S2346
Physics - Quality assurance and auditing
ESTRO 2026
integrated MLC QA method with high robustness is proposed by incorporating log files. Material/Methods: A series of electronic portal imaging device (EPID) PF test patterns were designed by varying strip width (range: 7-13 mm, increment: 1 mm) and spacing (range: 10-20 mm, increment: 2 mm) to quantify the influence of strip beam radiation scatter on three algorithmic parameters (full width at half maximum, peak height, and peak area). This influence was partially compensated by establishing a scatter library. Subsequently, three MLC QA algorithms (FWHM- based, peak height - based, and peak area - based) were developed using the respective parameters derived from log files. To enhance accuracy and robustness, two integrated approachesa voting-based method and a linear fitting-based methodwere proposed by combining the three algorithms. Additionally, the effects of different data processing strategies and strip spacing on the positional accuracy of the developed algorithms were analyzed to identify the optimal algorithm development paradigm. Results: With increasing strip spacing, the FWHM and peak area of strips at all positions in the PF test images increased, while the peak height decreased. At a fixed strip spacing, both FWHM and peak height gradually decreased from the central strip toward the peripheral strips. Except for the outermost strips, the peak area remained nearly constant across strips. Developing algorithms independently for each strip and each leaf pair yielded the highest positional accuracy. When the strip spacing was no less than 16 mm, the overall absolute positional error remained relatively consistent (0.155 ± 0.168 mm). The respective absolute errors for each algorithm were: 0.294 ± 0.213 mm (peak height method), 0.100 ± 0.086 mm (peak area method), 0.071 ± 0.054 mm (FWHM method), 0.057 ± 0.062 mm (voting integration method), and 0.036 ± 0.528 mm (fitting integration method). The main results are showed in Figure 1-3.
Conclusion: A minimal data pipeline (SQL plus a commercial visualization tool or internal web dashboards) can support real-time feedback and operational learning in radiation oncology. By shifting from retrospective audits to proactive workflow monitoring, departments can identify bottlenecks early, implement targeted improvements, and enhance both efficiency and quality of care. Automated dashboards with individualized feedback reports were effective in increasing accountability and improving contour timeliness. The approach is generalizable and reproducible across clinical sites and system vendors. Keywords: Quality Improvement, Workflow Efficiency, Auditing Digital Poster 1496 MLC QA based on EPID picket fence test: An integrated method to partially eliminate radiation scatter from strip beams Qianxi Ni, Chenlu Liu Radiation Oncology, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University/Hunan Cancer Hospital, Changsha, China Purpose/Objective: Current multi-leaf collimator (MLC) quality assurance methods based on the picket fence (PF) test are influenced by radiation scatter from strip beams, which may potentially lead to insufficient analytical accuracy. This study aims to quantify the impact of inter-strip radiation scatter on algorithmic parameters and partially mitigate this effect. Subsequently, an
Made with FlippingBook - Share PDF online