Despite its success in advanced melanoma, Bristol Myers Squibb’s fixed-dose PD-1/LAG-3 combo, Opdualag, failed to move the needle in resected skin cancer.

After revealing in February that Opdualag failed as a postsurgery adjuvant treatment for patients with completely resected stage 3 and 4 melanoma, BMS has peeled back full results from the negative Relativity-098 trial at the 2025 American Society of Clinical Oncology annual meeting.

At a minimum follow-up of 23.4 months, Opdualag—a combo of Opdivo and the LAG-3 antibody relatlimab—showed no difference from Opdivo alone in terms of patients' risk of disease recurrence or death. About 39% of patients in both arms of the trial had experienced recurrence or passed away.

At the two-year mark, 62% of patients in the Opdualag arm remained alive without recurrence, versus 63.6% in the Opdivo group. The median recurrence-free survival time was not reached for the Opdualag arm, while it was 33.1 months for the control group.

Treatment with Opdualag was associated with a higher rate of grade 3 or 4 treatment-related adverse events, 19%, compared with 8% for Opdivo alone. Two treatment-related deaths happened in the 547-patient Opdualag arm, compared with one case in the 546-subject Opdivo arm.

The adjuvant melanoma flop again brings the utility of LAG-3 antibodies into question and draws attention to BMS’ somewhat controversial decision last year to push Opdivo and a higher dose of relatlimab into phase 3 testing in the all-important setting of first-line non-small cell lung cancer (NSCLC).

But to BMS Chief Medical Officer Samit Hirawat, M.D., findings from a study in adjuvant melanoma have little bearing on the combo's promise in metastatic NSCLC.

Relatlimab potentiates the immune system and the working of Opdivo in the environment when the tumor is present, attracting T cells to work for Opdivo, Hirawat explained.

Relatively-098, meanwhile, was conducted in resected melanoma, where patients had no tumors present.

In the metastatic setting, where the tumor is present, “we do see that patients who have a non-small cell lung cancer with expression of PD-1 greater than 1%, those are the patients that benefit,” Hirawat said.

The BMS exec was referring to data from the phase 2 Relativity-104 trial, which didn’t impress in a first-line NSCLC all-comers population. But when sharing the data at last year’s European Society of Medical Oncology Congress, BMS pointed to improvements in nonsquamous patients whose PD-L1 expression was between 1% to 49% to justify its decision to launch the phase 3 Relativity1093 trial in that population. The study was later edited to include all PD-L1-positive patients after the company examined the PD-L1-high data further. 

BMS’ decision was not very well received. Critics argued that the company was trying too hard to find a positive signal and, in the process, cutting the data too thin. The control arm in the phase 2 trial also appeared to have underperformed historical data, raising questions about the validity of the positive signals that BMS saw. 

Despite the latest adjuvant melanoma setback, “we remain pretty confident about conducting” the first-line NSCLC trial, Hirawat said.