肿瘤免疫微环境评分预测免疫检查点抑制剂为基础的方案在晚期非小细胞肺癌中的疗效

Immune checkpoint inhibitor (ICI)-based regimens have become the standard first-line treatment for advanced non-small cell lung cancer (NSCLC). However, response rates vary widely, emphasizing the need for a predictive model to optimize patient selection and improve treatment outcomes.

We retrospectively analyzed 96 treatment-naïve patients with advanced NSCLC who received first-line ICI-based regimens (anti-PD-1 ± platinum chemotherapy) at Hunan Cancer Hospital. Pre-treatment tumor biopsies underwent bulk RNA sequencing to identify differentially expressed genes associated with progression-free survival (PFS). A five-gene Cox proportional hazards model, the Prediction model of Immunotherapy Efficacy (PIE), was developed and validated in two independent cohorts: ORIENT-11 (n = 113; chemo-ICI) and OAK (n = 344; ICI monotherapy). Multiplex immunofluorescence staining (CD8, CD56, CD16, Pan-CK, PD-L1, MAGEA2) was performed to validate immune-cell infiltration patterns and candidate biomarkers across PIE-defined risk groups.

Patients in the long-survival group exhibited enrichment of IL7R + NK cells, SLC4A10 + CD8 + T cells, and dendritic cells, which were significantly associated with longer PFS. PIE integrated transcriptomic signatures of five biomarkers (CD274, KRT14, FOLR2, SLC31A2, and EFCAB14) to predicting PFS. Compared with patients having low-risk PIE scores, those classified as high-risk demonstrated significantly worse PFS (HR = 2.37, p < 0.001). PIE demonstrated predictive accuracy for 24-month PFS (AUC = 0.768). Additionally, MAGEA2 and MAGEA12 were identified as potential therapeutic targets.

PIE represents a clinically relevant, transcriptome-based predictive model for evaluating the benefit of ICI treatment in advanced NSCLC. With further prospective validation, PIE may facilitate personalized patient selection and guide the development of novel strategies to overcome primary resistance.