Preliminary translational immune and stromal correlates in a randomized phase II trial of pembrolizumab with or without defactinib for resectable pancreatic ductal adenocarcinoma (PDAC)

4024 Background: PDAC is an aggressive cancer and refractory to immunotherapy due to its immunosuppressive tumor microenvironment (TME). Focal adhesion kinase (FAK) is a master regulator of the TME and associated with TME immune suppression. Our current randomized phase II trial evaluates the use of pembrolizumab, a programmed cell death 1 (PD-1) immune checkpoint inhibitor, with or without defactinib, a FAK inhibitor (FAKi), as sequential neoadjuvant and adjuvant therapy in patients with high risk resectable PDAC. We hypothesize that the patients receiving pembrolizumab and defactinib treatment will exhibit a decrease in immunosuppressive fibroblasts and myeloid subtype populations, leading to increased CD8+ T-cell infiltration into the TME compared to patients receiving pembrolizumab alone. Methods: We performed quantitative analysis of multiplex immunohistochemistry (mIHC) using 40 biomarkers, evaluating immune and stromal cell types on 14 pre-treatment biopsies and post-treatment resections of PDAC patients enrolled in our platform neoadjuvant clinical trial (NCT03727880). All patients received 2 cycles of gemcitabine+nab-paclitaxel neoadjuvant chemotherapy and underwent a biopsy after completion of chemotherapy. Patients randomized to Arm A received 2 cycles of pembrolizumab 200 mg IV every 3 weeks and defactinib 400 mg PO BID, and those randomized to Arm B received pembrolizumab alone, followed by surgical resection. Image cytometry was used to quantify immune cell populations and colocalize biomarker expression in distinct cell types. Cell populations were compared using unpaired T-tests. Results: Lower FAP+ fibroblast density was significantly associated with higher CD8+ T-Cell infiltration in Arm A (p=0.002), but not in Arm B. Patients in Arm A demonstrated a 5.44-fold average increase in CD8+ T-cell percentage between pre-immunotherapy treatment and post-immunotherapy treatment specimens compared to a 2.01-fold increase in patients in Arm B (p=0.02, p=0.09, respectively). Only patients in Arm A showed a significant increase in M1 macrophage cell density (p=0.02) after treatment. Both arms increased in CXCR4+ cell percentage (p=0.011, p=0.06) following neoadjuvant immunotherapy; however, the increase in Arm A was less than in Arm B (0.72 and 0.84-fold, respectively). Conclusions: In this analysis, pembrolizumab combined with defactinib was associated with lower fibroblast infiltration, higher anti-tumor M1 macrophage expression and increased CD8+ T-cell infiltration into the TME, versus pembrolizumab alone. The increased expression of CXCR4 across both treatment arms may represent a resistance mechanism and supports CXCR4 as an additional TME target. These preliminary findings warrant continued research into FAK inhibition and immune checkpoint combinatorial strategies. Clinical trial information: NCT03727880 .