Combining Commercial Cancer Protein Biomarkers and Benign Fungal Antibodies Improves Diagnostic Accuracy in Pulmonary Nodules.

Abstract

To evaluate whether adding cancer and histoplasma biomarkers to existing clinical prediction models improves diagnostic accuracy for malignancy in indeterminate pulmonary nodules (IPNs). IPNs remain a diagnostic challenge, particularly in regions where granulomatous disease exists. Clinical prediction models such as the Mayo model perform moderately well but only use clinical and radiographical information. Blood-based biomarkers for cancer proteins and fungal exposure may enhance diagnostic discrimination. Two geographically distinct cohorts with 6-30 mm IPNs were studied from the Ohio River Valley (ORV) and the Mountain West (MW). Mayo scores, serum levels of four cancer proteins (CYFRA 21‑1, CEA, CA‑125, and HE‑4), and histoplasma IgG and IgM antibodies were measured. Area under the receiver operating curve (AUC) and 95% confidence intervals were calculated for individual components and combinations. Multivariable logistic regression models were built for pairwise combinations and for all variables together (full model). Risk reclassification was assessed in each cohort. We studied 366 patients: 282 from the ORV cohort and 84 from the MW cohort. Mayo model AUCs were 0.74 (0.67-0.80) and 0.68 (0.59-0.78), respectively. Combining cancer biomarkers with fungal antibodies yielded AUCs of 0.74 (0.68-0.80) and 0.76 (0.64-0.86). The full model achieved AUCs of 0.80 (0.75-0.87) and 0.79 (0.70-0.88). In intermediate-risk nodules, the full model correctly upgraded malignancy risk in 46.3% and 42.9% of patients in each cohort. Integrating cancer biomarkers, fungal antibodies, and clinical prediction tools improves diagnostic performance for IPNs across distinct geographic regions. A combined biomarker-clinical model offers a feasible strategy to better differentiate benign from malignant nodules.