PD-L1 expression is commonly used to predict response to immune checkpoint inhibitor therapy. However, PD-L1 density could be a better biomarker, primary results of the PIONeeR trial suggest.


PD1 and PD-L1 inhibitors deeply changed the management of advanced and metastatic non-small cell lung cancer (NSCLC). However, up to 30% of patients in the first-line setting and up to 50% of patients in the second-line setting are progressing in the first 6 months of treatment without any biomarker to predict this progression. Although there are some proposed mechanisms to predict primary resistance to immune checkpoint inhibitors, like poor immunogenicity, downregulation of chemokines, and induction of IDO, the distribution of these mechanisms in advanced/metastatic NSCLC is unknown; as is the predictive value of these mechanisms for the efficacy of new checkpoint inhibitors.


The PIONeeR project aims to predict response and/or resistance to PD1/PD-L1 immune checkpoint inhibitors in advanced NSCLC patients through comprehensive agnostic multiparametric and longitudinal biomarkers assessment. Prof. Fabrice Barlesi (Aix-Marseille University, France) presented the first results of the project based on 100 enrolled patients [1].

Eligible for the project are patients with advanced/metastatic NSCLC with available archived tumor tissue. Patients were treated with either nivolumab, pembrolizumab, or atezolizumab alone (second-line) or combined with chemotherapy (first-line) according to current standards. All patients were systematically re-biopsied and blood-sampled at 6 weeks of treatment. Responses were assessed by RECIST 1.1 every 6 weeks.


A 6-weeks biopsy was available in 46% of cases. At data cut-off, 123 different biomarkers were analyzed in the baseline biopsies. Of 100 patients, 33 progressed before the 6 weeks milestone, 35 progressed between 6 and 24 weeks,10 patients progressed between 24 and 52 weeks, and 21 patients had not progressed at data cut-off. Median progression-free survival (PFS) was 2.99 months and median overall survival (OS) was 11.01 months, which is in line with what can be expected in a real-world setting.


Baseline parameters positively associated with response were: a higher percentage PD-L1-positive cells (33% in responders vs 14% in non-responders), cytotoxic lymphocytes in the tumor (383 vs 298 cells/mm2), cytotoxic lymphocytes at the tumor-stroma interface (511 s 178 cells/mm2), and – surprisingly – a higher percentage regulatory T cells in the stroma (70 vs 18 cells/mm2). Response was also associated with a higher percentage neutrophils in the tumor at 6 weeks (73 vs 16 cells/mm2). Both PFS and OS were positively correlated to PD-L1 expression >1%. However, the correlation was even stronger for PD-L1-positive cell density in the tumor and/or stroma. PD-L1-positive cell density ≥22 cells/mm2 was correlated with improved PFS (P=0.014), whereas PD-L1 positive cell density ≥546 cells/mm2 was correlated with improved OS (P=0.015). In line with this, non-responders who had high PD-L1-expression (>15%) all had tumors with low PD-L1-positive cell density. In addition, non-response despite adequate PD-L1-positive cell density was found to be associated with through levels of the checkpoint inhibitor below the required threshold in 5 patients.


Barlesi F, et al. Precision immuno-oncology for advanced non-small cell lung cancer (NSCLC) patients (pts) treated with PD1/L1 immune checkpoint inhibitors (ICIs): A first analysis of the PIONeeR study. ESMO 2020 Virtual Meeting, abstract LBA53.