MethylScan Blood Test Achieves 93.8% Accuracy in Detecting Four Cancer Types Through Novel DNA Methylation Analysis
Key Insights
MethylScan (search), a novel blood test using methylation-sensitive restriction enzymes, achieved 93.8% accuracy in detecting liver, lung, ovarian, and stomach cancers in a study of 1,061 individuals.
The test demonstrated 63.3% sensitivity at 98% specificity for all cancer stages and 55.3% sensitivity for early-stage cancers, with particularly strong performance in liver cancer surveillance at 79.6% sensitivity.
By selectively removing healthy blood DNA background, MethylScan (search) reduces sequencing costs by 5.1-fold while detecting cancer signals at tumor fractions as low as 0.05%.
A breakthrough blood test called MethylScan (search) has demonstrated remarkable accuracy in detecting four major cancer types while simultaneously identifying liver diseases and organ damage, according to research published in PNAS. The novel diagnostic platform achieved an area under the receiver operating characteristic curve (AUROC) of 0.938 for multicancer detection across liver, lung, ovarian, and stomach cancers in a validation study of 1,061 individuals.
Revolutionary Approach to Cancer Detection
MethylScan (search) represents a significant advancement in liquid biopsy technology by addressing a fundamental challenge in cell-free DNA (cfDNA (search)) analysis. Approximately 85% of circulating cfDNA originates from healthy blood cells, creating overwhelming background noise that masks disease signals. During early-stage cancer development, tumor-derived cfDNA may constitute less than 0.1% of the total sample, requiring expensive deep sequencing approaches that have limited clinical feasibility.
The MethylScan (search) method employs specialized methylation-sensitive restriction enzymes (MSREs) - specifically HpaII (search) and HhaI (search) - that selectively cleave DNA only when recognition sites are hypomethylated, a characteristic state of background cfDNA (search) from white blood cells. This innovative approach effectively strips away healthy DNA noise, enriching rare disease signals while reducing sequencing costs by 5.1-fold compared to conventional methods.
Clinical Performance Across Cancer Types
In the validation cohort comprising 460 cancer patients and 601 non-cancer individuals, MethylScan (search) demonstrated robust diagnostic performance. At a specificity of 98.0%, the assay achieved 63.3% sensitivity across all cancer stages and 55.3% sensitivity for early-stage (Stage I/II) cancers. For Stage I cancers specifically, the AUROC reached 0.906 with 54.4% sensitivity.
The test's ability to distinguish between the four cancer types proved equally impressive, with an accuracy of 91.7% for all stages and 89.8% for early-stage cancers. Analytical validation using liver tumor DNA dilution series confirmed the assay's capability to detect cancer signals at tumor fractions as low as 0.05% with strong linear correlation between expected and estimated fractions (R² = 0.983).
Multi-Disease Detection Capabilities
Beyond cancer detection, MethylScan (search) demonstrated versatility in identifying various liver conditions. The assay classified patients with Hepatitis B, Hepatitis C, alcoholic liver disease, and metabolic dysfunction-associated steatotic liver disease (MASLD) with an overall accuracy of 84.7%. For liver cancer surveillance specifically, the test achieved an AUROC of 0.927 with 79.6% sensitivity at 90.4% specificity.
The platform's tissue deconvolution capabilities revealed significantly elevated liver-derived cfDNA (search) fractions in patients with liver cancer (p = 4.1 × 10⁻¹⁶) and liver disease (p = 5.6 × 10⁻³) compared to general hospital visitors, highlighting its potential for organ-specific damage assessment.
Technical Innovation and Cost-Effectiveness
The MethylScan (search) protocol utilizes a custom-designed capture panel comprising 154,028 target regions totaling 1,600,725 CpG sites, selected based on consistent hypomethylation patterns across white blood cell and healthy plasma samples. The method requires only 10 ng of cfDNA (search) per sample and incorporates 300 control regions for read count normalization.
To optimize model performance and prevent overfitting, researchers applied singular value decomposition for dimensionality reduction before training Linear Support Vector Machine classifiers. This computational approach, combined with the enzymatic background depletion, enables the detection of multiple diseases from a single blood sample while maintaining cost-effectiveness for population-level screening.
Implications for Clinical Practice
The research findings suggest that comprehensive multi-disease screening through a single blood test may become increasingly accessible and scalable. The ability to simultaneously detect multiple cancers, liver diseases, and organ damage could transform current screening paradigms and enable earlier clinical intervention.
The study's demonstration of methylation pattern analysis for ancestry prediction, achieving 97.1% accuracy in distinguishing White and Asian participants, further illustrates the platform's analytical sophistication and potential for personalized medicine applications.
As the field moves toward holistic diagnostic approaches, MethylScan (search)'s capacity for continuous data stacking positions it well for future machine learning and artificial intelligence applications, potentially expanding its diagnostic scope beyond the current four cancer types and liver conditions.