Impact of Patient Characteristics, Prior Therapy, and Sample Type on Tumor Cell PD-L1 Expression in Patients with Advanced NCSLC Screened for the ATLANTIC Study

Abstract Introduction We evaluated the impact of patient characteristics, sample types, and prior non-immunotherapy treatment on tumor cell (TC) programmed cell death ligand-1 (PD-L1) expression using samples from patients with advanced NSCLC. Methods 1,590 patients screened for the ATLANTIC study submitted a recently acquired (≤3 months) or archival (>3 months to >3 years old) tumor sample for PD-L1 assessment (VENTANA PD-L1 (SP263) Assay; TC ≥25% cut-off). Samples were acquired either prior to or after the ≥2 treatment regimens required for study entry and sample age varied among patients. 123 patients provided recent and archival samples. Results 517 (32.5%) of 1,590 patients had TC ≥25%: prevalence was greater in smokers versus non-smokers ( p = 0.0005) and those with EGFR− versus EGFR + tumors ( p = 0.0002); these effects were independent. Prevalence of TC ≥25% was increased in recent metastatic versus primary ( p = 0.005) and recent versus archival ( p = 0.039) samples. Chemotherapy or radiotherapy, but not tyrosine kinase inhibition, before sampling were associated with significantly increased TC ≥25% prevalence. PD-L1 status (TC ≥25% cut-off) remained unchanged in 74.0% of patients with recent and archival samples; where PD-L1 status changed, it was more likely to increase than decrease over time or with intervening treatment. Conclusions Several factors potentially impact PD-L1 TC ≥25% prevalence in advanced NSCLC, although no characteristic can be considered a surrogate for PD-L1 expression. Fresh biopsy may provide more accurate assessment of current tumoral PD-L1 expression where a low/negative result is seen in an archival sample, especially if the patient has received intervening therapy.