Abstract 1344P
Background
Immune checkpoint inhibitors have markedly improved survival in various cancers, including non-small cell lung cancer (NSCLC). However, effective prediction of individual responses and therapeutic outcomes is an unresolved clinical challenge.
Methods
This study enrolled 105 NSCLC patients (pts), stage IIIB-IV, had no driver mutations, and received 1st immuno-chemotherapy (IC). Blood was collected before therapy for analysis of plasma proteins and peripheral blood mononuclear cells (PBMC). The Olink proximity extension assay was used to analyze 92 immune-related proteins in plasma, and transcriptome sequencing was conducted on PBMC.
Results
Median PFS and OS were 18.2 and 25.3 months, respectively. Levels of eight proteins (CD27, TWEAK, CD8A, GZMH, IL6, PGF, TIE2, and TNFRSF12A, Table) in baseline plasma were significantly associated with OS (p < 0.05). Of note, higher levels of CD27 and TWEAK associated with favored OS; The other six proteins were related to dismal outcome. A model predicting therapeutic efficacy of IC was constructed. Both univariate (U) and multivariate (M) Cox regression analyses confirmed that this model independently predicted poorer PFS and OS (U HRs: 3.0 for PFS, 4.9 for OS; M HRs: 3.1 for PFS, 4.8 for OS, p < 0.05). Kaplan-Meier analysis confirmed significant differences in PFS and OS (p < 0.05). 10-fold cross-validation confirmed that the risk score predicts 12-month OS and 12-month PFS following IC, with AUCs of 0.80 and 0.72, respectively. In addition, in baseline PBMC, significantly higher enrichment of B cells, CD4+ naïve T cells, and M1 macrophages (p < 0.05) in the low-risk group, while the high-risk group exhibited significantly higher levels of M2 macrophages (p < 0.1).
Conclusions
A plasma protein-based prognostic model for immunotherapy was developed. Low-risk pts may exhibit stronger systematic immune responses, whereas high-risk pts may be immunosuppressed, as indicated by PBMC profiles. Table: 1344P