[P025] The use of radlex playbook to manage CT big data: An italian multicentre study

• DOI: 10.1016/j.ejmp.2018.06.353
• Publication Date: 2018-08-19T22:29:23Z
• Authors: Francesca Calderoni; Cristina De Mattia; Stefania Nici; Federica Rottoli; Marina Sutto; Mauro Campoleoni; Paola Enrica Colombo; Alberto Torresin
• Keywords: 03 medical and health sciences; 0302 clinical medicine; 3. Good health

Purpose To manage and analyze dosimetric data provided by 12 different CT scanners, installed in four Italian hospitals, through the standardization of RadLex Playbook. Methods The interest in CT dose management has been growing during last years because of the increasing awareness of risks related to high radiation dose exams. A task of the Italian project PREP (Procedure Radiodiagnostiche in Eta Pediatrica), is to create a dosimetric database analyzing exposure values from the CT installed in 4 Italian different hospitals. These data were collected in a cloud-based database provided by NEXO[DOSE]; a Radiation Dose Index Monitoring software (Bracco, Italy). Until now, about 300,000 CT exams were collected: in 2017, the exams were about 76,000. For each exam, NEXO[DOSE] reports patients’ demographic information, scan protocol values (as total DLP and CTDI vol ) and single phase parameters. When it is necessary to compare procedures from different devices/PACS systems, some problems arise from the different names used for the same procedure. It is thus indispensable to cluster them in homogeneous sets according to scan region and acquisition type (scan with or without intravenous contrast). For each single device we selected the DICOM tag which was more appropriate to describe the procedure and we associated this description to a label from RadLex playbook, a standard system for naming radiology procedures written by RSNA. We used 100 RadLex lables, including almost 600 procedures. Results 84% of all the CT exams are included in 10 labels. Since each label has homogenous exposure data, a risk assessment for the whole group can be performed. Preliminary results show that single-phase head CT is the most common exam (38%) followed by single-phase chest (9%), multiphase abdomen-pelvis (9%), and multiphase chest-abdomen-pelvis (9%). Conclusions RadLex is a useful tool to cluster exams from different hospitals and devices. This can be helpful for statistical and dosimetric purposes. Our future applications will be to use RadLex also for other modalities (Magnetic Resonance, Mammography, Angiographic procedures).