Author: Joke Binst, Hilde TC Bosmans, Niki Fitousi, Bram Miseur, Dimitar Petrov, Kwinten Torfs, Janne Vignero 👨🔬
Affiliation: Qaelum NV, Department of radiology, UZLeuven, Department of imaging and pathology, KULeuven 🌍
Purpose: Routine quality monitoring in CT abdomen is crucial for ensuring optimal image quality and patient safety, but the simultaneous evaluation of all aspects of importance is challenging. The purpose of this study is to use big data for the creation of protocol specific reference levels that visualize - in a single summary plot - the impact of CTDIvol, global noise level (GNL), water equivalent diameter (WED) and kVp, to facilitate and enable scanner comparisons. The initial focus was on venous phase abdominal scans.
Methods: Abdomen venous CT data of 46000 exams across six CT scanners, characterized by CTDI, GNL, WED and kVp, were collected via a dose management platform. For each venous phase acquisition protocol, . Reference levels and standard deviations of CTDI and GNL were calculated as a function of WED for different kVp settings using a running average approach. The data was visualized in a comprehensive summary plot for each protocol. Additionally, a comparison of CT scanners was conducted for each kVp.
Results: All parameters were successfully and meaningfully integrated in a single summary plot. The protocol summary plots demonstrate the influence of kVp on the reference GNL, reducing it up to 70% going from 80 to 140 kVp. The detection of suboptimal scans was improved as the patient size was considered to identify noise or/and dose outliers. Reasons for outliers included incorrect table positioning, arms in the beam and blind scanning. Comparing scanners illustrated the impact of differences in kV/mA modulation settings.
Conclusion:Comprehensive CT protocol quality management is possible from metrics provided by a dose management system. This enables more sensitive detection of outliers by factoring in patient size, kVp, and protocol settings when benchmarking quality and dose.