Enspyre：一种新型富集技术实现超灵敏ctDNA检测并将测序需求降低98%

Background: Circulating tumor DNA (ctDNA) liquid biopsies show promise for minimal residual disease (MRD) detection, but clinical implementation is limited by high sequencing costs and the need for ultra-deep coverage to detect ctDNA at levels ≤100 parts per million (ppm). Current tumor-informed assays require >500 million reads per sample, necessitating sample batching and high-throughput platforms, which limits accessibility and increases turnaround times. Methods: We evaluated Enspyre (Enrichment by selective pyrophosphorolysis and release) , a novel enrichment technology that enables selective enrichment of specific variant molecules rather than just target regions. Eight lung cancer patients underwent whole genome sequencing of FFPE tumor tissue (median coverage 118x) for personalized probe design targeting a median of 1,995 somatic variants. Patient plasma samples were diluted to create a concentration series (5-1000 ppm) using healthy donor plasma. Enspyre enrichment was performed on low-input cfDNA samples (median 7.67 ng) followed by sequencing on NextSeq 550 with only 9.6 million read pairs per sample. ctDNA detection and quantification were performed using a Bayesian MRD estimation model, with analyses conducted blinded to ground truth concentrations. Results: Enspyre demonstrated exceptional analytical performance across 72 patient samples and 8 controls. The assay achieved 100% sensitivity at 10 ppm (6/6 samples detected) and 20% sensitivity at 5 ppm (1/5 samples detected) without molecular barcodes. 100% specificity was maintained with no false positives in control samples (10/10 correctly called negative). Quantitative ctDNA estimates showed strong linear correlation with expected values (r=0.90, p<2.2×10⁻¹⁶), with estimates averaging 1.22-fold of target concentrations. Performance was maintained despite low DNA inputs, with successful detection from as little as 0.84 ng cfDNA. Compared to standard hybrid capture methods, Enspyre achieved equivalent sensitivity with a 98% reduction in sequencing depth (10M vs 500M reads per sample). Conclusions: Enspyre enables ultra-sensitive ctDNA detection at 10 ppm with dramatically reduced sequencing requirements, addressing key barriers to clinical implementation. The technology's ability to maintain performance with low DNA inputs and simplified workflows makes ctDNA testing accessible on benchtop sequencers, potentially enabling broader adoption in clinical trials and community oncology. Sample throughput increases from 1 to 40 samples per NextSeq run represent a 40-fold improvement in accessibility. These results support Enspyre's potential to democratize ctDNA-based MRD detection while maintaining the analytical rigor required for clinical decision-making.