FOXP3-based immune risk model for recurrence prediction in small-cell lung cancer at stages I-III.

Background: Immunotherapies may prolong the survival of patients with small-cell lung cancer (SCLC) to some extent. The role of forkhead box protein P3 (FOXP3) in tumor microenvironment (TME) remains controversial. We aimed to examine FOXP3-related expression characteristics and prognostic values and to develop a clinically relevant predictive system for SCLC.

Methods: We enrolled 102 patients with histologically confirmed SCLC at stages I-III. Through immunohistochemistry, we determined the expression pattern of FOXP3 and its association with other immune biomarkers. By machine learning and statistical analysis, we constructed effective immune risk score models. Furthermore, we examined FOXP3-related enrichment pathways and TME traits in distinct cohorts.

Results: In SCLC, FOXP3 level was significantly associated with status of programmed death-ligand 1 (PD-L1), programmed cell death protein 1 (PD-1), CD4, CD8, and CD3 (p=0.002, p=0.001, p=0.002, p=0.030, and p<0.001). High FOXP3 expression showed longer relapse-free survival (RFS) than the low-level group (41.200 months, 95% CI 26.937 to 55.463, vs 14.000 months, 95% CI 8.133 to 19.867; p=0.008). For tumor-infiltrating lymphocytes (TILs), subgroup analysis demonstrated FOXP3 and PD-1, PD-L1, lymphocyte activation gene-3, CD3, CD4, or CD8 double positive were significantly correlated with longer RFS. We further performed importance evaluation for immune biomarkers, constructed an immune risk score incorporating the top three important biomarkers, FOXP3, TIL PD-L1, and CD8, and found their independently prognostic role to predict SCLC relapse. Better predictive performance was achieved in this immune risk model compared with single-indicator-based or two-indicator-based prediction systems (area under the curve 0.715 vs 0.312-0.711). Then, relapse prediction system integrating clinical staging and immune risk score was established, which performed well in different cohorts. High FOXP3-related genes were enriched in several immune-related pathways, and the close relationships of interleukin-2, , basic excision repair genes , , , and oxidative phosphorylation related gene cytochrome c oxidase subunit 8A with FOXP3 expression were revealed. Moreover, we found low-immune risk score group had statistically higher activated CD4 memory T cells (p=0.014) and plasma cells (p=0.049) than the high-risk group. The heterogeneity of tumor-infiltrating immune cells might represent a promising feature for risk prediction in SCLC.

Conclusion: FOXP3 interacts closely with immune biomarkers on tumor-infiltrating cells in TME. This study highlighted the crucial prognostic value and promising clinical applications of FOXP3 in SCLC.