Use of LIF and LIFR expression to characterize survival and tumor microenvironment composition in lung adenocarcinoma.

8553 Background: In lung adenocarcinoma (LUAD), EGFR-directed therapy has transformed clinical practice, yet heterogeneity in treatment response persists. Novel biomarkers are needed to refine therapeutic selection and identify new treatment targets. Leukemia inhibitory factor (LIF) and its receptor (LIFR) have emerged as potential mediators of the tumor microenvironment (TME), but their clinical significance in LUAD remains poorly defined. This study sought to investigate LIF and LIFR expression in LUAD, especially EGFR -mutant (EGFRm), and their TME composition and impact on survival outcomes. Methods: 10,041 LUAD tumors underwent next-generation sequencing of DNA (592-gene panel or whole exome sequencing) and RNA (whole transcriptome) at Caris Life Sciences. Tumors were stratified by LIR and LIFR RNA expression quartiles (Q1, low; Q4, high). QuantiSEQ was used to profile TME. Statistical analyses were performed using chi-square and Mann–Whitney U tests. Overall survival (OS) was estimated from insurance claims data using Cox proportional hazards models to calculate hazard ratios and log-rank tests to determine p-values. Results: In LUAD, LIF Q4 tumors demonstrated lower abundance of CD8⁺ T cells (fold change (FC) 0.75, p < 0.001), but higher abundance of NK cells, M1 macrophages, neutrophils, Tregs and B cells vs LIF Q1 tumors (FC 1.1-1.4, p < 0.001). LIFR Q4 tumors exhibited higher abundance of dendritic cells, NK cells and M2 macrophages vs Q1 tumors (FC: 1.2-1.5, p < 0.001). For LIF, Q4 tumors had higher median MAPK activation (1.73 vs −0.93, p < 0.001) and T-cell–inflamed scores (94 vs −2, p < 0.001). LIFR Q4 tumors demonstrated higher MAPK activation (1.51 vs −0.83, p < 0.001), and higher T-cell–inflamed scores (105 vs −29, p < 0.001). IFN-γ signaling was modestly lower in LIF Q4 and LIFR Q4 tumors. Longer OS was observed in LIF Q1 vs Q4 (HR 0.79, 95% CI: 0.75-0.84, 25.1 vs 18.1 m, p < 0.001) but shorter OS in LIFR Q1 vs Q4 (HR 1.59 , 95% CI: 1.5-1.7, 14.0 vs 27.9 m, p < 0.001). EGFR mutations were significantly more frequent in LIF Q1 vs Q4 (21.2% vs 12.8%, p < 0.001) and LIFR Q4 vs Q1 (27.2% vs 7.6%, p < 0.001). In EGFR m tumors treated with osimertinib, OS differences based on LIF expression were accentuated:(LIF Q1 vs Q4, HR 0.681, 95% CI: 0.58-0.80, 35.7 vs 24.7 m, p < 0.00001). Conclusions: Findings demonstrate distinct transcriptional immune activation profiles between LIF- and LIFR-driven states, with superior survival in low LIF and high LIFR expressors. Despite elevated T-cell–inflamed scores, differences in immune cell composition suggest qualitative differences in immune activation between LIF- and LIFR-driven states. These findings support their relevance as prognostic and biologically informative biomarkers, and identify the LIF–LIFR axis as a key stratifier of immune state and survival heterogeneity in LUAD, with additional survival differences observed in EGFRm LUAD treated with osimertinib.