Back to 2026 Abstracts
Combining Commercial Cancer Protein Biomarkers and Benign Fungal Antibodies Improves Diagnostic Accuracy in Indeterminate Pulmonary Nodules
*Hudson M. Holmes
1, *Kevin C. McGann
1, *Sheau-Chiann Chen
2, *Cole Welch
1, *Yong Zhou
1, *Sanja Antic
3, *Yency Forero
3, *Samson Argaw
3, *Heidi Chen
2, *Alex Kaizer
4, *Anna Baron
5, *Robert M. Meguid
6, *Fabien Maldonado
3, *Stephen Deppen
1,
Eric Grogan11Department of Thoracic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee; 2Department of Biostatistics and Epidemiology, Vanderbilt University Medical Center, Nashville, Tennessee; 3Department of Pulmonology, Vanderbilt University Medical Center, Nashville, Tennessee; 4Department of Biostatistics and Informatics, University of Colorado Anschutz Medical Campus, Aurora, Colorado; 5Department of Public Health, University of Colorado Anschutz Medical Campus, Aurora, Colorado; 6Department of Surgery, University of Colorado Anschutz Medical Campus, Aurora, Colorado
Objective(s): To evaluate whether combining commercial cancer protein biomarkers and benign fungal antibodies improves diagnostic accuracy in indeterminate pulmonary nodules. We sought to determine whether biomarkers in isolation and with clinical (age, smoking, cancer history) and radiographic variables (nodule size, shape, and location) in the Mayo Clinic Model improve diagnostic accuracy in regions with historically high and low rates of histoplasmosis.
Methods: Patients from two geographically distinct cohorts (Ohio River Valley [ORV] and Mountain West [MW] regions) with indeterminate pulmonary nodules (IPNs) 6-30mm in diameter were included for retrospective analysis. All patients had Mayo Clinic model variables, and serum measurements of four cancer proteins (CYFRA 21-1, CEA, CA-125, and HE-4) and histoplasma immunoglobulins G (IgG) and M (IgM) available. Serum samples were collected prospectively and bio-banked at -80C in the Thoracic Biorepository. Subsequent assays were performed on commercial platforms (Roche Cobas e411, MiraVista ELISA) in CAP-certified laboratory environments. All variables were treated as continuous data: cancer biomarker concentrations were measured in ng/mL, fungal antibody levels were reported in optical density, and Mayo scores were given as probability of malignancy. The area under the receiver operating characteristics curve (AUC) was calculated for the cancer biomarkers, fungal biomarkers, and their combination based on final diagnosis. Diagnosis was adjudicated by either tissue pathology confirming benign or malignant disease or clear documentation of at least two years of stable radiologic follow-up consistent with benign disease. Multivariable logistic regression models were fitted for each pairwise combination and the combination model including all variables (full model).
Results: A total of 366 patients were included. The ORV cohort comprised 282 patients, and the MW cohort contained 84 patients. The AUC of the cancer biomarkers alone was 0.71 (95% CI, 0.64-0.76) in the ORV cohort and 0.77 (95% CI, 0.66-0.88) in the MW cohort. The AUC of the fungal antibodies alone was 0.61 (95% CI, 0.55-0.69) and 0.63 (95% CI, 0.55-0.74) in these cohorts, respectively. The combination of cancer biomarkers and fungal antibodies had an AUC of 0.74 (95% CI, 0.68-0.79) and 0.78 (95% CI, 0.66-0.88). The Mayo Model alone had an AUC of 0.74 (95% CI, 0.67-0.80) and 0.67 (95% CI, 0.59-0.77). The full combination model had an AUC of 0.79 (95% CI, 0.74-0.85) and 0.79 (95% CI, 0.71-0.88), respectively.
Conclusions: The combination of commercial cancer protein biomarkers and fungal antibodies improves diagnostic accuracy in IPNs in both the Ohio River Valley and Mountain West cohorts. A full combination model integrating cancer protein biomarkers, fungal antibodies, and an established clinical prediction tool provides a feasible approach to better discriminate benign from malignant nodules.
Back to 2026 Abstracts
