P. Esquinas, Yen-Fu Luo, P. Farzam, Tyler Baldwin, Moshe Raboh, T. Binder, Arkadiusz Sitek, O. Sakhi, Yi-Qing Wang, Sameer Suman, G. Palma, Paul Dufort, Benedikt Graf
{"title":"Evaluation of an Automated Method to Detect Missed Focal Liver Findings In Single-Phase CT Images of The Abdomen","authors":"P. Esquinas, Yen-Fu Luo, P. Farzam, Tyler Baldwin, Moshe Raboh, T. Binder, Arkadiusz Sitek, O. Sakhi, Yi-Qing Wang, Sameer Suman, G. Palma, Paul Dufort, Benedikt Graf","doi":"10.1109/ISBI52829.2022.9761517","DOIUrl":null,"url":null,"abstract":"In the present study, an automated method to identify potential missed focal liver lesions in abdominal CT scans is described and evaluated. The method analyzes radiology reports and DICOM data via natural language processing and deep-learning based imaging algorithms, respectively, aiming to detect and classify liver lesions in studies where the original radiologist found no evidence of them. The proposed approach was evaluated on a cohort of 13500 contrast-enhanced abdominal CT studies and yielded a total of 25 potential missed liver lesions which were subsequently reviewed by 5 independent radiologists. On average, 48.8% of studies flagged by the method contained actual liver lesions not reported by the original radiologist, and 15.2% of all findings were deemed to be clinically significant. The proposed method could be a valuable tool to inform radiologists of potential missed focal liver lesions.","PeriodicalId":6827,"journal":{"name":"2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI)","volume":"2015 1","pages":"1-5"},"PeriodicalIF":0.0000,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISBI52829.2022.9761517","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
In the present study, an automated method to identify potential missed focal liver lesions in abdominal CT scans is described and evaluated. The method analyzes radiology reports and DICOM data via natural language processing and deep-learning based imaging algorithms, respectively, aiming to detect and classify liver lesions in studies where the original radiologist found no evidence of them. The proposed approach was evaluated on a cohort of 13500 contrast-enhanced abdominal CT studies and yielded a total of 25 potential missed liver lesions which were subsequently reviewed by 5 independent radiologists. On average, 48.8% of studies flagged by the method contained actual liver lesions not reported by the original radiologist, and 15.2% of all findings were deemed to be clinically significant. The proposed method could be a valuable tool to inform radiologists of potential missed focal liver lesions.