Abstract BACKGROUND: Metastatic non-small cell lung cancer (NSCLC) tumors have adopted methods to evade immune detection and/or clearance. This can occur via overexpression of programmed cell death ligand 1 (PD-L1). Response rate, progression free survival and overall survival with PD-1 inhibitors are greater in tumors with high tumor PD-L1 expression (Garon et al, NEJM 2015, Paz-Ares et al, ASCO 2015). There has been interest in using PD-L1 tumor expression as a treatment selection criterion. Currently available methods of screening involve invasive tumor biopsy followed by histological grading of PD-L1 levels. Biopsies allow sampling from limited sections of the tumor, which may miss heterogeneity. CTC PD-L1 levels could aid in screening patients, and could supplement tissue PD-L1 biopsy results by evaluating a representative sampling from multiple tumor sites which shed cells into the blood. Additionally, following PD-L1 levels on CTCs serially over time may potentially yield information about modulation of tumor PD-L1 expression in the presence of PD-1 or PD-L1 antibodies or other anti-cancer agents. METHODS: We have developed a microfluidic device for rapid, size-based capture of CTCs from blood called Vortex HT chip. The Vortex HT chip utilizes inertial microfluidic flows to isolate CTCs with capture efficiency up to 40% and high purity (>80%). We used the Vortex HT device to capture CTCs from NSCLC patients undergoing PD-1 immunotherapies, both prior to initiating treatment and during treatment. PD-L1 expression was evaluated on CTCs using immunofluorescence staining. CTC number and PD-L1 expression were correlated with treatment response as evaluated by immune related response criteria (irRC) on serial CT scans. We also measured cell size and intensity levels of cytokeratin (CK) and CD45 on the collected cells. We developed a semi-automated algorithm to quantify fluorescence for these different markers on DAPI positive cells collected from each patient sample. We compared these results to PD-L1 expression on the initial tumor biopsy sections, as assayed by immunohistochemical staining and expression quantified with HALO software (Indica Labs). RESULTS: In patients receiving anti-PD-1 antibodies, PD-L1 expression on CTCs could be quantified and compared to a pre-treatment tumor biopsy, as well as to radiographic treatment response. Evaluating patient CTC count and PD-L1 expression at baseline and over time may lead to simple and non-invasive methods to predict response to immunotherapies. As an assay amenable to repeat testing, CTC PD-L1 levels may also be an important pharmacodynamics marker to assess the synergistic potential of combination immunotherapies. Further work is continuing to better understand this predictive biomarker. At the meeting, we will present correlation data between CTC and tumor biopsy PD-L1 expression assays and between the CTC assay and radiographic response to treatment. Citation Format: Jonathan W. Goldman, Manjima Dhar, James Che, Edward B. Garon, Siwen Hu-Lieskovan, Melissa Matsumoto, Brian R. Wolf, James M. Carroll, Matthew J. Crabtree, D. Andrew Tucker, Jennifer Strunck, Elodie Sollier, Rajan Kulkarni, Dino Di Carlo. Serial evaluation of PD-L1 expression on circulating tumor cells (CTCs). [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr B98.
Read full abstract