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.
McGann et al. (Wed,) studied this question.