Development and validation of predictive models combining cell-Free DNA motifs and protein biomarkers for early detection of esophageal squamous cell carcinoma and precancerous lesion.

backgroundDetecting and treating precancerous lesions can lower the incidence of esophageal squamous cell carcinoma (ESCC), making it a key preventive strategy. Although endoscopic screening and intervention can significantly reduce mortality associated with ESCC, they have certain shortcomings. We aimed to develop three predictive models: the motif, eight-protein, and combined motif-protein models to identify ESCC and its precancerous lesions.

methodsPlasma samples were collected for cfDNA sequencing, and nine commonly used clinical protein biomarkers related to the digestive system were measured. Using a total cohort of 199 patients with ESCC, 91 patients with esophageal squamous precancerous lesions (ESPL), and 201 controls, we developed an integrative model based on selected multi-omics biomarkers.

resultsThe motif-protein model, integrating 20 principal components of cfDNA terminal motifs with six protein features, outperformed both the motif model and the eight-protein model in distinguishing patients with ESCC and ESPL (area under the curve = 0·90). It achieved an overall sensitivity of 88·5% and a specificity of 75·4%. Notably, it successfully identified 90·9% of high-grade intraepithelial neoplasia cases, 86·8% of stage I ESCC cases, and 87·8% of HGIN or T1aN0 stage ESCC (subset of Stage I) cases who were eligible for endoscopic treatment, highlighting its potential as an effective tool for early diagnosis.

conclusionsThe motif-protein model may serve as an effective tool for the early diagnosis of esophageal lesions. Our findings underscore the clinical potential of the multi-omics liquid biopsy test as a non-invasive method for detecting early esophageal lesions.