A low-cost bioassay for multi-cancer detection in cfDNA using quantitative methylation-specific PCR

Liquid biopsy multi-cancer detection tests hold significant promise to improve patient outcomes by enabling earlier intervention. However, emerging sequencing-based assays face practical challenges due to high costs, complex workflows, and long turnaround times. In contrast, quantitative methylation-specific PCR (qMSP) enables rapid turnaround times with low-cost, low-complexity, and scalable workflows, while accommodating flexible batch sizes and decentralized testing. Harbinger Health's proprietary methylation biomarkers were applied to Harbinger Health's proprietary qMSP method to enable low-cost multi-cancer detection. Harbinger Health's biomarkers are associated with the initiation of oncogenesis and are ubiquitously observed across multiple cancer types. This redundancy enabled the selection of six highly pan-cancer-informative methylation patterns, reducing the genomic footprint while largely maintaining cancer detection performance. Despite the informative nature of these biomarkers, background methylation from processes like aging can attenuate disease detection. To address this limitation, Harbinger developed a qMSP method that incorporates locked nucleic acid (LNA) blockers to suppress probe mishybridization to stochastically methylated somatic cell-free DNA (cfDNA), improving selectivity for defined methylation patterns in low abundance circulating tumor DNA (ctDNA). The qMSP panel underwent analytical validation during development and demonstrated high sensitivity, specificity, and reproducibility for detecting cfDNA fragments with defined methylation patterns. The panel was applied to 130 cfDNA samples obtained from the CORE-HH clinical study (NCT05435066), representing 62 non-cancer samples and 68 cancer samples across nine cancer types. Samples were selected to reflect a similar range of methylation signals observed in the CORE-HH study cohort. The cancer samples represented a balanced stage distribution. cfDNA samples were processed into bisulfite libraries and analyzed by qMSP using 5 ng of library per reaction. The cycle threshold (Ct) of each methylation marker was normalized to an internal reference assay and delta Ct values were used to measure methylation. Using target-specific detection thresholds established on 62 non-cancer samples to achieve an equivalent of 97% specificity, qMSP detected at least one positive target for 30 of 68 (44%) cancer samples. This work represents an initial proof of concept for a low-cost multi-cancer detection assay. qMSP reactions in this study cost less than $5 per sample and were completed in under five hours, addressing key barriers to population-scale implementation such as cost and turnaround time. Future refinements, including expansion of the marker panel, optimization of detection thresholds, and improved assay design, are expected to further improve clinical performance.