Blood test could reduce unnecessary biopsies of lung nodules

A risk-assessment test that combined levels of plasma proteins and clinical factors accurately determined whether pulmonary nodules were benign or malignant, according to a prospective, multicenter observational trial.

Using this test in a clinical setting could result in 40% fewer biopsies on benign nodules, the Pulmonary Nodule Plasma Proteomic Classifier, or PANOPTIC, study showed.

In the National Lung Screening Trial, CT imaging identified concerning pulmonary nodules among 24% of screened patients, but only 4% of the nodules were ultimately determined to be malignant.

Nodules identified by chest CT scan require further diagnostic evaluation, which may be invasive and cause patients anxiety.

According to Gerard A. Silvestri, MD, MS, FCCP, lung cancer pulmonologist at Medical University of South Carolina, and colleagues, patients with intermediate-risk nodules may benefit from risk stratification tools that identify those in need of more aggressive evaluation.

Silvestri and colleagues assessed the sensitivity, specificity and negative predictive value of BDX-XL2 (Biodesix) — a pulmonary nodule plasma proteomic integrated classifier — to determine its feasibility to potentially avoid invasive biopsies of lung nodules.

The integrated classifier used multiple reaction monitoring mass spectrometry to assess the relative abundance of proteins LG3BP and C163A in combination with an assessment of clinical risk factors — including age; smoking status; and nodule diameter, shape and location — to identify likely benign nodules.

The analysis included 178 patients (median age, 65.52 years; 53.37% men) with 8 mm to 30 mm lung nodules who had physician-assessed pretest probability of cancer of 50% or less.

Among this cohort, researchers reported a 16% prevalence of lung cancer.

The integrated classifier demonstrated a sensitivity of 97% (CI, 82-100) in distinguishing benign nodules from malignant nodules.

Researchers also found the test had a specificity of 44% (CI, 36-52) and a negative predictive value of 98% (CI, 92-100).

With an area under the curve of 0.76, the integrated classifier performed better than PET (0.58), validated lung nodule risk models (VA model, 0.6; Mayo Clinic model, 0.69) and physician cancer probability estimates (0.69; P < .001).

Silvestri and colleagues also estimated that if used to direct care, the integrated classifier would lead to 40% fewer procedures performed on benign nodules, although 3% of malignant nodules would be misclassified as benign.

The limitations of the study included the retrospective nature of the care impact assessment, underrepresentation of community practices, and use of 1-year — not 2-year — outcomes to determine nodule status.

In an accompanying editorial, Nawar Al Nasrallah, MD, and Catherine R. Sears, MD, of the Indiana University School of Medicine, discussed the integrated classifier’s strengths and potential.

“Advantages of this integrated classifier are that the serum biomarkers are easily obtained, the clinical characteristics are readily available, and it could be combined with other risk stratification characteristics available now and in the future,” they wrote. “This trial represents an important step in development of a molecular profile that will aid in classification of intermediate-risk nodules and hopefully avoid unnecessary procedures, anxiety and costs.” – by Cassie Homer

Disclosures: Silvestri reports research grants from Exact Sciences, Integrated Diagnostics and Olympus. Please see the study for all other authors’ relevant financial disclosures. Nasrallah and Sears report no relevant financial disclosures.