Ideaya's Melanoma Drug Clears Trial. The Number That Matters Most Is Being Left Out.

When Ideaya Biosciences and Servier announced Tuesday that their drug combination had cleared the primary endpoint in a Phase 2/3 trial for metastatic uveal melanoma, the wire led with progression-free survival: 6.9 months versus 3.1 months on standard therapy. That's a meaningful result. It's also not the most important number in the data.

The number that should be getting attention is 21.1 months.

That is the median overall survival Ideaya reported last October from an earlier Phase 2 trial called OptimUM-01, using the same darovasertib-and-crizotinib combination in first-line metastatic uveal melanoma patients. The historical median overall survival for this disease is approximately 12 months, based on a published meta-analysis. The combination is approaching a near-doubling of survival in a cancer with zero approved treatment options and a median time-to-death after metastasis of roughly a year.

Uveal melanoma is rare — about 7,095 new cases diagnosed annually in the United States — but the biology is well-understood. More than 90% of uveal melanomas carry mutations in GNAQ or GNA11, genes that drive the cancer's growth through the PKC signaling pathway. Darovasertib is a PKC inhibitor. Crizotinib, originally approved for ALK-positive lung cancer, is a cMET inhibitor that appears to work synergistically with darovasertib in this context. The combination has FDA Breakthrough Therapy Designation in neoadjuvant primary uveal melanoma and Fast Track Designation for metastatic disease.

The Phase 2/3 trial, OptimUM-02, enrolled 435 first-line patients who were HLA-A02:01-negative — a population that cannot use tebentafusp, the only approved systemic therapy for metastatic uveal melanoma, which requires HLA-A02:01 positivity. Ideaya retained US commercial rights; Servier holds ex-US rights under a September 2025 licensing deal worth $210 million upfront plus up to $320 million in milestones and royalties.

The primary endpoint — progression-free survival by blinded independent central review — was met with a hazard ratio of 0.42. That's a 58% reduction in the risk of progression or death. In a disease where three months of PFS is the floor, six months of median PFS is genuinely meaningful. The overall survival data from OptimUM-02 is not yet mature; the OS signal from OptimUM-01 is the best evidence available that this drug combination is doing something deeper than delaying progression.

What makes this story complicated is the timeline. The 21.1-month OS came from a single-arm Phase 2 trial. Randomized Phase 3 data on OS is still accruing. It's possible — even likely — that the magnitude of the OS benefit narrows when the randomized data mature. A near-doubling of survival in an uncontrolled study is hypothesis-generating, not practice-changing. The FDA will want the Phase 3 OS data before granting full approval, and the company will need those results to differentiate from tebentafusp in the HLA-A*02:01-positive population.

But for the 40 to 50% of uveal melanoma patients who develop metastatic disease — predominantly in the liver, with limited options beyond locoregional therapy — the OS signal is not a footnote. It is the point. A therapy that doubles median survival in a disease with no approved alternatives would represent a fundamental shift in what metastatic uveal melanoma means for patients. That is worth covering before the randomized data land, because the community of patients and oncologists already knows what they are seeing.

The deal structure also bears watching. Ideaya retained US rights. If the Phase 3 OS data hold, Ideaya is sitting on a rare disease asset with a clear regulatory path, Breakthrough Therapy Designation already in hand, and a partner handling ex-US commercialization. At a moment when oncology deal flow is slowing, a validated Phase 2/3 readout in an indication with no approved systemic therapy is unusual.

What happens next: the companies plan to file for approval based on the PFS endpoint. FDA will review. OS data will continue to mature. The real answer — does this combination actually extend life, or does it primarily delay progression — won't be known for another 12 to 18 months. That uncertainty should be stated honestly. But the 21.1-month number is not noise. It is the reason researchers in the field are paying attention.