IGF::OT::IGF MACHINE LEARNING TO AUTOMATE CASE CONSOLIDATIONPOP: 09/19/2016 THROUGH 09/18/2017

University Of Utah, Salt Lake City UT

Investigators

Abstract

Cancer registry work continues to rely heavily on human decision-making to produce a completed record for each cancer case. The continuing need to add new, clinically-relevant required data items poses substantial challenges to the registry staff to learn and apply rules needed to produce highly reliable and accurate data. Application of information technology to cancer registration has the potential to address formidable issues associated with data collection by reducing the amount of human decision-making needed to produce a given number of completed records. The objectives of this study are: 1) Develop algorithms for consolidating cancer TNM stage information using machine learning, 2) Validate the performance of machine learning algorithms for TNM stage consolidation through comparison to the consolidation decisions made by cancer registrars.