Acceptable colorectal cancer detection but poor detection of advanced precancerous lesions.
Background
Colorectal cancer (CRC) screening reduces incidence and mortality but remains underused. A blood-based screening option may improve adherence compared with colonoscopy or stool-based tests, but requires validation in average-risk populations against colonoscopy with histopathology.
Patients
- Asymptomatic, average-risk adults aged 45–85 years undergoing routine CRC screening colonoscopy.
- Key exclusions: strong family history, hereditary GI cancer syndromes, prior CRC/adenomas, inflammatory bowel disease, recent CRC screening tests.
- Evaluable cohort: 27,010 participants (median age 57 years; 55.8% women).
- Race/ethnicity: 73.0% White, 11.2% Black/African American, 8.8% Asian; 11.8% Hispanic/Latino.
Intervention
Investigational blood-based circulating tumor DNA (ctDNA) test assessing plasma CpG methylation patterns associated with advanced colorectal neoplasia. Prespecified, locked classifier; testing performed blinded to colonoscopy findings.
Control
Reference standard: screening colonoscopy with histopathology (central pathology review with tiered adjudication), performed within 120 days of blood draw; all parties blinded to blood test results.
Outcome
- Primary endpoints: Sensitivity for CRC; specificity for advanced colorectal neoplasia (CRC or advanced precancerous lesions); negative predictive value (NPV) and positive predictive value (PPV) for advanced colorectal neoplasia.
- Secondary endpoint: Sensitivity for advanced precancerous lesions (eg, ≥1 cm adenomas, villous features, high-grade dysplasia/carcinoma in situ, sessile serrated lesions ≥1 cm, traditional serrated adenomas).
Study Design
Prospective, multicenter, cross-sectional diagnostic accuracy study at 201 centers across 49 US states and the UAE (May 2020–April 2022). Participants, site staff, laboratory personnel, and pathologists were blinded.
Level of Evidence
Level 1 diagnostic accuracy evidence (prospective validation with appropriate reference standard and blinding; Oxford CEBM for diagnostics).
Follow up period
Cross-sectional assessment from blood draw to colonoscopy within 120 days (initially 90 days; extended due to COVID-19).
Results
Primary outcomes
- Sensitivity for CRC: 79.2% (57/72; 95% CI, 68.4%–86.9%).
- Stage-specific sensitivity: Stage I 57.1% (16/28); Stage II 100% (15/15); Stage III 82.4% (14/17); Stage IV 100% (11/11).
- Specificity for advanced colorectal neoplasia: 91.5% (22,306/24,371; 95% CI, 91.2%–91.9%).
- NPV for advanced colorectal neoplasia: 90.8% (22,306/24,567; 95% CI, 90.7%–90.9%).
- PPV for advanced colorectal neoplasia: 15.5% (378/2,443; 95% CI, 14.2%–16.8%).
Secondary outcomes
- Sensitivity for advanced precancerous lesions: 12.5% (321/2,567; 95% CI, 11.3%–13.8%).
- High-grade dysplasia/carcinoma in situ: 29.1% (32/110; 95% CI, 21.4%–38.2%).
- Safety: No serious adverse events related to phlebotomy or testing.
- NNT: Not applicable for a diagnostic accuracy study.
Limitations
- One-time diagnostic accuracy only; no head-to-head adherence or effectiveness outcomes versus other modalities.
- Uncertain optimal screening interval for this blood test.
- Lower sensitivity for advanced precancerous lesions limits preventive impact without further improvements.
- Modeling needed to quantify long-term effects on CRC incidence/mortality considering adherence and follow-up colonoscopy completion.
- Specificity decreased with age; generalizability to older subgroups may vary.
Funding
Funded by Freenome Holdings Inc., which contributed to study design; data collection, management, analysis, and interpretation; manuscript preparation; and publication decisions.
Citation
Shaukat A, Burke CA, Chan AT, Grady WM, Gupta S, Katona BW, et al.; for the PREEMPT CRC Investigators. Clinical Validation of a Circulating Tumor DNA–Based Blood Test to Screen for Colorectal Cancer. JAMA. 2025;334(1):56-63. doi:10.1001/jama.2025.7515. Published online June 2, 2025. ClinicalTrials.gov: NCT04369053.