Abstract 226P
Background
Yttrium-90 (Y90)-resin microspheres radioembolization followed by nivolumab (Nivo) has shown encouraging response rate in advanced hepatocellular carcinoma (HCC), but only a subset of patients benefit from it. Predictive biomarkers help to identify patients for the treatment.
Methods
Longitudinal plasma and peripheral blood mononuclear cells (PBMCs) collection from HCC patients who received sequential Y90-RE followed by nivolumab (CA209-678; NCT03033446) allowed assessment of dynamics changes. Predictor selection procedure includes individual timepoint comparison, average analysis, trajectory description, responsiveness prediction, and combined modalities validation. 65-plex Human ProcartaPlex Luminex panel was employed to examine the concentration of cytokines and chemokines in 187 plasma samples collected at eight timepoints (responder [R]=11, non-responder [NR]=22). Immunomonitoring of PBMCs (NR=12, R=5) was assessed with a 39-plex Cytek panel followed by dimension reduction analysis. CXCL9 was reported impacting the immune treatment response, thus we focused on CXCL9+ CD8 cluster. Mixed models were fitted for responsiveness represented by Luminex analytes and CD8 phenotypes. Logistic regression and Cox model were fitted for the association between analytes and treatment response and 3-year overall survival (OS). Biomarkers showed significant association (P<0.05) and AUC > 0.7 were selected as the potential predictors.
Results
All Luminex biomarkers showed difference between the R and NR groups 14 days after Y90 treatment (P<0.05). The higher increment of IL-18, IL-12p70, and CCL24 after Y90 treatment related to better treatment response and 3-year OS. CXCL9+ and CXCR3+ CD8 clusters differed between R and NR on the 21st and 35th day after Y90-RE, respectively. IL-18 and IL-12p70 were significantly associated with CXCL9+ and CXCR3+ CD8 clusters in the Rs, but not NRs.
Conclusions
IL-18, IL-12p70, and CCL24 were potential predictors for treatment response as well as 3-year OS for HCC patients treated by Y90-Nivo.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.
Resources from the same session
185P - Real-world data analysis of genomic profiling-matched targeted therapy outcomes in patients with fusion-positive NSCLC
Presenter: Jyoti Patel
Session: Poster session 01
186P - Pooled efficacy and safety data of alectinib (A) vs. crizotinib (C) from the randomized phase III ALEX and J-ALEX trials
Presenter: Marco Tagliamento
Session: Poster session 01
187P - Ipatasertib and atezolizumab in cancers with increased PI3K-AKT pathway activity: First results from the CRAFT trial
Presenter: Christoph Heilig
Session: Poster session 01
188P - The landscape of SMARCA2 genomic alterations in Chinese cancer patients
Presenter: Chen Jiaqi
Session: Poster session 01
189P - Design and enrollment for a classifier development study for a blood-based multi-cancer early detection (MCED) test
Presenter: Christopher Douville
Session: Poster session 01
190P - Quantitative serum tumor markers (CEA, CA19-9, and CA-125) are independently predictive of survival in patients with appendiceal adenocarcinoma
Presenter: John Paul Shen
Session: Poster session 01
191P - Novel approach to proficiency testing demonstrates wide gaps in biomarker quality for colon cancer treatment
Presenter: Kassandra Bisson, Brandon Sheffield
Session: Poster session 01
192P - Impact of oncogenic fibroblast growth factor receptor (FGFR) alterations in patients with advanced solid tumors in a real-world setting
Presenter: Hussein Sweiti
Session: Poster session 01
194P - Early kinetics of C-reactive protein for cancer-agnostic prediction of therapy response and mortality in patients treated with immune checkpoint inhibitors: A multi-center cohort study
Presenter: Dominik Barth
Session: Poster session 01
195P - Identification of biomarkers associated with checkpoint inhibitor pneumonitis based on serum proteomic approach and construction of an online interactive visual prediction model
Presenter: Xiaohui Jia
Session: Poster session 01