Darovasertib Plus Crizotinib Delivers PFS Gains in HLA-A2–Negative Metastatic Uveal Melanoma

The combination of darovasertib and crizotinib (Xalkori) led to a clinically meaningful and statistically significant improvement in progression-free survival (PFS) vs investigator’s choice of therapy in patients with HLA-A*02:01–negative metastatic uveal melanoma, meeting the primary end point of the phase 2/3 OptimUM-02 trial (NCT05987332).¹

Data from the trial, which were presented at the 2026 American Society of Clinical Oncology Annual Meeting with a median follow-up of 7.4 months, revealed a median PFS of 6.9 months (95% CI, 5.6-8.3) in the combination arm vs 3.1 months (95% CI, 1.8-4.2) in the control arm per blinded independent central review ([BICR] HR, 0.42; 95% CI, 0.30-0.59; P < .0001). Median investigator-assessed PFS was 6.7 months (95% CI, 5.6-8.2) and 2.7 months (95% CI, 1.7-4.1), respectively (HR, 0.36; 95% CI, 0.26-0.50; P < .0001).

Additionally, PFS favored the combination in subgroup analysis, regardless of age, gender, ECOG performance status, largest single-target lesion size, or metastatic pattern.

“The OptimUM-02 results support a potential new therapeutic standard for a disease with limited treatment options and poor prognosis,” lead study author Marlana M. Orloff, MD, the Alexander & Johnston Family Endowed Clinical Director in Uveal Melanoma Professor at Thomas Jefferson University, and an associate professor at Jefferson Health in Philadelphia, Pennsylvania, said in a presentation of the data.

Why investigate PKC inhibition in uveal melanoma?

Uveal melanoma is the most common ocular cancer in adults, but metastatic disease carries a poor prognosis with a median PFS of 2.8 months and median overall survival (OS) of 10 to 12 months.^2,3 Mutations in GNAQ or GNA11 that activate protein kinase C (PKC) are detected in more than 95% of uveal melanomas, making PKC a rational therapeutic target. No FDA-approved treatments currently exist for patients with HLA-A*02:01–negative metastatic uveal melanoma, further underscoring a target population with high unmet clinical need.

Darovasertib is a first-in-class, oral PKC inhibitor with demonstrated clinical activity across both HLA-A*02:01–positive and –negative metastatic uveal melanoma.^4,5 The agent is being evaluated in combination with crizotinib, an oral MET inhibitor that has shown complementary antitumor activity with darovasertib both preclinically and clinically, based on a rationale of parallel pathway activation through hepatocyte growth factor and MET overexpression.⁵

What was the trial’s design?

The phase 2/3, multicenter, multi-arm, multistage, open-label study enrolled patients with HLA-A*02:01–negative metastatic uveal melanoma. In phase 2a, patients were randomly assigned 2:2:1 to first-line therapy with darovasertib at 300 mg plus crizotinib at 200 mg twice daily, darovasertib at 200 mg plus crizotinib at 200 mg twice daily, or investigator’s choice of pembrolizumab (Keytruda), ipilimumab (Yervoy) plus nivolumab (Opdivo), or dacarbazine. In phase 2b and phase 3, patients were randomly assigned 2:1 to darovasertib at 300 mg plus crizotinib at 200 mg twice daily, or investigator’s choice of the aforementioned 3 therapies.

The primary end point of phase 2b was PFS by BICR. Secondary end points included PFS by investigator, objective response rate (ORR) by both BICR and investigator, duration of response (DOR), 12-week disease control rate (DCR), and safety. The primary end point of the phase 3 portion is OS, which will be shared once the data mature.

All patients in the intention-to-treat population or the phase 2b portion of the trial were evaluated for efficacy per BICR and investigator: 210 in the 300-mg darovasertib arm and 103 patients in the investigator’s choice arm. Safety was evaluated in all patients in phase 2a and 2b who received at least 1 dose of the 300-mg darovasertib regimen (n = 239) or the control (n = 100).

How did statistical assumptions shape the trial design?

A total of 130 PFS events were required to detect an HR of 0.55 with 90% power. A stratified log-rank test was used with a 1-sided alpha of 2.5%. The median PFS in the control arm was approximated at 3 months. With a target HR of 0.55, representing a 45% reduction in the risk of disease progression or death, the median PFS was expected to be approximately 5.5 months in the investigational arm, factoring in exponential distributions.

What were the baseline characteristics of the trial population?

In the combination arm, the mean age was 61.1 years (range, 22-85), and 48.1% of patients were female (n = 101). Patients were White (n = 177; 84.3%), non-White (n = 4; 1.9%), or did not have reported data (n = 29; 13.8%). ECOG performance status was 0 in 162 patients (77.1%) and 1 in 48 (22.9%). Baseline lactate dehydrogenase (LDH) levels were below the upper limit of normal (ULN) in 120 patients (57.1%), between the ULN and less than twice the ULN in 65 (31.0%), and at least twice the ULN in 25 (11.9%).

Within the investigational arm, the median time since first metastatic diagnosis was 0.18 years (range, 0.0-8.1). The size of the largest metastatic lesion was no more than 3 cm in 119 patients (56.7%), between 3.1 and 8.0 cm in 73 patients (34.8%), and at least 8.1 cm in 18 patients (8.6%). Metastatic sites were hepatic (n = 127; 60.5%), extrahepatic (n = 11; 5.2%), or both (n = 72; 34.3%).

In the control arm, the mean age was 61.5 years (range, 29-85), and most patients were female (n = 52; 50.5%). Eighty-three patients were White (80.6%), 4 were non-White (3.9%), and 16 had no reported data (15.5%). ECOG performance status was 0 (n = 87; 84.5%) or 1 (n = 16; 15.5%), and baseline LDH levels were below the ULN in 62 patients (60.2%), between the ULN and less than twice the ULN in 29 (28.2%), and at least twice the ULN in 12 (11.7%). Additionally, 79 patients (76.7%) received ipilimumab plus nivolumab and 24 (23.3%) received pembrolizumab; no patients received dacarbazine.

In the control arm, the median time since first metastatic diagnosis was 0.19 years (range, 0.0-4.6). The size of the largest metastatic lesion was no more than 3 cm in 67 patients (65.0%), between 3.1 and 8.0 cm in 31 (30.1%), and at least 8.1 cm in 5 (4.9%). Metastatic sites were hepatic (n = 65; 63.1%), extrahepatic (n = 7; 6.8%), or both (n = 31; 30.1%).

How did the combination perform in terms of responses?

With respect to secondary end points, the ORR according to BICR was 37.1% (n = 78; 95% CI, 30.6%-44.1%) in the combination arm vs 5.8% (n = 6; 95% CI, 2.2%-12.3%) in the control arm (OR, 10.8; 95% CI, 4.4-26.4; P < .0001). The median DOR was 6.8 months (95% CI, 5.5-11.3) vs not evaluable (NE) in the investigational and control arms, respectively. The 12-week DCRs were 73.3% (n = 154; 95% CI, 66.8%-79.2%) and 31.1% (n = 32; 95% CI, 22.3%-40.9%) in the respective arms (OR, 7.7; 95% CI, 4.3-13.5; P < .0001).

Best overall responses per BICR in the combination and control arms consisted of complete responses (CRs; 2.4% vs 0%), partial responses (PRs; 34.8% vs 5.8%), stable disease (SD; 51.0% vs 41.7%), and progressive disease (PD; 6.2% vs 28.2%), respectively. One patient was NE by RECIST in the control arm, and 12 (5.7%) and 24 (23.3%) patients in the combination and control arms, respectively, had no postbaseline assessments available and were therefore not evaluable.

When investigator assessment was applied, the ORR was 39.5% (n = 83; 95% CI, 32.9%-46.5%) in the combination arm vs 1.9% (n = 2; 95% CI, 0.2%-6.8%) in the control arm (OR, 34.6; 95% CI, 8.1-146.9 ; P < .0001). The median DOR was 6.8 months (95% CI, 4.8-9.7) vs NE in the investigational and control arms, respectively. The 12-week DCRs were 74.3% (n = 156; 95% CI, 67.8%-80.1%) and 27.2% (n = 28; 95% CI, 18.9%-36.8%) in the respective arms (OR, 9.5; 95% CI, 5.3-17.0; P < .0001).

Best overall responses per investigator assessment in the combination and control arms consisted of PRs (39.5% and 1.9%), SD (47.1% and 37.9%), and PD (7.1% and 33.0%), respectively; 1 (0.5%) and 4 (3.9%) patients were NE by RECIST in the combination and control arms, respectively, and 12 (5.7%) and 24 (23.3%) patients had no postbaseline assessments available.

Orloff noted that OS data, although not mature, demonstrated an early trend toward improvement with the combination.

What was the safety profile of the regimen?

The median relative dose intensity was 91.0% for darovasertib, 77.1% for crizotinib, and 100% for investigator’s choice of therapy. Treatment-related serious adverse effects (AEs) were less frequent with darovasertib plus crizotinib at 9.2% (n = 22) compared with 25.0% (n = 25) for investigator's choice therapy; events leading to treatment withdrawal occurred in 0.8% (n = 2) and 5.0% (n = 5) of patients, respectively.

Treatment-related AEs (TRAEs) leading to darovasertib or crizotinib withdrawal occurred in 2.5% (n = 6) and 10.0% (n = 24) of patients, respectively. Nineteen (19.0%) patients discontinued investigator’s choice therapy because of a TRAE. TRAEs leading to dose reduction of either darovasertib or crizotinib occurred in 23.4% (n = 56) and 26.4% (n = 63) of patients, respectively. No dose reductions owing to TRAEs were required in the investigator’s choice arm.

Any-grade TRAEs that occurred in at least 20% of patients in the combination arm (n = 235; 98.3%) were diarrhea (n = 203; 84.9%), nausea (n = 179; 74.9%), peripheral edema (n = 160; 66.9%), vomiting (n = 119; 49.8%), dermatitis acneiform (n = 100; 41.8%), fatigue (n = 94; 39.3%), hypotension (n = 79; 33.1%), hypoalbuminemia (n = 61; 25.5%), dysgeusia (n = 61; 25.5%), dizziness (n = 61; 25.5%), and decreased appetite (n = 57; 23.8%).

Any-grade TRAEs reported in the control arm (n = 88; 88.0%) were diarrhea (n = 29; 29.0%), increased alanine aminotransferase (ALT) levels (n = 25; 25.0%), increased aspartate aminotransferase (AST) levels (n = 25; 25.0%), fatigue (n = 25; 25.0%), nausea (n = 25; 25.0%), and pruritus (n = 24; 24.0%).

“The AEs were consistent with [the regimen’s] prior safety profile, with a low rate of treatment-related serious AEs and discontinuations due to TRAEs for darovasertib and crizotinib,” Orloff reported.

In the combination arm the most common grade 3/4 TRAEs that occurred in at least 2% of patients (n = 97; 40.6%) were diarrhea (n = 24; 10.0%), syncope (n = 17; 7.1%), hypotension (n = 9; 3.8%), nausea (n = 7; 2.9%), peripheral edema (n = 7; 2.9%), and anemia (n = 6; 2.5%).

In the control arm the most common grade 3/4 TRAEs that occurred in at least 2% of patients (n = 37; 37.0%) were increased ALT levels (n = 7; 7.0%), increased AST levels (n = 7; 7.0%), diarrhea (n = 6; 6.0%), hepatitis (n = 5; 5.0%), colitis (n = 4; 4.0%), increased lipase (n = 3; 3.0%), autoimmune hepatitis (n = 3; 3.0%), increased amylase (n = 3; 3.0%), tubulointerstitial nephritis (n = 2; 2.0%), hypophysitis (n = 2; 2.0%), and hyperthyroidism (n = 2; 2.0%).

Grade 5 TRAEs occurred in 0.4% (n = 1) and 1.0% (n = 1) of patients in the combination and control arms, respectively.

Disclosures: Orloff disclosed employment from Pfizer (I); honoraria from Immunocore; consulting or advisory roles with Delcath Systems, IDEAYA Biosciences, Immunocore, Replimune, and TriSalus Life Sciences; speakers bureau participation with Immunocore; and research funding from Bristol Myers Squibb (Inst), Delcath Systems (Inst), IDEAYA Biosciences (Inst), Immunocore (Inst), iOnctura (Inst), Iovance Biotherapeutics (Inst), Linnaeus Therapeutics (Inst), Novelwise (Inst), Regeneron (Inst), Replimune (Inst), and TriSalus Life Sciences (Inst).

References

1. Orloff M, Ramelyte E, Butler MO, et al. Darovasertib plus crizotinib vs investigator’s choice as first-line treatment for patients with HLA-A2 negative metastatic uveal melanoma: primary results from the OptimUM-02 trial. J Clin Oncol. 2026;44(suppl 17):LBA9503. 10.1200/JCO.2026.44.17_suppl.LBA9503
2. Khoja L, Atenafu E, Leyvraz S, et al. Updated analyses from a global meta-analysis in metastatic uveal melanoma (mUM) to determine progression free and overall survival benchmarks by line of therapy: an international rare cancers initiative (IRCI) ocular melanoma study. J Clin Oncol. 2025;43(suppl 16):9539. doi:10.1200/JCO.2025.43.16_suppl.9539