestingly, however, late metastases were not confined to the highest-risk group. Of the 23 patients diagnosed with metastases after five years, nine had T1 tumors and eight had T2 disease, compared with only one T4 case. That nuance complicates any simple argument for stopping surveillance purely on the basis of initial aggressiveness.
For ophthalmic oncologists, the message is not necessarily to abandon long-term follow-up, but to rethink current one-size-fits-all protocols. This study supports a more personalized strategy for patients – one that weights T stage and tumor location, and potentially integrates genomic risk markers in the future. In a healthcare environment increasingly focused on value, the data make a strong case that five years may be the point at which routine surveillance should be reconsidered, rather than automatically prolonged.
