Clinical validation of Northstar Response for monitoring mono- and combination immunotherapy efficacy in a multi-cancer cohort.

Abstract

e14548 Background: Methylated circulating tumor DNA (ctDNA) enables non-invasive, real-time treatment response monitoring compared to conventional imaging. Unlike traditional ctDNA approaches that track specific mutations requiring tissue biopsies, methylation analysis offers enhanced sensitivity by increasing the amount of traceable loci. This study evaluates the clinical validity of Northstar Response, a tissue-free, methylated ctDNA assay that quantifies changes across methylated loci to generate a Tumor Methylation Score (TMS) and its association with therapeutic response. Methods: This retrospective analysis evaluated matched plasma and buffy coat samples from patients treated at Allegheny Health Network with either immuno-monotherapy or immuno-combination (IC) therapy. All patients had baseline (pre-treatment) samples and at least 1 on-treatment sample collected prior to imaging assessment. Samples were analyzed using the Northstar Response assay to generate a TMS. Using a pre-defined threshold, molecular response was defined as no change or decrease in TMS versus baseline; molecular non-response was defined as an increase in TMS score from baseline. Associations with progression-free survival (PFS) and overall survival (OS) were evaluated using Cox proportional hazard ratios (HR). Results: Among 54 evaluable patients, 61% had lung cancer, 19% melanoma, and 20% other cancers, with 41% receiving immunotherapy monotherapy and 59% IC therapy. High baseline TMS predicted shorter PFS (HR = 3.8, p < 0.001) and OS (HR = 3.1, p = 0.009), while molecular response at 3 months predicted improved PFS (HR = 0.26, p < 0.001) and OS (HR = 0.18, p < 0.001). Combined analysis showed patients with low baseline TMS and molecular response had the best outcomes (HR = 0.19, p = 0.01), while those with high baseline TMS without response had the worst (HR = 5.8, p = 0.007) when compared to low baseline, molecular non-responders as reference. Disease control rate (DCR) assessment showed significantly different TMS fold-changes (FC) between progressive disease (mean FC = 2.0, n = 11) and non-progressive disease (mean FC = 0.4, n = 33; p < 0.004). Conclusions: TMS analysis provides prognostic and predictive information for immunotherapy outcomes across multiple cancer types. The combination of baseline TMS and early on-treatment molecular response effectively stratifies patient outcomes. Additionally changes in the TMS are significantly associated with DCR. Cumulatively these data suggest that the TMS is a clinically meaningful biomarker that associates with patient survival and radiographic assessment.