在液体处理平台上进行自动化高体积cfDNA提取，以96样本通量实现与手动高投入量磁珠工作流程的分析等效性

Introduction: High-volume cfDNA extraction improves sensitivity for liquid biopsy studies but is limited by manual hands-on time and variability. We evaluated whether implementing a high-volume magnetic-bead cfDNA workflow on an automated liquid-handling platform maintains analytical performance while increasing throughput and reducing labor. Methods: Human plasma collected in K2EDTA tubes was processed using double centrifugation (low- then high-speed). Urine was preserved in Tris-EDTA pH 8 and clarified by high-speed centrifugation. Plasma and urine inputs (1-20 mL) were extracted manually or on the automated platform using identical magnetic-bead chemistry; only wash/elution steps were automated. Replicates were run across multiple reagent lots and days. Endpoints included cfDNA yield (Qubit HS), fragment metrics (Agilent cfDNA ScreenTape), and locus-specific qPCR. Pre-specified non-inferiority margins were ±10% for yield and fragment attributes. Equivalence was assessed with paired statistics, TOST, Bland-Altman, and variance components. Operational metrics included batch size, hands-on time (HOT), turnaround time (TAT), and walk-away time. Cross-contamination was assessed using checkerboard layouts analyzed by Qubit and qPCR. End-to-end traceability used unique barcodes on capture tubes readable by the automation platform. Results: Automation preserved analytical performance across matrices and input volumes. The automated/manual yield ratio was 0.93 (95% CI ±7.8%), meeting the ±10% non-inferiority criterion (TOST p < 0.05). Fragment metrics-including modal size (~170 bp), 50-700 bp distribution, and mono:di-nucleosome ratios-were equivalent, differing by ~1% with between-run CV ≈1%. qPCR copy numbers showed minimal bias (mean 3.35 copies/reaction) with acceptable limits of agreement; limits of detection were identical for all loci. All 96-sample runs met QC thresholds. Checkerboard assays showed no detectable cross-contamination. Automation increased batch capacity from ~24-48 to 96 samples/run, reduced HOT by ~70-85% (to ~2 hours per batch), and shortened TAT by ~40-60%, providing ~2.5 hours of walk-away time. Conclusions: Automating a high-volume magnetic-bead cfDNA workflow delivers analytical equivalence to manual processing while enabling 96-sample throughput and major reductions in labor and turnaround time. These findings support automation as a scalable solution for high-volume cfDNA programs requiring consistent, high-sensitivity performance. Portions of this text were generated with AI and were reviewed, edited, and approved by the authors.