Why Uveal Melanoma Surveillance Is Harder Than It Looks
Uveal melanoma accounts for roughly 5% of all melanoma diagnoses in the United States. Its death rate is higher than expected for how often it occurs. The liver is where it spreads most often, and once cancer reaches the liver, treatment options narrow sharply. A study published in the British Journal of Cancer tracked 1,086 patients at a UK center from 2006 to 2022. It found that 315 patients (29%) developed cancer throughout their body - with 93% of those cases found within five years of eye treatment.
The eye tumor is usually controlled. The real challenge comes later, in the months and years after eye treatment, when the question shifts from controlling the tumor to detecting whether cancer has spread.
What Gene Expression Profiling Tells Clinicians
Gene expression profiling (GEP) - also called a gene expression profiling assay, or GPA - is now the standard tool for measuring the risk of spread after uveal melanoma treatment. The 15-gene profile classifies tumors into three groups: Class 1A (lowest risk), Class 1B (intermediate risk), and Class 2 (highest risk).
For Class 2 tumors, research suggests about 50% of patients stay free of spread for three years and about 28% for five years. A 2024 study in the Journal of Clinical Oncology - the COOG2.1 trial - tested a classifier that combines 15-GEP with PRAME RNA expression. This improves accuracy but adds complexity. Monitoring systems that use just a high/low flag won't work for this kind of multi-variable risk.
Research shows that 93% of high-risk Class 2 patients are referred to medical oncology for follow-up, compared with 51% of Class 1 patients. The GPA result drives what happens next - which is an operational challenge.
The Velocity Gap: Where the Process Breaks Down
GPA results don't automatically create surveillance schedules. That's the core problem.
When a GEP result reaches a lab system or an EHR, several steps must happen: the result is reviewed and classified, the clinician picks a surveillance protocol, an imaging order is placed, and the patient is scheduled. In a busy oncology clinic, each step involves a person, a queue, and competing priorities.
The velocity gap is the delay between a GPA result and the first scheduled surveillance scan. For Class 2 patients, experts recommend a 9-month interval to minimize the gap between a negative and a positive scan while limiting total exams. But liver metastasis tumors in untreated uveal melanoma can double in size every 30 to 80 days, according to a 2025 systematic review of uveal melanoma surveillance guidelines. A missed 6-month imaging window in a Class 2 patient isn't just an administrative problem. It can mean catching a tumor when it's one to two times larger than it would have been.
High-risk patients typically need liver imaging - alternating between MRI and ultrasound - every 6 months for up to 10 years. For the highest-risk Class 2 patients, experts recommend imaging every 3 to 4 months, according to research on standard versus enhanced surveillance protocols for high-risk uveal melanoma patients. Tracking these intervals by hand across 30 to 50 active cases is impossible for one clinic coordinator managing multiple disease groups.
Imaging Sensitivity and the Monitoring Ceiling
Ultrasound detects liver cancer in 96% of cases (95% confidence interval: 80% to 99%) and correctly identifies non-cancerous findings 88% of the time, based on a comparative analysis of surveillance methods. That's strong detection performance - if the scan happens on schedule.
Detection rates assume scans happen when planned. A test that works 96% of the time in a study loses real value when many scheduled scans are delayed by two months or more. The problem isn't the imaging type. It's the scheduling system.
There's also an issue with overlooking low-risk cases. Some Class 1A patients have additional risk factors - larger tumors, ciliary body involvement, or PRAME expression - that may call for closer follow-up than their GPA class alone suggests. The COOG2.1 study addressed this directly. A clinic that uses GPA class as the only trigger for imaging intervals may miss a group of patients whose individual tumor characteristics warrant more frequent scans.
Protocol Divergence Across Providers
The same systematic review found major differences across institutions in imaging type (MRI vs. ultrasound vs. alternating), how long to scan (5 years vs. 10 years vs. lifelong), and how often to image. Some national guidelines call for lifelong liver imaging every 6 months for high-risk patients. Others stop recommending it after 10 years.
These differences create handoff problems. Uveal melanoma patients get care from ophthalmology, medical oncology, and radiology. Without one person or system owning the follow-up schedule, each provider may assume another has booked the next scan. In a British Journal of Cancer study, patients who got imaging on schedule had better outcomes. Adherence dropped when responsibility for follow-up was unclear across departments.
For hospital administrators, this is an operational risk as much as a clinical one. Uveal melanoma is rare enough that most health systems don't make it a priority for standardized protocols. Most large oncology centers see only a handful of new cases each year. That rarity makes manual tracking seem doable - until a surveillance gap appears during a retrospective review.
For related scheduling issues in melanoma care, see our post on closing melanoma follow-up gaps in oncology clinics, which covers the scheduling problems that affect post-treatment surveillance across melanoma types.
What a Structured Surveillance Platform Can Address
A platform designed for oncology clinic workflow can reduce velocity gaps at several points.
- Result-triggered scheduling: When a Class 2 result arrives, the system schedules the next imaging without a coordinator manually checking protocol documents.
- Interval tracking across the active panel: A clinic managing 30 to 50 active post-treatment patients needs a live view of who is on schedule and who is overdue.
- Composite risk inputs: As prognostic tools add PRAME, tumor diameter, and cytogenetic markers to GPA class, scheduling systems need to handle multi-variable rules, not just single-test flags.
For similar challenges in other cancers, see our posts on LDH monitoring in mantle cell lymphoma and AFP velocity tracking in hepatocellular carcinoma surveillance. The mechanics are similar, and both focus on liver-based monitoring.
The Administrative View
Uveal melanoma's rarity means it rarely shows up on health-system dashboards as a compliance risk. Most centers see only a few cases per year. But the stakes per case are high, and the surveillance window for Class 2 patients is narrow. A single missed imaging window can mean catching a tumor two scans later when it's much larger - exactly the kind of event that triggers protocol reviews.
The answer isn't more paperwork. It's a scheduling system that automatically maps GPA results to imaging intervals, assigns responsibility, and flags missed appointments.
Demos are 30 minutes. We'll show you how GPA results flow to scheduling on your clinic's live data, including multi-year interval tracking for Class 2 patients. Book a demo to see Rucja in action.