Plasma-guided adaptive first-line chemoimmunotherapy for non-small cell lung cancer (NSCLC).

Abstract

8515 Background: There is clinical uncertainty as to which patients with advanced/metastatic NSCLC require first-line treatment with chemoimmunotherapy (ChemoIO) versus treatment with immune checkpoint inhibitor (ICI) monotherapy. While PD-L1 expression can predict for ICI response, it is an imperfect biomarker. Cell-free DNA (cfDNA) clearance is a dynamic biomarker that predicts for benefit to immunotherapy in metastatic NSCLC. Here we evaluate the use of cfDNA clearance after initial cycles of first-line ICI monotherapy to guide early treatment intensification via addition of platinum-based chemotherapy. Methods: In this prospective clinical trial patients with advanced/metastatic PD-L1 positive (TPS ≥ 1%) NSCLC were treated with two cycles of pembrolizumab monotherapy and then assessed for plasma and radiographic response. Plasma response was defined as ≥50% reduction in the maximum variant allele fraction and/or persistent low-shedding status at cycle two day one compared to pretreatment using an amplicon-based plasma NGS assay. Patients with radiographic response or with radiographic stable disease with plasma response continued pembrolizumab monotherapy. Patients with RECIST stable disease without plasma response were intensified to carboplatin doublet (paclitaxel squamous/pemetrexed nonsquamous) plus pembrolizumab. Those with radiographic progression ended study treatment. Results: Forty patients were enrolled across six sites. 56.8% (n=21) had nonsquamous histology and 37.5% (n=15) were PD-L1 low (TPS 1-49%). 36 patients (90%) completed a C2D1 plasma response assessment, with plasma response to ICI monotherapy in 58.3% (n=21) patients (57.1% in PD-L1 low, 59.1% in PD-L1 high [TPS ≥ 50%]). At cycle 3, 52.8% (n = 19) continued pembrolizumab monotherapy (7 with PR, 12 with SD and plasma response). 19.4% (n=7) had radiographic SD and plasma non-response and received intensification to ChemoIO (pemetrexed in 6, paclitaxel in 1). 27.8% (n=10) went off study treatment for PD (n=4), death (n=1), adverse event (n=3), or patient/physician decision (n=2). 20% (n=3) of PD-L1 low patients and 16% (n=4) of PD-L1 high patients received treatment intensification to ChemoIO. The ORR to this adaptive treatment strategy was 50%, the median PFS was 11 months (95% CI 3.4-15.9 months), and the median OS was 14.9 months (95% CI 8.2-27.2 months). Median PFS was higher in patients with plasma response (16.4 vs 4.8 months; HR 0.34; 95% CI 0.12-0.92). Fewer patients were treated with platinum doublet chemotherapy than would have been predicted by PD-L1 status alone (17.5% vs 37.5%). Conclusions: The use of a cfDNA-guided adaptive treatment design resulted in a median progression-free survival which compared favorably to historical controls with less upfront exposure to platinum doublet chemotherapy. Further study within a randomized prospective trial design is needed to validate this treatment strategy. Clinical trial information: NCT04166487 .