Clinical Validation and Diagnostic Performance of Targeted Cell-Free DNA Sequencing for Detection of Multiple Cancer Types

Background/Objectives: Blood-based multi-cancer detection has the potential to complement existing organ-specific screening by enabling simultaneous detection of tumour-associated signals across multiple malignancies. This study aimed to clinically validate an ultradeep targeted cell-free DNA (cfDNA) sequencing assay for cancer detection across 19 cancer types and clinical stages. Methods: This retrospective case-control study analysed 4,662 archived plasma samples collected between 2022 and 2025, comprising 2,667 participants with histologically confirmed, treatment-naïve malignancies and 1,995 cancer-free controls. cfDNA was ana-lysed using a custom hybridisation-capture next-generation sequencing panel, with unique molecular identifiers and ultradeep sequencing. Diagnostic performance was assessed across various cancer type and clinical stage. Results: Overall sensitivity was 88.6% (95% CI, 87.4-89.8%) and specificity was 99.2% (95% CI, 98.8-99.5%), with an overall accuracy of 93.2% (95% CI, 92.4-93.9%). Sensitivity exceeded 80% in 15 of 19 cancer types and was 90% or greater in 10 of 19 cancer types. Pan-cancer sensitivity was 74.3% (450/606; 95% CI, 70.6-77.6%) in stage I, 85.4% (573/671; 82.5-87.9%) in stage II, 96.2% (608/632; 94.4-97.4%) in stage III, and 98.7% (665/674; 97.5-99.3%) in stage IV. The test demonstrated particularly strong performance for the early detection of breast, colorectal, gastric, liver, ovarian, and pancreatic cancers, with combined stage I–II sensitivity exceeding 80% for these tumour types. This early-stage sensitivity was substantially higher than that reported for the corresponding conventional serum tumour markers. Conclusions: Ultradeep targeted cfDNA sequencing demonstrated broad multi-cancer detection across diverse tumour types, with substantial sensitivity in stages I-II disease. These findings support further prospective validation of this approach in intended-use screening populations and evaluation of its potential integration with existing cancer detection pathways.