Advanced Eye Cancer: What to Ask

Uveal melanoma spreads through the bloodstream, not through lymph nodes. NCI notes that node spread is rare. It is rare enough that sentinel node biopsy is not part of staging.

The first site found is usually the liver. Lung, bone, and tissue under the skin are also common. The COMS trials put numbers on it. They counted patients with metastases found during follow-up or at death. The liver was the only site found in 46%. Another 43% had liver plus other sites.

That pattern shapes almost every question below.

Tebentafusp is sold as KIMMTRAK. It is approved for HLA-A*02:01-positive adults with uveal melanoma that cannot be removed or has spread. The label calls it a bispecific gp100 peptide-HLA-directed CD3 T cell engager. In plain terms, it links a T cell to a target on the melanoma cell.

The FDA label is specific about selection. Patients are chosen based on a positive HLA-A*02:01 genotyping test. The test runs on a whole blood sample. The label points to FDA's companion diagnostics page for approved tests.

So the first question is simple. Has that blood test been sent, and when is the result due?

IMCgp100-202 enrolled 378 patients with metastatic uveal melanoma. All were HLA-A*02:01 positive by a central test. They were randomized 2 to 1. One arm got tebentafusp. The other got the doctor's choice of pembrolizumab, ipilimumab, or dacarbazine.

The entry rules matter for reading this. Prior systemic therapy for advanced disease was not allowed. Neither was prior liver-directed therapy. Prior surgery to remove a few metastases was allowed. Patients with untreated or symptomatic brain metastases were left out. Among those enrolled, 94% had liver metastases and 36% had a raised LDH.

The results:

• Overall survival. Median 21.7 months with tebentafusp versus 16.0 months with investigator's choice. Hazard ratio 0.51.
• Progression-free survival. Median 3.3 months versus 2.9 months. Hazard ratio 0.73.
• Objective response rate. 9.1% versus 4.8%. One complete response in the tebentafusp arm.

Note the gap between those numbers. Survival improved a lot. Tumor shrinkage stayed uncommon. That gap is worth knowing before treatment starts. Scans may not look much better even when the drug is working.

The schedule is fixed, and it steps up over the first three doses before settling into a weekly dose. Each infusion into a vein takes about 15 to 20 minutes. The label sets the amounts, so there is nothing for you to work out. Ask your team to write out your own schedule.

The monitoring is the part to plan around. The first three infusions go in an appropriate healthcare setting. Monitoring runs during the infusion and for at least 16 hours after. The reason is cytokine release syndrome. It carries a boxed warning on the label and can be serious or life-threatening.

Then it can get easier. If grade 2 or worse low blood pressure does not occur during or after the third infusion, later doses can move to an outpatient setting. Monitoring there is a minimum of 30 minutes. Ask about that switch early. It changes the logistics of every week that follows.

This is worth raising directly. NCI's clinical summary for intraocular melanoma states that no effective method of systemic treatment has been identified for metastatic ocular melanoma, and that clinical trials are an option.

The FDA has since approved tebentafusp for HLA-A*02:01-positive adults with disease that cannot be removed or has spread. So the two documents do not agree. We are not in a position to settle that for you.

What we can say is what to do with it. Take both to your appointment. Ask which one your team is working from, and whether the approved drug is an option in your case.

Two points remain accurate and useful.

Extraocular extension means tumor growing outside the wall of the eye. The outlook is poor. Gross involvement of the orbit requires orbital exenteration. But NCI states plainly that no evidence shows such radical surgery prolongs survival. Most patients with limited or walled-off extension do not have it. NCI calls the subject controversial.

Surgery to remove metastases is the other point. Single-center case series of carefully picked patients have reported longer survival. NCI's caution is direct. How much of that came from selection is unclear, so the approach is not standard.

Ask what the pathology said about cell type. It is the most used predictor after the eye is removed. Spindle-A cell melanomas carry the best outlook. Epithelioid cell melanomas carry the least favorable. Most tumors are a mixture. NCI notes there is no clear agreement on how many epithelioid cells make a tumor mixed rather than epithelioid.

NCI lists other factors too. They are tumor size, where the front edge of the tumor sits, how much the ciliary body is involved, and extraocular extension. Mitotic activity and lymphocyte infiltration can also affect outlook.

For scale: the 5-year mortality rate after metastasis from ciliary body or choroidal melanoma is about 30%, compared with 2% to 3% for iris melanomas.

Uveal melanoma is rare. Mean age-adjusted US incidence is about 4.3 new cases per million people a year. It is higher in men at 4.9 than in women at 3.7. The rate has been stable since the early 1970s. It peaks near age 70.

At that rarity, most cancer doctors see very few cases. So two questions are fair, not awkward. Does the center run uveal melanoma trials? Will the case go to a tumor board? Our page on eye cancer covers the earlier stages.

• Call 911 or go to an emergency department if you are wheezing or struggling to breathe, or you faint, in the hours after a tebentafusp infusion. Cytokine release syndrome carries a boxed warning on the label, and it can be life-threatening.
• Call your care team the same day if fever, chills, nausea, vomiting, headache, dizziness, light-headedness or a rash come on after an infusion. These are the other listed signs of cytokine release syndrome. Most of them start the same day as a dose.
• Call your care team within a day or two if a rash is spreading, or is bad enough to keep you awake. Rash affected 83% of people in the label's safety data. It is treated with antihistamines or steroids, so it is worth reporting rather than enduring.

• National Cancer Institute, Intraocular (Uveal) Melanoma Treatment (PDQ) - Health Professional Version, updated May 16, 2025, accessed August 6, 2026
• DailyMed, KIMMTRAK (tebentafusp-tebn) injection, prescribing information, SPL version 8, published May 8, 2025, accessed August 6, 2026