Is a weekly qualitative picket fence test sufficient? A proposed alternate EPID‐based weekly MLC QA program

Chaitanya Kalavagunta, Huijun Xu, Baoshe Zhang, Sina Mossahebi, Michael MacFarlane, Kai Jiang, Sung‐Woo Lee, Shifeng Chen, Amit Sawant, Arun Gopal, B YiJournal of Applied Clinical Medical Physics 2022 · DOI

TL;DR

This paper proposes a more comprehensive weekly quality assurance (QA) program for multileaf collimators (MLCs) in radiation therapy, moving beyond the standard qualitative picket fence test. By analyzing over 20,000 QA images, they demonstrate that their proposed program significantly reduces the risk of MLC failures, as quantified by a Failure Mode and Effects Analysis (FMEA), making it essential for ensuring accurate radiation delivery.

What Problem Does This Paper Address?

Ensuring the accuracy of radiation therapy delivery is paramount, and a critical component of this is the quality assurance (QA) of the multileaf collimator (MLC). The MLC is responsible for shaping the radiation beam, and even small deviations in its leaf positioning can lead to significant under- or over-dosing of the target tumor or surrounding healthy tissues. Current guidelines, such as the AAPM TG-142 report, recommend a weekly qualitative "picket fence" test for MLCs. However, the authors question whether this simple test is sufficient, especially with the increasing use of hypofractionated radiation therapy (higher doses per fraction) and complex treatment techniques like Volumetric Modulated Arc Therapy (VMAT). They hypothesize that a more comprehensive weekly QA program is needed to catch potential failure modes that the basic picket fence test might miss, thereby improving patient safety.

Methods

The researchers developed and evaluated a "Comprehensive Weekly" (C-Weekly) MLC QA program. This program consists of five specific tests designed to analyze various aspects of MLC performance:

  1. Absolute MLC Leaf Position: Verifies the precise position of individual MLC leaves.
  2. Interdigitation Leaf Position: Assesses how well opposing MLC leaves interdigitate, a critical maneuver for complex beam shaping.
  3. Static Gantry Picket Fence: A variation of the standard picket fence test, but with intentional leaf overruns and underruns to check reproducibility.
  4. Minimum Leaf Gap Setting: Examines the smallest gap between MLC leaves, crucial for VMAT.
  5. VMAT Test: Evaluates the accuracy of VMAT delivery using gamma analysis.

These tests were performed using Electronic Portal Imaging Devices (EPIDs) on 11 linear accelerators across five clinical sites over a seven-year period. To quantify the effectiveness of their C-Weekly program compared to the standard weekly picket fence (PF-Weekly) test, they employed a Failure Mode and Effects Analysis (FMEA) approach, as recommended by AAPM TG-100. This involved scoring the probability of occurrence (O), severity of effect (S), and detectability (D) of potential MLC failure modes. The Risk Priority Number (RPN = O × S × D) was then calculated for each failure mode under both QA programs. A target RPN of less than 125 was used as the threshold for acceptable risk.

Datasets

The study analyzed data from 20,226 QA images collected over 16,855 tests (approximately 3,371 tests per year) from 11 Varian linear accelerators (7 C-series, 3 TrueBeam, and 1 TrueBeam-Edge) across 5 clinical sites between May 2014 and June 2021. This dataset represents a substantial real-world experience with weekly MLC QA.

Key Insights and Ideas

  • The standard picket fence test is insufficient: The study strongly suggests that the qualitative PF-Weekly test, while easy to perform, misses several critical failure modes that a more comprehensive approach can detect.
  • FMEA is a powerful tool for QA program comparison: The FMEA methodology provided a quantitative and objective way to compare the risk associated with different QA programs, moving beyond subjective assessments.
  • VMAT delivery is a significant source of potential failures: The VMAT test within the C-Weekly program identified the highest number of failures (55.9%), highlighting the complexity and potential pitfalls of this advanced treatment modality.
  • A comprehensive weekly QA program is clinically necessary: The significantly lower RPN scores for the C-Weekly program indicate a substantially reduced risk of MLC-related treatment errors, justifying its implementation.
  • Practicality of the C-Weekly program: The proposed program can be performed within a reasonable timeframe (10 minutes for beam delivery, 5 minutes for analysis) and utilizes readily available EPID technology, making it feasible for busy clinical environments.

Key Results

The analysis of 16,855 tests revealed a total of 143 failures (0.9%). The C-Weekly program demonstrated significantly lower RPN scores for all identified failure modes, ranging from 48 to 96, all below the acceptable threshold of 125. In contrast, the PF-Weekly program's RPN scores ranged from 60 to 192, with several modes exceeding the 125 threshold. Specifically, the static field MLC position deviation had an RPN of 168 for PF-Weekly, and the VMAT/static IMRT dose difference had an RPN of 192 for PF-Weekly, both indicating a higher risk.

Limitations and Future Work

The authors acknowledge that FMEA is inherently subjective and user-dependent, meaning there's a chance of underestimating risks if not all failure modes are considered. They also note that FMEA might not be ideal for systems with multiple simultaneous failures due to its inability to account for correlations.

Future work could involve further refining the C-Weekly program to explicitly include sliding-window IMRT testing, which is not fully covered by the current five tests. The authors are in the process of adding a step-and-shoot delivery test to compare with VMAT and sliding-window images. They also suggest that while their program is designed for conformal therapy, step-and-shoot IMRT, and VMAT, further investigation into its applicability to other advanced techniques might be warranted.

Summary updated

Abstract

RPNs for the 5 failure modes of MLC QA tests were quantitatively determined and analyzed. A comprehensive weekly MLC QA is imperative to lower the RPNs of the 5 failure modes to the desired level (<125); those from the PF-Weekly MLC QA program were found to be higher (>125). This supports the clinical necessity for comprehensive weekly MLC QA.
BibTeX
@article{kalavagunta2022is,
  author = {Chaitanya Kalavagunta and Huijun Xu and Baoshe Zhang and Sina Mossahebi and Michael MacFarlane and Kai Jiang and Sung‐Woo Lee and Shifeng Chen and Amit Sawant and Arun Gopal and B Yi},
  title = {{Is a weekly qualitative picket fence test sufficient? A proposed alternate EPID‐based weekly MLC QA program}},
  journal = {Journal of Applied Clinical Medical Physics},
  year = {2022},
  doi = {10.1002/acm2.13699},
}
Figure