SEER 2015-04; Expanding SEER to Include Molecular Profiling in Non-Small Cell Lung Cancer (NSCLC)

Title: Expanding SEER to Include Molecular Profiling in Non-Small Cell Lung Cancer (NSCLC) Principal Investigator: Eric B. Durbin, DrPH, MS Abstract This study addresses a current deficiency in the lack of tumor molecular profile data (TMPD) on Non-Small Cell Lung Cancer (NSCLC) in SEER registries. Specifically, the study seeks to leverage electronic pathology (ePath) reports to automate the annotation of NSCLC lung cases with results from epidermal growth factor receptor (EGFR) and Anaplastic Lymphoma Kinas (ALK) when available. The specific objectives of this study are to 1) develop a Natural Language Processing (NLP) algorithm to process ePath reports to ascertain the use and results of EGFR and ALK testing in stage IV non-squamous NSCLC cases included in SEER registries; and 2) conduct a multiphase validation study of the NLP algorithm initially involving cases included in the Kentucky SEER registry, and posteriorly validating the algorithm in other SEER Contractor registries. A specified cohort of eligible cases will be identified from each registry followed by the generation of a gold standard set of EGFR and ALK results through certified tumor registrar review of each case. The NLP method specific to these biomarkers will be developed based upon prior NLP investigations by the research team. The final NLP algorithm will be tested and validated as measured by precision, recall, sensitivity, specificity, and F-Score.