Monitoring PD-L1 Expression on Circulating Tumor-Associated Cells in Recurrent Metastatic Non-Small-Cell Lung Carcinoma Predicts Response to Immunotherapy With Radiation Therapy

Abstract

Purpose: Current diagnostic methods to determine programmed death 1 (PD-1) receptor and its ligand (PD-L1)/PD-1 immunotherapy (immune checkpoint inhibitor [ICI]) efficacy in recurrent or metastatic non-small-cell lung carcinoma (rmNSCLC) are imprecise. Although previously shown that patients with high tumor PD-L1 (≥ 50%) demonstrate clinical benefit in the form of disease reduction and improved survival, patients with low PD-L1 (< 50%) sometimes benefit from treatment. Since the PD-L1/PD-1 pathway is dynamic, monitoring PD-L1 levels during treatment may be more accurate than a static baseline tumor biopsy; however, rebiopsying the primary or metastatic disease is rarely feasible. Liquid biopsies that measure the upregulation of PD-L1 on tumor-associated cells (TACs), ie, cancer-associated macrophage-like cells and circulating tumor cells, have been performed, but their predictive value for ICI therapy efficacy is unknown.

Materials and methods: We initiated a single-blind prospective study to evaluate TAC PD-L1 expression changes in rmNSCLC from blood samples before (T0) and after (T1) treatment with ICI (ICI, n = 41) or without ICI (no ICI, n = 41). Anonymized blood was filtered to isolate TACs, which were then quantified for high/low PD-L1 expression. Progression-free survival (PFS) or overall survival (OS) hazard ratios (HRs) were evaluated at 18 and 24 months by censored univariate analysis.

Results: Increased TAC PD-L1 expression between T0 and T1 in patients who were not treated with ICI had no relationship with PFS or OS. However, increased TAC PD-L1 expression between T0 and T1 in patients treated with ICI had significantly better PFS (HR, 3.49; 95% CI, 1.5 to 8.3; P = .0091) and OS (HR, 3.058; 95% CI, 1.2 to 7.9; P = .0410).

Conclusion: Blood-based monitoring of dynamic changes in PD-L1 in TACs appears to identify patients with rmNSCLC who may benefit from ICI.

FIG 1.

Comparing the survival of patients on the basis of high/low PD-L1 expression after treatment (T1). (A) PFS of ICI patients with high PD-L1 expression (solid blue) v ICI patients with low PD-L1 expression (solid red). Median PFS 4.8 v 17 months. PFS of no ICI patients with high PD-L1 expression (dotted blue) v no ICI patients with low PD-L1 expression (dotted red). Median PFS 8.1 v 9.6 months. (B) OS of ICI patients with high PD-L1 expression (solid blue) v ICI patients with low PD-L1 expression (solid red). Median OS 5.9 v 21 months. OS of no ICI patients with high PD-L1 expression (dotted blue) v no ICI patients with low PD-L1 expression (dotted red). Median OS 9.9 v 16 months. ICI, immune checkpoint inhibitor; OS, overall survival; PD-L1, programmed death 1 receptor and its ligand; PFS, progression-free survival.

FIG 2.

Comparing the survival of patients based on change in PD-L1 expression between T0 and T1. (A) PFS of ICI patients with increased PD-L1 expression (solid green) v ICI patients with no increase in PD-L1 expression (solid orange). Median PFS 4.8 v 20 months. PFS of no ICI patients with increased PD-L1 expression (dotted green) v no ICI patients with no increase in PD-L1 expression (dotted orange). Median PFS 8.1 v 10 months. (B) OS of ICI patients with increased PD-L1 expression (solid green) v ICI patients with no increase in PD-L1 expression (solid orange). Median OS 5.9 v 22 months. OS of no ICI patients with increased PD-L1 expression (dotted green) v no ICI patients with no increase in PD-L1 expression (dotted orange). Median OS 9.9 v 16 months. ICI, immune checkpoint inhibitor; OS, overall survival; PD-L1, programmed death 1 receptor and its ligand; PFS, progression-free survival.