Data from: High specificity in circulating tumor cell identification is required for accurate evaluation of programmed death-ligand 1

Background: Expression of programmed-death ligand 1 (PD-L1) in non-small
cell lung cancer (NSCLC) is typically evaluated through invasive biopsies;
however, recent advances in the identification of circulating tumor cells
(CTCs) may be a less invasive method to assay tumor cells for these
purposes. These liquid biopsies rely on accurate identification of CTCs
from the diverse populations in the blood, where some tumor cells share
characteristics with normal blood cells. While many blood cells can be
excluded by their high expression of CD45, neutrophils and other immature
myeloid subsets have low to absent expression of CD45 and also express
PD-L1. Furthermore, cytokeratin is typically used to identify CTCs, but
neutrophils may stain non-specifically for intracellular antibodies,
including cytokeratin, thus preventing accurate evaluation of PD-L1
expression on tumor cells. This holds even greater significance when
evaluating PD-L1 in epithelial cell adhesion molecule (EpCAM) positive and
EpCAM negative CTCs (as in epithelial-mesenchymal transition (EMT)).
Methods: To evaluate the impact of CTC misidentification on PD-L1
evaluation, we utilized CD11b to identify myeloid cells. CTCs were
isolated from patients with metastatic NSCLC using EpCAM, MUC1 or Vimentin
capture antibodies and exclusion-based sample preparation (ESP)
technology. Results: Large populations of CD11b+CD45lo cells were
identified in buffy coats and stained non-specifically for intracellular
antibodies including cytokeratin. The amount of CD11b+ cells misidentified
as CTCs varied among patients; accounting for 33–100% of traditionally
identified CTCs. Cells captured with vimentin had a higher frequency of
CD11b+ cells at 41%, compared to 20% and 18% with MUC1 or EpCAM,
respectively. Cells misidentified as CTCs ultimately skewed PD-L1
expression to varying degrees across patient samples. Conclusions:
Interfering myeloid populations can be differentiated from true CTCs with
additional staining criteria, thus improving the specificity of CTC
identification and the accuracy of biomarker evaluation.