TL;DR
This paper introduces an automated system for managing and monitoring CT scanner protocols, addressing the labor-intensive and error-prone nature of manual reviews. The system successfully identified a significant number of inappropriate protocol modifications, highlighting the need for such automated quality control to ensure patient safety and consistent image quality.
What Problem Does This Paper Address?
In modern medical imaging, particularly with Computed Tomography (CT), a large number of protocols are used to optimize image acquisition for various diagnostic tasks and scanner models. Manually reviewing these protocols is a monumental task, prone to human error, and costly. This manual process is crucial because incorrect protocol parameters can lead to increased patient radiation dose, potentially causing injury, or result in suboptimal image quality, compromising clinical diagnoses. Regulatory bodies like the ACR and Joint Commission mandate protocol reviews, and some states even require them. However, the sheer volume of protocols, continuous optimization efforts, and the potential for accidental or unauthorized changes make manual oversight increasingly challenging and inefficient. Furthermore, differences in scanner models and manufacturers add another layer of complexity to maintaining protocol consistency across an institution.
Methods
The authors developed a software system to automate the CT protocol management and monitoring workflow. The system's core function is to verify that protocols programmed on CT scanners match a defined set of "master protocols." The workflow involves several key steps:
- Master Protocol Maintenance: Establishing and maintaining an accurate electronic repository of institutional master protocols.
- Protocol Retrieval: Automatically or manually retrieving current protocols from the CT scanners.
- Protocol Comparison: Comparing the retrieved scanner protocols against the master protocols.
- Discrepancy Review: Displaying flagged differences between scanner and master protocols for human review.
- Protocol Updating: Updating either the scanner or master protocols based on the review decisions.
The software was primarily developed using MATLAB and Excel. The comparison step involves matching scanner protocols to master protocols, ideally through a unique identifier system, or by relying on protocol names, which can be prone to errors due to variations in spacing or capitalization. The system monitors a defined set of acquisition and reconstruction parameters deemed clinically relevant and user-adjustable.
Datasets
The study evaluated the system using abdomen and pelvis protocols from two Siemens Definition Flash scanners over a 24-week period. This evaluation focused on both single- and dual-energy, adult, and pediatric protocols. The "dataset" in this context refers to the actual CT scanner protocols and the master protocols maintained by the institution. No new benchmark datasets were introduced.
Key Insights and Ideas
- High Frequency of Protocol Modifications: CT scanner protocols are modified much more frequently than one might assume, with hundreds of modifications occurring over a relatively short period.
- Prevalence of Inappropriate Modifications: A significant portion of these modifications (around 14% in this study) were deemed inappropriate, meaning they were not intentional or were incorrectly implemented. This underscores the critical need for automated oversight.
- Series Description as a Common Error Point: The "series description" parameter was the most frequently inappropriately modified parameter. This is significant because it plays a crucial role in image organization and hanging protocols for radiologists.
- Impact on Radiation Dose and Image Quality: While many inappropriate modifications were to less critical parameters, the study flagged two instances of inappropriate changes to tube current, a parameter directly impacting radiation dose and image quality. This highlights the system's potential to catch critical safety issues.
- Need for a Centralized Master Protocol: The pilot study revealed limitations in using baseline scanner protocols as master protocols. A centralized, institution-wide master protocol system is essential for ensuring consistency across all scanners.
Key Results
During a 24-week pilot study on two scanners, the system detected 309 modified protocols and a total of 712 protocol parameter modifications. Of these, 614 (86%) were appropriate modifications, reflecting ongoing optimization efforts. However, 98 (14%) were identified as inappropriate modifications. The most common inappropriate modification was to the series description (47 times). Two inappropriate modifications were to the tube current. The majority of appropriate modifications were to reconstruction parameters (83%), while acquisition parameters accounted for 17%.
Limitations and Future Work
- Manual Review Still Necessary: While the system automates comparison and flagging, the decision-making process for determining whether a discrepancy is appropriate or inappropriate still requires human review by protocol experts.
- Limited Scope of Pilot Study: The pilot study focused only on abdomen and pelvis protocols and a specific scanner model.
- Lack of Reason Analysis: The study did not investigate the underlying reasons for the decisions made during the review of flagged differences.
- Future Work: The authors suggest that the system could be expanded to include other body regions and scanner models, and that further development of the DICOM standard for CT protocols could enable fully automated remote updating of scanner protocols. They also propose that a more detailed analysis of the reasons behind protocol modification decisions could further optimize the review process.
Summary updated