Purpose: The Lung CT Screening Reporting and Data System (Lung-RADS™) is an algorithm that can be used to classify lung nodules in patients with significant smoking histories. It is published in table format but can be implemented as a computer program. The aim of this study was to assess the efficiency and accuracy of the use of a computer program versus the table in categorizing lung nodules.
Methods: The Lung-RADS algorithm was implemented as a computer program. Through the use of a survey tool, respondents were asked to categorize 13 simulated lung nodules using the computer program and the Lung-RADS table as published. Data were gathered regarding time to completion, accuracy of each nodule's categorization, users' subjective categorization confidence, and users' perceived efficiency using each method.
Results: The use of a computer program to categorize lung nodules resulted in significantly increased interpretation speed (80.8 ± 37.7 vs 156 ± 105 seconds, P < .0001), lung nodule classification accuracy (99.6% vs 76.5%, P < .0001), and perceived confidence and efficiency compared with the use of the table. There were no significant differences in accuracy when comparing thoracic radiologists with the remainder of the group.
Conclusions: Radiologists were both more efficient and more accurate in lung nodule categorization when using computerized decision support tools. The authors propose that other institutions use computerized implementations of Lung-RADS in the interests of both efficiency and patient outcomes through proper management. Furthermore, they suggest the ACR design future iterations of the Lung-RADS algorithm with computerized decision support in mind.
Keywords: Lung-RADS; decision support; lung cancer screening.
Copyright © 2015 American College of Radiology. Published by Elsevier Inc. All rights reserved.