A Study of Biomarkers for Cancer Immune Checkpoint Inhibitor Therapy

Abstract

Immune Checkpoint Inhibitors (ICI) have revolutionised the treatment of solid tumours and brought in a new era of oncology. However, only 40% of patients derive a significant benefit and an equal number develop autoimmune related adverse events (irAE). Current biomarkers such as Programmed Death Ligand 1 positivity (PD-L1) and Tumour Mutational Burden (TMB) lack predictive power in most cases. The APC125 study aimed to identify novel biomarkers of immune checkpoint inhibitor response in blood, faecal microbiome, and tumour tissue in patients with unresectable solid tumours treated with PD-1/PD-L1 inhibitors. This was a single institution observational study. Patients were also treated with standard of care combination cytotoxics where appropriate. Patients aged over 18 years with unresectable or metastatic solid tumours measurable by iRECIST criteria and scheduled to start first line anti PD-1/PD-L1 ICI therapy with or without cytotoxic chemotherapy or tyrosine kinase inhibitors (TKIs) were invited to enrol in an observational study. Baseline white cell count (WCC) and differential, platelet count, LDH, ESR and C-Reactive Protein (CRP) were recorded as were serial WCC and differential at the second treatment visit (3 to 6 weeks later) and at the 3-month treatment and restaging Computed Tomography (CT) scan visit. The Systemic Inflammatory Index (SII) was calculated from the Full Blood Count (FBC). (SII = Neutrophils x Platelets/Lymphocytes). Serum samples were collected at all 3 time points for proteomic analysis and Kynurenine/Tryptophan ratio. PAXgene® peripheral blood mRNA tubes were collected at baseline and at the second visit only. Faecal samples were collected at baseline and second visit for microbiome analysis. Archived FFPE samples were tested for transcriptome mRNA expression. Pre-specified clinical endpoints included disease response rates as assessed by CT scans reported to iRECIST, progression free survival (PFS), overall survival (OS), and Grade 2-5 irAE. The FACT-ICM questionnaire was used to measure changes in Quality of Life (QoL) between baseline and restaging CT visit (Visit 3). Some patients used digital tablets to complete baseline and Visit 3 Cambridge Neuropsychological Test Automated Battery (CANTAB) testing.

Between August 2021 and December 2022, 20 patients were enrolled. The breakdown was as follows: 12 Non-Small Cell Lung Cancer (NSCLC) (7 adenocarcinoma, 5 squamous cell), 3 Extensive Stage Small Cell Lung Cancer (ES-SCLC), 2 oesophageal adenocarcinoma, 1 malignant melanoma, 1 endometrial carcinoma and 1 with triple negative breast cancer (TNBC). The median PFS and OS were 20.9 weeks and 34 weeks for the entire population and 20.1 weeks and 35.9 weeks for the NSCLC cohort. Four patients experienced Grade 3-4 irAE. Two patients with ongoing response had ICI held for multiple Grade 2 toxicities. Higher baseline levels of CRP correlated with worse PFS (p = 0.01,) as did Systemic Inflammatory Index (p = 0.20) when analysed as single variables. Using multivariate analysis Neutrophil to Lymphocyte Ratio correlated with worse OS with a Hazard Ratio of 1.69 (95% C.I. 1.05 to 3.21). When outliers were removed using Grub’s test there was a significant rise (p= 0.03) in Kynurenine/Tryptophan Ratio (K/T) between the pooled average of baseline and second visit blood draws which may enable identification of patients who would benefit from modification of T cell metabolism. Principle component analysis (PCA) of baseline serum immune proteins using the 96 Immune-Oncology Olink Proximity Extension Assay identified candidate biomarkers of worse OS including elevated levels of circulating Platelet Derived Growth Factor Subunit Beta (PDGF BB) and Vascular Endothelial Growth Factor Receptor 2 (VEGF R2). Although not significant it is possible that patients with high levels of serum PDGF BB and VEGF R2 would benefit from simultaneous targeting of angiogenesis and immune checkpoints. mRNA analysis suggests that upregulation of Extracellular Membrane gene pathways and downregulation of immune activation correlated with better response as per first restaging CT scan imaging measured by iRECIST. Faecal microbiome alpha diversity increased on treatment and beta diversity decreased but did not correlate with clinical outcomes.