A panel of 10 serum biomarkers for lung cancer could offer more accurate interpretation of nodules detected on computed tomography, avoiding invasive biopsies and radiographic follow-up.

"CT-screening detection of an indeterminate pulmonary nodule, a nonspecific but frequent finding in high-risk subjects with a smoking history, creates a diagnostic dilemma," wrote investigator William L. Bigbee, Ph.D., and his colleagues in the April issue of the Journal of Thoracic Oncology.

"Although the biomarker model we described could not detect every lung cancer, it offers a significant clinical improvement over CT imaging alone ... Also, patients with nodules not identified as cancer by the model would continue to receive follow up clinical monitoring and would be biopsied if the nodules grew in size, which is the current standard of care," (J. Thorac. Oncol. 2012;7:698-708).

Dr. Bigbee of the University of Pittsburgh and his colleagues cite results of the National Lung Screening Trial (NLST), published in June 2011, which showed for the first time that low-dose CT screening of heavy smokers could reduce lung cancer mortality by 20%. But, as Dr. Bigbee et al. note in the current study, the "vast majority" of positive results in the NLST program turned out to be false after diagnostic evaluation. Moreover, smaller nodules are least likely to be malignant and least likely to be considered for biopsy or surgery.

For the current study, the researchers initially looked at a "training" set of 56 patients with non–small cell lung cancer in the University of Pittsburgh Cancer Institute Georgia Cooper Lung Research Registry. These cases were matched with 56 controls from the Pittsburgh Lung Screening Study (PLuSS), a volunteer cohort at high risk for lung cancer. All controls were known to be cancer free. The authors then analyzed serum samples from both groups for the presence of 70 potential cancer-associated biomarkers.

"Together, these biomarkers incorporate a wide range of host and tumor derived factors that allow a broad analysis of the lung cancer/host interaction, and includes a number of previously described epithelial cell cancer-associated serological markers," wrote the investigators. "The initial goal of this discovery study was to identify the most robust subset of these biomarkers to discriminate lung cancer and matched control samples."

The researchers, using a rule-learning algorithm, whittled the field of potential biomarkers down to eight: prolactin, transthyretin, thrombospondin-1, E-selectin, C-C motif chemokine 5, macrophage migration inhibitory factor, plasminogen activator inhibitor 1, and receptor tyrosine-protein kinase erbB-2.

"This rule model distinguished the lung cancer case samples from the control samples in the training set with a sensitivity of 92.9% and specificity of 87.5%," they reported.

Ultimately, two additional biomarkers were added to the panel – cytokeratin fragment 19-9 and serum amyloid A protein – and an additional set of cases and controls, 30 in each cohort, was assessed, in a blinded "verification" set.

In this set, the authors calculated an overall classification performance of 73.3% sensitivity and 93.3% specificity. Only 10 misclassifications occurred among 60 predictions made. Moreover, when looking at accuracy according to patient demographic factors, the researchers found that the 10-biomarker panel was equally good at distinguishing males and females as either cases or controls and that neither current smoking status nor airway obstruction skewed the results.

"Age overall was not a significant factor in misclassification of cases or controls, although two of three cases aged 38-44 [years] were misclassified as controls by the 10-biomarker model," the authors concede. "This inaccuracy may result from the absence of younger subjects in the training set that included no cases younger than 46 years at diagnosis and no controls younger than 50 years."

Nor did the presence of nodules visible on CT scan confound the biomarkers’ predictive value. "In fact, those PLuSS subjects with a suspicious nodule were more often correctly classified as controls than those with no nodule or a benign nodule," wrote the authors.

They add that all nodules found in controls remained clinically noncancerous at least 3 years after initial detection, with either resolution or no further growth on subsequent CT scans.

Finally, Dr. Bigbee assessed the model’s accuracy when confronted with early- vs. late-stage tumors.

"Among stage I/II lung tumors, the 10-biomarker panel misclassified 15% of stage I/II tumors in the verification set, compared to 50% of the stage III/IV tumors, suggesting the model performs well in discriminating early-stage lung cancer," he wrote. "With a specificity of 93.3%, the 10-biomarker model [balanced accuracy] was 89.2% in stage I/II disease."

The authors conceded that the biomarker panel presented here would not suffice for general population screening. However, in a clinical context, among high-risk patients, the model "may provide clinical utility in guiding interpretation of screening CT scans, even in tobacco-exposed persons with COPD or emphysema," they wrote.

"Formal validation in larger patient cohorts will be needed to confirm these initial findings."

The authors disclosed that funding for this study was supplied by grants from the National Cancer Institute. Dr. Bigbee stated that there were no personal disclosures.

Ten Biomarkers May Aid Lung Cancer Detection