A team of researchers from Geneseeq and a network of Chinese academic hospitals has successfully validated a blood-based test capable of detecting multiple cancers with high accuracy using cell-free DNA (cfDNA). Published in Nature Medicine, the study presents a promising tool for early cancer diagnosis—a critical area where traditional screening methods still fall short.
High Sensitivity and Specificity in Validation Studies
The multi-cancer early detection (MCED) test demonstrated 87.4% sensitivity and 97.8% specificity in an independent validation cohort. Moreover, the test correctly predicted the tissue of origin in 83.5% of cases and up to 91.7% when considering the top two likely sources. These findings highlight the test’s potential to significantly enhance diagnostic accuracy and clinical decision-making.
Addressing the Gaps in Early Detection
Early-stage cancer often remains undetected due to the limitations of current screening methods, especially for cancers without established protocols. Researchers focused on cfDNA—fragments of DNA released into the bloodstream by tumors—as a noninvasive and promising biomarker. However, until now, detecting early-stage and rare cancers using cfDNA has been challenging.
Advanced Technology Powers the Detection Framework
As reported by medicalxpress, the team developed a whole-genome sequencing (WGS)–based blood test and applied machine learning models to interpret cfDNA fragmentation patterns. Two supervised classifiers were trained: one to detect the presence of cancer, and another to predict the tissue of origin. These models utilized multidimensional fragmentomics features, such as fragment size, copy number variation, nucleosome positioning, and inferred methylation profiles.
Robust Study Design Ensures Data Integrity
The researchers used data from 3,076 cancer patients and 3,477 non-cancer controls for model training. Validation involved two cohorts: an internal group of 1,746 participants and an independent group of 1,465. Importantly, the study followed a double-blind protocol, keeping data analysts and clinical teams unaware of each other’s findings. A standardized bioinformatics pipeline ensured consistency across all stages of testing.
Test Performance Across Cancer Types and Stages
In the independent validation cohort, the MCED test showed impressive results:
- Stage I sensitivity: 79.3%
- Stage II sensitivity: 86.9%
- Liver and bile duct cancers: 100% sensitivity
- Lung cancer: 94.5%
- Ovarian cancer: 90.5%
- Colorectal cancer: 82.3%
- Pancreatic cancer: 76.9%, including 58.3% for stage I
Prospective Study Confirms Test’s Real-World Potential
The prospective Jinling cohort enrolled 3,724 asymptomatic individuals at two Chinese medical centers. Within one year, the test identified 43 cancer cases with:
- Sensitivity: 53.5%
- Specificity: 98.1%
- Sensitivity for 13 targeted cancers: 62.1%
- Positive predictive value: 25%
- Negative predictive value: 99.4%
- Tissue-of-origin accuracy: 63.2% (top prediction), 89.5% (top two predictions)
Remarkably, nearly half of the cancers detected in this cohort were missed by standard screenings or physical exams.
Implications for Future Cancer Screening
The MCED test’s high sensitivity for difficult-to-detect cancers—such as liver, ovarian, and pancreatic—demonstrates its clinical value. Its ability to predict tissue of origin further strengthens its role in guiding early interventions and treatment planning.
Conclusion: A Promising Step Toward Early Detection
The researchers concluded that their findings strongly support the MCED test’s potential to transform early cancer detection. By offering accurate, noninvasive diagnosis and tissue identification, the test may soon become a vital tool in the global fight against cancer.