ParticipantsDenise Oldenburg, MD, Essen, Germany (Presenter) Nothing to Disclose
Sophronia Yu, San Francisco, CA (Abstract Co-Author) Nothing to Disclose
Jason K. Luong, San Francisco, CA (Abstract Co-Author) Nothing to Disclose
Philip Chu, San Francisco, CA (Abstract Co-Author) Nothing to Disclose
Yifei Wang, San Francisco, CA (Abstract Co-Author) Nothing to Disclose
Axel Wetter, Essen, Germany (Abstract Co-Author) Nothing to Disclose
Rebecca Smith-Bindman, MD, San Francisco, CA (Abstract Co-Author) Nothing to Disclose
Radiation doses for Computed Tomography (CT) examinations between patients, institutions, and countries are highly variable. Diagnostic reference levels (DRLs), and achievable doses (ADs) are often created to help reduce unnecessary variation. The European Society of Radiology has identified common indications for CT named EUCLID (European Study on Clinical DRLs) in order to create benchmarks for these examination types. We generated DRLs and ADs for these examinations.
METHOD AND MATERIALSStandardized data from > 2.3 million CT examinations in adults 18 years of age and older were collected between January 2016 and December 2018 from 155 institutions across 7 countries in a large, multinational CT Dose Registry. Two dose metrics were evaluated: CT-dose index (CTDIvol), and dose-length product (DLP).
RESULTSAD (50% in dose distribution) and DRL (75% in dose distribution) are summarized as follows (CTDIvol (mGy)/ DLP (mGy-cm), sample size n): chronic sinusitis (15 and 21 / 250 and 373, n= 57070), stroke to detect and exclude hemorrhage (47 and 53/ 872 and 1076, n= 14040), cervical spine trauma (19 and 30/ 450 and 962, n= 111397), pulmonary embolism (10 and 15/ 372 and 558, n= 112784), coronary calcium scoring (4 and 7/ 66 and 102, n= 22579), coronary angiography (21 and 31/ 497 and 915, n= 3176), lung cancer first and follow-up (11 and 15/ 556 and 858, n= 7064), hepatocellular carcinoma (9 and 14/ 1304 and 2016, n= 4289), colic/abdominal pain (10 and 14/ 519 and 773, n= 64724), and appendicitis/routine abdomen (11 and 16/661 and 1059, n= 721263). Most CT scans for clinical indications showed large differences in radiation dose compared to routine CT scans, e.g. sinusitis scans were >60% lower in both CTDIvol and DLP compared to routine head CT scans. Further, there were large differences in the DRLs and ADs across facilities.
CONCLUSIONDRLs and ADs for the clinical indications of EUCLID were presented and showed differences to routine CT scans. Dose metrics from large multi-center studies can help create representative DRLs and ADs that can be used for dose optimization, institutional evaluation, and indication-specific dose-optimized protocols.
CLINICAL RELEVANCE/APPLICATIONDRLs and ADs for clinical indications are essential due to high variation of CT radiation doses and for dose optimization, institutional evaluation, and indication-specific dose-optimized protocols.