Inter and intra-tumor heterogeneity of PD-L1 and MET expression in metastatic renal cell carcinoma (mRCC)

Abstract only 4569 Background: Although inhibition of PD-1/PD-L1 and MET receptors have clinical efficacy in mRCC, their expression is not a predictive biomarker. Heterogeneity between the sites of disease might be one explanation. The aim of our study was to evaluate PD-L1 and MET expression in pri...

Full description

Saved in:
Bibliographic Details
Published inJournal of clinical oncology Vol. 35; no. 15_suppl; p. 4569
Main Authors Derosa, Lisa, Le Teuff, Gwenael, Khordahi, Manal, Chanez, Brice, Zalcman, Emanuelle, Gravis, Gwenaelle, Negrier, Sylvie, Radulescu, Camelia, Albiges, Laurence, Merabet, Zahira, Escudier, Bernard J.
Format Journal Article
LanguageEnglish
Published 20.05.2017
Online AccessGet full text

Cover

Loading…
More Information
Summary:Abstract only 4569 Background: Although inhibition of PD-1/PD-L1 and MET receptors have clinical efficacy in mRCC, their expression is not a predictive biomarker. Heterogeneity between the sites of disease might be one explanation. The aim of our study was to evaluate PD-L1 and MET expression in primary and metastases (brain (BM)/pancreas (PM)) RCC lesions and their correlation with clinicopathologic characteristics. Methods: RCC specimen from different institutions were collected. Clinicopathologic characteristics were assessed by revision of samples. PD-L1 and MET expression in tumor cells (TC) and immune cells (IC) ( > 1%) were assessed by immunohistochemistry. Results: 180resected RCC specimen were successful collected (42 primary tumors and 138 metastases (87 BM/51 PM)). Overall, 22%, 51% and 23% of patients had at least one specimen expressing PD-L1 TC, IC and MET, respectively. In primary tumours, the proportion was 12%, 50% and 0%, respectively. In metastasis, the proportion of PD-L1 TC was 22% (23% in BM vs 19% in PM, p = 0.631), PD-L1 IC was 48% (47% in BM vs 49% in PM, p = 0.821) and MET was 24% (35% in BM vs 2% in PM, p < 0.001). Comparing paired samples (primary tumour and metastasis) there was discordances of PD-L1 in TC or IC and of MET expression in 30%, 27% and 24% of samples, respectively. These two first disagreements seem varied over time. The discordance in PD-L1 TC or IC and MET between primary tumor and PM (BM) was 15% (40%), 33% (22%) and 0% (67%), respectively. Some correlations were observed between MET and PD-L1 and clinicopathologic characteristics. Conclusions: In this largest analysis, evaluating heterogeneity between primary tumor and metastases (brain/pancreatic lesions) in mRCC, PD-L1 and MET expression suggests that the assessment as predictive biomarkers may require analysis of metastatic lesions. [Table: see text]
ISSN:0732-183X
1527-7755
DOI:10.1200/JCO.2017.35.15_suppl.4569