ctDNA and TCR dynamics predict response toimmune checkpoint blockade in non-small cell lung cancer

The dynamic nature of the cancer-immune system crosstalk has rendered single biomarker approaches insufficient to predict outcome from immune-targeted agents. These therapies also pose a challenge to radiographic response assessments that may underestimate the therapeutic benefit. There is therefore an urgent clinical need to develop molecular assays of response and resistance.In order to capture the dynamic nature of the anti-tumor immune response under checkpoint blockade, we devised an integrative non-invasive approach that captures the evolving neoantigen landscape and the associated T-cell receptor (TCR) repertoire. We employed whole exome sequencing to determine the genomic landscape of tumors and identify tumor-derived alterations in subsequent analyses of circulating cell-free tumor DNA (ctDNA), for 14 patients with metastatic NSCLC treated with immune checkpoint blockade. Liquid biopsies were obtained prior to treatment, at week 4-6 and at additional timepoints until disease progression. We used the ultrasensitive targeted error correction sequencing approach to analyze 58 cancer driver genes in the circulation of these patients and assessed the value of longitudinal ctDNA monitoring as a surrogate for response. In parallel, we evaluated dynamics changes in the TCR repertoire by TCR next generation sequencing of serially collected peripheral T cells and tumor infiltrating lymphocytes for each patient. Radiographic response assessments were performed using the RECIST 1.1 criteria. These analyses revealed that ctDNA dynamics predict outcome significantly earlier than imaging. Patients that responded to therapy had a significant drop in ctDNA early during the treatment course whereas non-responders had either limited changes or significant increase in ctDNA. For patients with acquired resistance, ctDNA kinetics revealed clearing of ctDNA at the time of response followed by a new peak at the time of progression. ctDNA detection early during the treatment course was a significant prognostic factor for progression-free and overall survival (log rank p=.004 and .002 respectively) and ctDNA elimination was superior to tumor mutation burden as a predictor of response to therapy. In parallel, we identified changes in the TCR repertoire that are predictive of outcome: peripheral T cell expansion of a subset of intra-tumoral clones was noted at the time of response whereas there was no evidence of T cell expansion in non-responders. Peripheral T cell expansion of a subset of intra-tumoral clones was noted to peak at the time of response and decrease to baseline levels at the time of resistance for patients with acquired resistance. Our findings suggest that a dynamic assay able to capture the tumor-immune system equilibrium would be superior to conventional analyses of static time points. Such an integrated approach would be highly relevant to tailored cancer immunotherapy strategies. Citation Format: Valsamo Anagnostou, Patrick Forde, Jarushka Naidoo, Kristen Marrone, Vilmos Adleff, James White, Jillian Phallen, Alessandro Leal, Carolyn Hruban, Ashok Sivakumar, Franco Verde, Rachel Karchin, Julie Brahmer, Victor Velculescu. ctDNA and TCR dynamics predict response toimmune checkpoint blockade in non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3668.