Real-World LAG3 Expression as A Biomarker in Advanced Melanoma: Implications for Treatment Selection.
BACKGROUND: Multiple first-line immune checkpoint regimens are approved for advanced melanoma, yet no biomarker reliably guides their selection. LAG-3, an inhibitory checkpoint enriched on exhausted immune cells, may reflect tumor-immune state beyond PD-L1 and guide therapeutic decisions. METHODS: We retrospectively analyzed Tempus platform data . Among 2,177 patients with advanced melanoma and RNA sequencing, 145 received first-line PD-1 monotherapy and 222 CTLA-4/PD-1 combination. LAG3 and CD274 (PD-L1) transcript levels were evaluated as dichotomous variables. Outcomes included objective response rate (ORR) at 90 days and real-world overall survival (rwOS). We assessed treatment-biomarker (LAG3/CD274) interactions and performed statistical de-correlation and fit treatment-specific multivariable Cox models including LAG3 expression, CD274 expression, and estimated CD8 T-cell infiltration . We also explored LAG3 association with immune infiltrate, tumor mutation - and neoantigen-burdens. RESULTS: Under monotherapy, high LAG-3 was associated with improved outcomes, ORR (67% vs 15%, p = 0.002); risk of death (HR 0.64; p < 0.001). In contrast, ORR and risk of death were similar for the combination cohort regardless of LAG-3 expression (HR 0.94, p = 0.6). After de-correlation, LAG-3 retained significant association with rwOS under monotherapy, whereas the association for CD274 was attenuated, indicating non-redundant biomarker information. In multivariable Cox analysis including continuous LAG3, CD274, and estimated CD8 T-cell infiltration, LAG3 remained independently associated with improved rwOS in the PD-1 monotherapy cohort (HR 0.54, 95% CI 0.32-0.91; p = 0.021), whereas neither CD274 nor CD8 infiltration remained significant. No biomarker significantly discriminated outcomes within the combination therapy cohort. CONCLUSIONS: These findings support further validation of LAG3 RNA expression as a candidate biomarker for frontline immunotherapy selection in advanced melanoma.
Authors
- Inderjit Mehmi (ORCID: https://orcid.org/0000-0002-2253-2206)
- Jacob Mercer (ORCID: https://orcid.org/0000-0001-8548-2800)
- Unnati Jariwala
- Adam Joseph Dugan (ORCID: https://orcid.org/0000-0001-5783-3781)
- Stamatina Fragkogianni (ORCID: https://orcid.org/0000-0002-5960-8921)
- Binyam Yilma (ORCID: https://orcid.org/0009-0007-8845-5390)
- Kathryn Blount Bollin (ORCID: https://orcid.org/0000-0001-7343-8925)
- Siwen Hu‐Lieskovan (ORCID: https://orcid.org/0000-0001-5836-4710)
- Omid Hamid
- Jonathan E Cohen
Institutions
- Hebrew University of Jerusalem (IL)
- Hadassah Medical Center (IL)
- Huntsman Cancer Institute (US)
- Angeles Clinic and Research Institute (US)
- Tempus Labs (United States) (US)
Publication Details
- Journal
- PubMed
- Published
- 2026-11-12
- DOI
- https://doi.org/10.1093/oncolo/oyag441
- Primary Topic
- Cancer Immunotherapy and Biomarkers
- Type
- article
- Field-Weighted Citation Impact
- 0.00