Dr Gandara on Expanding Beyond PD-L1 as a Biomarker in NSCLC

We don't measure PD-L1 the same way in non–small cell lung cancer that we do in almost every other tumor type. In every other tumor type, we use what's called a CPS composite score. It takes the PD-L1 level on the tumor cell and the PD-L1 levels in the immune infiltrating cells. In non–small cell lung cancer, we ignore the IC.

David R. Gandara, MD, professor emeritus, Division of Hematology and Oncology, and co-director, Center for Experimental Therapeutics, UC Davis Comprehensive Cancer Center, discussed the limitations of PD-L1 testing and the role of alternative biomarkers in non–small cell lung cancer (NSCLC).

At the IASLC 2026 World Conference on Lung Cancer (WCLC) in Seoul, Gandara and colleagues presented a genomic analysis of the phase 3 EMPOWER-Lung 1 trial (NCT03088540) of first-line cemiplimab (Libtayo) versus chemotherapy in PD-L1–high advanced NSCLC.1 Tumor mutational burden (TMB) of ≥10 mutations/megabase identified patients with an overall survival benefit from cemiplimab (n = 84; median, 53.8 vs 5.6 months; HR, 0.30; 95% CI, 0.17 - 0.52), but Gandara said that effect was dwarfed by the impact of KRAS G12C (HR, 0.18; 95% CI, 0.07 - 0.49). Co-occurrence of KRAS G12C with STK11 (2%) or KEAP1 (<1%) was rare.

Asked about unmet needs in PD-L1–low, KRAS G12C–mutant disease, Gandara said the first step is defining terms. Some clinicians consider any tumor proportion score of 1 - 49% "low," but the formal category is PD-L1 less <1%, meaning the pathologist did not see a single stained cell.

That threshold, he said, is complicated by a historical mistake. Nearly every other tumor type uses a combined positive score that counts PD-L1 on both tumor cells and immune infiltrating cells; NSCLC scores tumor cells only. After 15 years, Gandara said, the field is unlikely to go back and fix it, but it explains in part why a PD-L1–negative result does not rule out immunotherapy benefit.

The practical implication, Gandara said, is a need for biomarkers beyond PD-L1. TMB is one, as the EMPOWER-Lung 1 analysis showed. Another is plasma proteomics: PROphetNSCLC, now commercially available in the United States, measures ?7000 proteins reflecting tumor-derived, host, and tumor microenvironment biology.

OncoHost presented data at WCLC 2026 on the assay's performance in samples collected after the first treatment cycle.

Gandara said published work shows the proteomic score complements PD-L1, and a positive PROphet result in a patient with PD-L1 expression supports benefit from immunotherapy plus chemotherapy.

Together, he said, these tools will change how trials are interpreted and designed.