98P - Liquid biopsy in NSCLC: A promising tool to predict immunotherapy response

Background

Lung cancer represents the leading cause of cancer-related death worldwide, with non-small cell lung cancer (NSCLC) accounting for ∼85% of total cases. Although immunotherapy using immune checkpoint inhibitors (ICI) has shown promising results, only a subset of patients benefits from it and there is an urgent need for reliable predictive response biomarkers. Furthermore, traditional tissue biopsies are limited in size and provide a biased representation of the tumour. Therefore, we aimed to explore the possibility to obtain predictive biomarkers of response to ICI in a non-invasive way using liquid biopsy (LB).

Methods

To counter this, we analysed samples from 25 patients from advanced NSCLC treated with pembrolizumab at first line. Blood and tissue samples were collected at baseline to obtain cell-free DNA (cfDNA) and genomic DNA from plasma and PBMCs respectively, and tumour DNA from FFPE primary tumour. All samples underwent a target NGS panel to call somatic mutations in circulating tumour DNA (ctDNA) and primary tumour, evaluating their potential as predictive biomarkers.

Results

Through the integration of three different bioinformatic approaches for variant calling and ROC curve analysis, we identified mutations in key cancer-related genes associated with a 3-month radiological response. We also assessed a promising mutational signature in ctDNA comprising 10 genes with a high ability to predict response to ICI (AUC = 0.98). Although we explored copy number alterations and tumour mutational burden in LB, results did not yield statistical significance, suggesting that assessment of point mutations in particular genes and the proposed signature hold higher potential as biomarkers. Moreover, we compared ctDNA analysis with tissue biopsies, revealing a high concordance between ctDNA and primary tumour, enhancing the reliability of ctDNA analysis as a good alternative for tumour assessment.

Conclusions

These findings highlight the value of ctDNA analysis as a non-invasive predictive biomarker for ICI response and its correlation with mutations found in primary tumours. We also proposed a mutational signature to predict response at baseline for advanced NSCLC ICI treated patients, which could be readily implemented in the clinical management of the disease.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

Translational Molecular Oncology Unit. IIS Galicia Sur.

Funding

Vigo Contra el Cáncer; GAIN; Xunta de Galicia (ED481A-2020/214); Miguel Servet program (CP20/00188).

Disclosure

All authors have declared no conflicts of interest.