Read Part II here.
Time is of the essence in treating acute stroke victims. Streamlined procedures can shave critical minutes from diagnostic evaluation. Radiologists and medical physicists at the University of Wisconsin-Madison analyzed the speed of computed tomography (CT) imaging of stroke patients to identify clinical variables and areas where improvements could be made.
Their analysis was published online in February in the Journal of the American College of Radiology. A related article in the April issue of the Journal of Digital Imaging described the detailed methodological framework the authors used to define and monitor the CT imaging acquisition workflow.
Medical physicist Timothy P. Szczykutowicz, PhD, and colleagues analyzed exam data of 494 patients who underwent a comprehensive CT stroke imaging exam over 18 months at one of four CT scanners, three at the main campus and one at a satellite facility.
The hospital’s protocol for acute stroke CT consists of a localizer radiograph plus four series. These include noncontrast head, CT angiography (CTA), and perfusion, and with-contrast head. The protocol for the technologist performing the scan is to send each series to PACS as soon as images are reconstructed to expedite interpretation by emergency radiologists. Upon completion of the perfusion scans, data are sent to automated software or to a workstation for processing.
The authors collected data that identified the scanner, date and time of the exam, name of the technologist, and the perfusion processing method. Time stamps from DICOM image headers and the PACS database were used to measure workflow, specifically the length of time for series setup, acquisition for each stage, and to transfer acquired images to the PACS for interpretation.
The analysis revealed that performance was consistent regardless of the time or day of the week that stroke CT imaging was performed. However, at the satellite facility, CT scans were acquired and processed more slowly than at the main campus. Comparisons of anonymized CT technologists showed statistically significant differences in acquisition times (4.75-15.28 minutes) and median times for full examination availability in PACS (15.4-51.8 minutes). More experienced technologists were more efficient. The data showed improvements could be made in technologist training and that an investigation of the satellite facility CT scanner was merited.
A policy change also showed a dramatic increase in efficiency. The change required technologists to proceed directly to CTA after acquiring a noncontrast scan. Previously, they waited for the attending neurologist to exclude hemorrhage on the head noncontrast examination before proceeding to CTA. Implementing this policy decreased the time from noncontrast head scan acquisition to CTA acquisition from 7.18 minutes to 2.46 minutes.
Automated perfusion processing software also improved efficiency. During the analysis, manual perfusion processing software was used initially, requiring technologists to process the perfusion images on a separate workstation. When this was replaced by an automated system, mean processing time fell from 14.6 minutes to 11.7 minutes.
“We found our median time for acquiring a comprehensive acute stroke CT exam varied by technologist from less than 5 to over 15 minutes,” lead author Christina Brunnquell, MD, told Applied Radiology. “This result confirms that we are providing rapid stroke response, but also shows that we have a huge opportunity to share best practices among technologists.”
Time efficiency analysis for CT imaging of acute . Appl Radiol.