Abstract Introduction Lung cancer screening (LCS) via low-dose CT (LDCT) identifies a high rate of non-lung cancer incidental findings (IFs) ranging from benign to severe. Although the Lung Computed Tomography Screening Reporting and Data System (Lung-RADS) recommends the use of an S-modifier for IFs not reported on prior scans, it has been applied inconsistently. We sought to better understand radiologists’ practice paradigm and reasoning towards reporting and recommendations for IFs on LDCT. Methods A web-based survey was iteratively developed with input from the Veterans Affairs (VA) Lung Precision Oncology Lung Cancer Screening Working Group and VA radiologists. It was designed using RedCap and distributed to radiologists across the VA. The survey included three clinical vignettes about cardiopulmonary IFs and open-ended questions to evaluate radiologist perspectives and determine how artificial intelligence (AI) could be developed to support workflow. Results A total of 71 radiologists accessed the survey and after excluding incomplete responses and those that do not read LDCTs, 55 radiologists spanning 14 of 18 Veterans Integrated Service Networks provided responses for analysis. Most were over 10 years out of training (80%) and had additional subspecialty training (69%), and 20% reported thoracic subspecialty training. Most reported at least moderate familiarity with the ACR Quick Reference Guide for Incidental Findings on Lung Cancer Screening (73%) and reported referencing it either weekly (26%) or monthly (26%). Radiologists reported using the S-modifier depending on the level of severity for coronary artery calcification (65%) and pulmonary fibrosis (54%), or at any level (13% and 24%, respectively). In contrast, a large proportion of radiologists reported never using the S-modifier for emphysema (46%), see figure. Respondents felt reporting IFs benefits patients most of the time (61.2%) and were concerned about the workload for primary care providers (40.8%). When asked how AI could assist radiologists with reporting IFs, the most common responses were to generate follow-up recommendations (75.5%), generate a structured report (73.5%), and review the electronic medical record to determine which IFs were previously diagnosed (75.5%). Conclusions Responding radiologists endorse the use of the S-modifier depending on severity for coronary artery calcification and pulmonary fibrosis but show less agreement when it comes to reporting emphysema. Concerns include uncertain patient benefit from reporting of IFs and the workload for primary care providers. Future work is needed to improve the standardization and efficiency of reporting significant IFs on LDCT. This abstract is funded by: None
O’Laughlin et al. (Fri,) studied this question.