Dose-Tracking Software: A Retrospective Analysis of Dosimetric Data in CT Procedures

Abstract

The increasing use of ionizing radiation in healthcare is causing growing alarm about radiation protection of patients and the doses they receive during procedures. Radiation dose assessment for patients in radiodiagnostic procedures is the subject of interest in view of the recent Italian D.Lgs 31 July 2020, n. 101 (Decreto Legislativo 31 luglio 2020, n. 101) and one of its most important focuses is the prescription to provide patient exposure information as an integral part of the examination report. Dose monitoring systems are therefore essential for the collection of the dosimetric data. In order to analyse potential and critical issues of these software, different systems, adopted at the Antonio Cardarelli Hospital in Naples, were employed. Data extracted from the DoseWatch software (GE Healthcare) and Gray Detector (EL.CO. S.r.l. Healthcare Solutions, Italy) and relating to several protocols adopted for computed tomography (CT), were retrospectively analysed for the purpose of identifying critical issues in the data acquisition and recording phase, comparing with Italian nationwide diagnostic reference levels (DRLs), as provided for in regulatory provisions for radiation safety. Multiphase examinations were also included in this study. Once the distributions of volumetric CT Dose Index (CTDIvol) and dose-length product (DLP) were determined for each acquisition phase and total DLP (DLPtot) for each examination, the 25th, 50th and 75th percentiles were calculated for each distribution and then compared with the relevant Italian nationwide DRLs. In addition, to improve protocol optimization and dose reduction the magnitude of the CT acquisition settings chosen in each procedure was evaluated. In conclusion, these systems allow accurate analysis of radiation dose according to equipment and protocol over time. For the application of optimization measures, a constant use of the dose tracking software is required, which can be translated into actions on scan parameters and prospective data analysis.