Automated tool to measure lung cancer risk factors in pulmonary nodule patients

Half of all people undergoing a computed tomography (CT) of the chest will have an incidental finding of a
pulmonary nodule. The incidence of lung nodules is expected to be rising because the use of CT is rapidly
increasing. Decisions surrounding the diagnostic evaluation and follow-up of pulmonary nodules are
complicated by a lack of high-level evidence supporting any one strategy, the influence of potentially
uninformed patient preferences, and clinician fear of litigation. Practice-guidelines attempt to mitigate this
complexity by recommending more or less aggressive evaluation and follow-up depending on the individual’s
risk of having lung cancer. Examples of lung cancer risk factors include tobacco exposure and radiographic
nodule characteristics. Unfortunately, there is no standard approach to lung cancer risk-assessment.
Moreover, it is unclear to what extent providers recognize the importance of ascertaining these risk factors
and how they incorporate this information into decision-making. There remains a fundamental knowledge
gap about how lung nodules are actually managed in the community-at-large. A significant barrier to
investigating nodule care is that there is no reliable method to identify a population of individuals with
pulmonary nodules and describe their risk for lung cancer. Investigators recently developed and validated an
automated tool utilizing administrative data, radiology reports, and natural language processing (NLP) to
identify enrollees with a lung nodule. This work is an important step forward in ultimately conducting multisite
investigations of care-delivery and outcomes. A logical next step in advancing this line of research is to
develop and validate an automated tool to identify a cohort with pulmonary nodules and measure
documented lung cancer risk factors. This approach assumes that clinicians document relevant lung cancer
risk factors—an assumption that has never been evaluated. Knowledge of commonly documented lung
cancer risk factors will inform the development of such a tool but is expected to also lead to interventions to
improve the delivery of nodule care.

<< Back to the list of CRN projects