Establishment of a Multi-Analyte Serum Biomarker Panel to Identify Lymph Node Metastases in Non-small Cell Lung Cancer

Abstract

In non-small cell lung cancer (NSCLC), the presence of locoregional lymph node metastases remains the most important prognostic factor and significantly guides treatment regimens. Unfortunately, currently-available noninvasive staging modalities have limited accuracy. The objective of this study was to create a multianalyte blood test capable of discriminating a patient's true (pathologic) nodal status preoperatively. Pretreatment serum specimens collected from 107 NSCLC patients with localized disease were screened with 47 biomarkers implicated in disease presence or progression. Multivariate statistical algorithms were then used to identify the optimal combination of biomarkers for accurately discerning each patient's nodal status. We identified 15 candidate biomarkers that met our criteria for statistical relevance in discerning a patient's preoperative nodal status. A 'random forest' classification algorithm was used with these parameters to define a 6-analyte panel, consisting of macrophage inflammatory protein-1alpha, carcinoembryonic antigen, stem cell factor, tumor necrosis factor-receptor I, interferon-gamma, and tumor necrosis factor-alpha, that was the optimum combination of biomarkers for identifying a patient's pathologic nodal status. A Classification and Regression Tree analysis was then created with this panel that was capable of correctly classifying 88% of the patients tested, relative to the pathologic assessments. This value is in contrast to our observed 85% classification rate using conventional clinical methods. This study establishes a serum biomarker panel with efficacy in discerning preoperative nodal status. With further validation, this blood test may be useful for assessing nodal status (including occult disease) in NSCLC patients facing tumor resection therapy.