2 Clarke Drive
Suite 100
Cranbury, NJ 08512
© 2024 MJH Life Sciences™ and OncLive - Clinical Oncology News, Cancer Expert Insights. All rights reserved.
In Partnership With:
Conference | SITC Meeting
The biomarker analysis of patients enrolled on part B of the CheckMate 914 trial explored the effects of PD-L1 and KIM-1 expression.
PD-L1 expression levels were found to be predictive of improved outcomes with nivolumab (Opdivo) and the CTLA-4 inhibitor ipilimumab (Yervoy) in patients with clear cell renal cell carcinoma (ccRCC), whereas high KIM-1 expression led to worse outcomes in this patient population with nivolumab alone or with ipilimumab, according to a biomarker analysis of part B of the phase 3 CheckMate 914 trial (NCT03138512) presented at the 2024 SITC Annual Meeting.1
Data showed that high KIM-1 levels were associated with worse disease-free survival (DFS) with nivolumab monotherapy vs placebo in patients with ccRCC, as well as a trend toward poor DFS with nivolumab/ipilimumab, suggesting that micro-metastatic disease releases KIM-1 into the circulation, as well as that serum KIM-1 is a potential minimally invasive biomarker for resected RCC recurrence. However, KIM-1 levels were not found to be predictive of response to nivolumab alone or with ipilimumab vs placebo.
When DFS was evaluated across study arms stratified by patients with PD-L1 expression lower than 1% (n = 663) and 1% or higher (n = 87), results showed that PD-L1 expression was determined to be predictive of response to nivolumab/ipilimumab. Higher baseline PD-L1 levels were significantly associated with better DFS, showing consistency with results from the phase 3 KEYNOTE-564 trial (NCT03142334) of adjuvant pembrolizumab (Keytruda) in patients with RCC.2
However, patients who had higher fibroblast levels within the tumor microenvironment had shorter DFS with nivolumab plus ipilimumab compared with nivolumab alone.
“In the adjuvant setting, tumor PD-L1 expression is potentially predictive of favorable nivolumab plus ipilimumab outcome, while high circulating KIM-1 levels are prognostic of worse clinical outcomes,” co-first study authors Sai Vikram Vemula, precision medicine/biomarker lead of immuno oncology at Bristol Myers Squibb, and Wenxin (Vincent) Xu, MD, an assistant professor of medicine at Harvard Medical School/Dana-Farber Cancer Institute in Boston, Massachusetts, wrote alongside coinvestigators in a poster presented during the meeting. “Patients with PD-L1 expression may be enriched with immune cell populations, contributing to better immunotherapy response.”
An unmet need exists for predictive biomarkers for immune checkpoint inhibitors in the treatment of patients with ccRCC. Prior data from part A of the CheckMate 914 trial showed that PD-L1 expression and serum KIM-1 levels are independently associated with treatment effect from nivolumab combined with ipilimumab vs placebo in patients with localized ccRCC.3
In the analysis presented at the 2024 SITC Annual Meeting, investigators further evaluated the role of PD-L1 and KIM-1 for association with nivolumab alone and with ipilimumab in part B of the CheckMate 914 trial; they also explored a novel human-interpretable image feature–based platform to characterize the cell and tissue composition from hematoxylin and eosin whole-slide images.1 This was to assess the platform’s potential to identify novel prognostic and predictive biomarkers for immunotherapy response in RCC.
Part B of CheckMate 914 comprised 825 patients with RCC with predominant clear cell histology; patients had radical or partial nephrectomy with negative surgical margins, an ECOG performance status of 0 or 1, and pathologic TNM staging of pT2a, G3 or G4, N0 M0; pT2b, G any, N0 M0; pT3, G any, N0 M0; pT4, G any, N0 M0; or pT any, G any, N1 M0. Patients could not have any clinical/radiological evidence of residual disease or distant metastases after nephrectomy as confirmed by blinded independent central review (BICR).
Stratification factors included pathologic Tumor, Node, Metastasis staging and type of nephrectomy. Patients were randomly assigned 2:1:1 to receive either nivolumab at 240 mg intravenously (IV) every 2 weeks for 12 doses plus placebo every 6 weeks for 4 doses (n = 411), placebo IV every 2 weeks for 12 doses plus placebo IV every 6 weeks for 4 doses (n = 208), or nivolumab plus ipilimumab at 1 mg/kg IV every 6 weeks for 4 doses (n = 206). Randomization occurred between 4 and 12 weeks after surgery.
The primary end point was DFS per BICR for nivolumab vs placebo, and secondary end points were DFS per BICR for nivolumab vs nivolumab/ipilimumab, overall survival (OS) for nivolumab vs placebo, OS for nivolumab vs nivolumab/ipilimumab, and safety.
Pretreatment tumor Nx samples were used to assess PD-L1 tumor cell expression levels via PD-L1 IHC 28-8 pharm Dx assay, and KIM-1 levels were assessed using Meso Scale Discovery immunoassay with serum samples (n = 704) collected prior to adjuvant dosing on cycle 1 day 1 and cycle 4 day 1. KIM-1 quartiles were defined across all patients, with the highest quartile corresponding to the highest levels. Human-interpretable image feature (n = 1043) was used to predict clinically relevant cell types and molecular phenotypes from histopathology hematoxylin and eosin whole-slide images. Investigators used a Kaplan-Meier and Cox proportional hazards analysis to determine the link between biomarkers and DFS.
The median follow-up was 27.0 months (range, 18.0-42.4).
Additional results showed that serum KIM-1 levels were linked with age and nephrectomy status. Higher levels of KIM-1 were more commonly reported in those who were older than 65 years of age and had undergone partial nephrectomy.
Investigators also explored the association of on-treatment changes in KIM-1 with response to nivolumab alone and with ipilimumab. Results showed that on-treatment increase in circulating KIM-1 was linked with shorter DFS, suggesting that KIM-1 could be an early predictor for response to single-agent nivolumab and nivolumab/ipilimumab in this setting.
Furthermore, patients with PD-L1 expression levels of 1% or higher had higher levels of macrophages, lymphocytes, and neutrophils, indicating a potentially inflamed tumor micronenviroment; endothelial cell levels were also lower in this subgroup.
The authors concluded that these observations should be explored in further clinical trials.