2026/05/12 by Kevin Connor McGann, Palina Woodhouse, Heidi Chen +26
paper · doi:10.1158/1055-9965.epi-25-2004
Abstract Background: Lung cancer remains the leading cause of cancer mortality, yet blood-based biomarkers are not routinely used in diagnosis. This study evaluated four commercial blood protein assays, originally validated for other indications, in indeterminate pulmonary nodules (IPN). Methods: Using a prospective specimen collection, retrospective blinded evaluation design, samples were collected from patients with screening-detected, incidental, or symptomatic IPNs. Cytokeratin 19 fragment (CYFRA 21-1), carcinoembryonic antigen (CEA), cancer antigen 125 (CA-125), and human epididymis protein 4 (HE-4) concentrations were quantified on commercial immunoassays. Logistic regression models were developed using internal training (Train), external testing (Test), and combined reestimation (Train + Test) cohorts and externally validated in an outcome-blinded multicenter cohort (Lung Team Project-2, LTP-2). Results: This study included 816 patients: 371 in Train, 166 in Test, and 279 in LTP-2. Malignancy rates were 54%, 44%, and 64%, respectively. In Train + Test, the area under the receiver operating curve (AUC) for lung cancer was 0.60 (95% confidence interval, 0.56–0.65) for CYFRA 21-1, 0.62 (0.58–0.67) for CEA, 0.60 (0.55–0.65) for CA-125, and 0.65 (0.60–0.70) for HE-4. In LTP-2, AUCs were 0.63 (0.56–0.70), 0.64 (0.57–0.70), 0.48 (0.40–0.55), and 0.61 (0.54–0.68), respectively. Combining all four biomarkers yielded an AUC of 0.70 (0.65–0.74) in Train + Test and 0.61 (0.54–0.68) in LTP-2. Conclusions: In the first biomarker study reporting external validation in LTP-2, CYFRA 21-1, CEA, CA-125, and HE-4 demonstrated diagnostic value in IPNs. Impact: By leveraging commercial assays, this study highlights opportunities to enhance lung cancer risk stratification using widely available diagnostics that could be rapidly integrated into clinical workflows.