In:
Journal of Clinical Oncology, American Society of Clinical Oncology (ASCO), Vol. 37, No. 15_suppl ( 2019-05-20), p. e14054-e14054
Abstract:
e14054 Background: Immunotherapy has revolutionized the treatment of NSCLC. However, response rate is variable, with a substantial failure rate. Thus, the identification of predictive biomarkers of response to immunotherapy is an area of great interest. Methods: Patients with locally advanced or metastatic NSCLC treated with nivolumab or pembrolizumab were enrolled. Disease response was defined following RECIST criteria (v. 1.1). Four ml of plasma were collected for the analysis of exosomal mRNA levels of PD-L1 (e-PD-L1) and IFN-γ (e-IFN-γ) at baseline and at the time of first radiological assessment. Exosome isolation and mRNA extraction was obtained by the exoRNeasy kit (Qiagen, Valencia, CA). e-PD-L1 and e-IFN-γ were evaluated by the QX100 ddPCR (Bio-Rad, Hercules, CA) and expressed as allele frequency (%). Chest computed tomography (CT) scan at baseline was used for the radiomic analysis. Tumor segmentation was performed on DICOM-formed images taken from the picture archiving, and the regions of interest were delineated manually and analyzed using the QUIBIM SL software. Survival was calculated stratifying patients based on e-PD-L1 and e-IFN-γ median values. Results: Nivolumab was given to 17 patients as 2nd line and to 8 subjects as further line of treatment, while 13 patients received 1st line pembrolizumab. Median PFS was 11 vs 16.2 months (mos) in patients with baseline e-PD-L1 of 〈 0.3% vs ≥0.3%, respectively (p = 0.16). e-PD-L1 significantly increased in disease progression (PD) vs partial response (PR) and disease stabilization (SD) (p = 0.01) after 2 mos of treatment. In patients with e-IFN-γ ≥4.1% vs 〈 4.1% at baseline, median PFS was 5.6 mos vs not reached, respectively (p = 0.003). The multiparametric radiomic analysis identified the Cluster Prominence Value (CPV, p = 0.012) and the Cluster Shade Value (CSV, p = 0.034) as significantly correlated with treatment outcome. Moreover, the D2d parameter and e-IFN-γ were inversely correlated (p 〈 0,0001). CPV and D2d reflect the intra-tumor architecture, including necrosis, cell proliferation, and angiogenesis. Conclusions: Liquid biopsy data correlate with radiomic parameters and predict response to immunotherapy.
Type of Medium:
Online Resource
ISSN:
0732-183X
,
1527-7755
DOI:
10.1200/JCO.2019.37.15_suppl.e14054
Language:
English
Publisher:
American Society of Clinical Oncology (ASCO)
Publication Date:
2019
detail.hit.zdb_id:
2005181-5
Bookmarklink