Clean sweep: Data-driven organization clears the cobwebs from CT scanner protocol lists

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - CT Scanner with patient

Reviewing CT protocols can be a time-consuming process, but researchers from Massachusetts General Hospital were able to use a data-driven process to lighten the load.

“Protocol review plays a critical role in CT quality assurance, but large numbers of protocols and inconsistent protocol names on scanners and in exam records make thorough protocol review formidable,” wrote the research team, including lead author Da Zhang, PhD, in a study published online Jan. 7 in the Journal of the American College of Radiology.

For large organizations with multiple scanners, reviewing all protocols is a daunting task due to the higher number of differing CT protocols, and the authors stated that having hundreds of protocols loaded on a single CT scanner is not uncommon.

“In this article, we report a data-driven process to catalog actively used protocols and to conduct housekeeping on the protocol pool, in the hope of alleviating the difficulty of protocol review,” Zhang and colleagues wrote.

The researchers collected scanner protocol and 18 months’ worth of exam records for 10 CT scanners. For the study, they developed algorithms that deconstructed protocol names on the scanner and within exam reports into core names.

Based on the core names developed by the algorithms, the team grouped the protocols into much smaller sets and linked exam records to the scanner protocols. Additionally, they calculated the most heavily used protocols in the scanners.

The data showed that, on average, 18, 33 and 49 core protocols per scanner covered 80, 90 and 95 percent, respectively, of all exams. The team also reported that duplicated and rarely used protocols were easily identified in the cataloging process.

While replicating the process at other organizations would likely yield similar results, the authors cautioned that the methods would need to be tuned to the naming conventions of individual institutions.

“Grouping similar scanner protocols into core protocols through decomposition, as well as calculating the frequency of usage for various core protocols, would help alleviate the difficulty of the protocol review and management process,” Zhang and colleagues wrote.