Study
PLCO
(Learn more about this study)
Project ID
2012-0222
Initial CDAS Request Approval
Oct 23, 2012
Title
Validation of Biomarker Panel for Classification of Indeterminate Pulmonary Nodules
Summary
Support for low-dose helical computed tomography (CT) screening for lung cancer has recently emerged from the National Lung Screening Trial (NLST). CT screening was shown to confer a 20.3% reduction in lung cancer mortality in a high risk population, relative to screening with chest x-ray. However, concerns remain regarding the low specificity of CT scanning and the resulting cost and morbidity associated with biopsy or resection of benign pulmonary nodules. Non-invasive lung cancer biomarkers may serve as a useful complement to imaging, providing a simple, cost-effective means to further clarify the diagnosis of patients with suspicious pulmonary nodules. We have identified a panel of candidate lung cancer protein biomarkers through liquid chromatography / mass spectrometry analyses of resected lung tumor specimens (n=13), lung cancer cell lines (n=17), and conditioned media from tumor cell lines (n=7). Differential expression of makers was verified through ELISA analyses performed in serum collected from subjects with stage I/II NSCLC (n=56) and healthy smoker controls (n=99). Subsequently, a 6-marker logistic regression model was developed which accurately resolved lung cancer cases in an independent validation study, AUC =0.83 [NSCLC stage I/II (n=82); smoker and COPD controls (n=136)]. To evaluate the performance of the test in studies where sample volume is limiting, a multiplex version of the assay was developed on the Luminex platform. Conditions were established that enabled robust dose response while minimizing cross-reactivity between analytes. In an independent validation study [(NSCLC (n=40), healthy smoker controls (n=40)], the multiplex test demonstrated comparable accuracy (AUC=0.87) to ELISA-based assays (AUC = 0.89) in classifying lung cancer specimens. Moreover, the multiplex assay utilized a significantly lower serum volume (30uL) than traditional ELISA assays (410uL). We now intend to evaluate the accuracy of the multi-marker test in classifying lung cancer in subjects with suspicious pulmonary nodules.
Aims
Imaging (Chest X-ray and CT) can detect lung tumors, but distinguishing uncommon cancers from common benign conditions is difficult and leads to unnecessary procedures, anxiety, cost and radiation exposure. In the PLCO trial less than 2% of the ~16% of positive findings were malignant [1]. Similarly, in the CT arm of the National Lung Screening Trial (NLST), less than 4% of the solitary nodules that were identified (~24%) turned out to be cancerous [2]. Blood based biomarkers may serve as a useful complement to imaging, clarifying the diagnosis in patients with suspicious pulmonary nodules identified by radiologic imaging. We have developed a multi-marker test that accurately classifies lung cancer cases from smoker/COPD controls. We now aim to evaluate the performance of the multi-marker test in differentiating benign from malignant pulmonary nodules. If test successfully distinguishes nodules identified through Chest X-Ray (PLCO) we intend to evaluate accuracy in classifying study CT-detected nodules (through an application to NLST).
Collaborators
Robert J. Lagier (Celera)
Fabien Maldonado (Mayo Clinic)
Related Publications
