Background: Cancers of the biliary tract are rising in incidence but have extremely limited treatment options and grim outcomes. The programmed cell death protein 1 receptor (PD-1) inhibitor, pembrolizumab, produced robust responses in a small subset of advanced biliary cancer patients in a phase I/II clinical trial. Based upon the increased activity of checkpoint inhibition in combination with granulocyte-macrophage colony-stimulating factor (GM-CSF) in other cancers, an investigator-initiated, phase II trial (the ?KEY-G? trial) was developed to study the novel combination of pembrolizumab plus GM-CSF in advanced biliary cancers. In other cancers, tumor PD-1 ligand-1 (PD-L1) expression, DNA mismatch repair defects, and overall ?mutational load? are associated with response to checkpoint inhibitor monotherapy. This project will use tumor samples and data from patients enrolled on KEY-G to test the hypotheses that: 1) pre-treatment PD-L1 expression is not required for response to the combination of pembrolizumab plus GM-CSF, based upon a putative mechanism of GM- CSF to induce tumor PD-L1 expression; and 2) tumor genetics such as mismatch repair defects and/or mutational load are associated with clinical outcomes, based upon predictive value in other tumors.
Specific Aims : 1) To determine the association of tumor PD-L1 expression to clinical outcomes on pembrolizumab plus GM-CSF in advanced biliary cancer patients; and 2) to determine the association of common somatic tumor mutations, mutations in DNA damage response genes, and overall mutational load to clinical outcomes on pembrolizumab plus GM-CSF in advanced biliary cancer patients. Design and Methods: This project will analyze archival tumor samples and de-identified clinical data from advanced biliary cancer patients treated on the KEY-G trial (N=24).
For Aim 1, tumor samples will be tested for PD-L1 expression (% cells) by immunohistochemistry using a validated assay. Association between tumor PD- L1 status (?1%) and progression-free survival and overall survival will be tested using Kaplan-Meier methods, and for RECIST 1.1 response, using Farrington Manning likelihood score. PD-L1 expression will also be explored as a continuous variable.
For Aim 2, the association between common somatic mutations, DNA damage response gene mutations, and clinical outcomes will be examined by the same methods as Aim 1. Mutational load (mutations variants of unknown significance/megabase) will be examined as a continuous variable for association with clinical outcomes using proportional hazards methods. Relationships between thes biomarkers and other clinical covariates (e.g. imaging features and anatomic subsite) will be explored. Impact: This project will leverage the infrastructure of the ongoing KEY-G trial to rapidly examine candidate biomarkers of PD-L1 and tumor genotype for association with response to the novel combination of pembrolizumab plus GM-CSF. These biomarker analyses are essential for analysis of the efficacy of KEY-G study and will inform future stratification and enrichment strategies in biliary cancer immunotherapy trials.
Cancers of the biliary tract including cholangiocarcinoma are rapidly rising in incidence but have limited standard treatment options and very grim outcomes. A subset of patients with biliary cancers has shown remarkable responses to immunotherapy with checkpoint inhibitors. This project will use previously-collected clinical data and archival tumor samples from patients enrolled on an investigator-initiated clinical trial of a novel immune checkpoint inhibitor combination to study potential biomarkers of response (including tumor PD- L1 expression and tumor DNA sequencing), with the future goal of identifying those patients most likely to benefit from immunotherapy treatment strategies.