Implementation and preliminary evaluation of a practical digital workflow for routine laser-radiation field...

To implement and preliminarily evaluate a practical digital acquisition-and-analysis workflow for routine laser-radiation field coincidence quality assurance (QA) on the CyberKnife M6 by integrating multileaf collimator (MLC) QA tool-based laser reference definition, QA StereoChecker (QASC)-based external digital image acquisition, and automated analysis with CKQA, an in-house CyberKnife QA software, and to compare this workflow with clinically realistic film-based alternatives.

Methods a and b were film-based measurements performed at the perch position. Method a used manual laser marking on film, whereas method b used the central metal marker of the MLC QA tool to define the laser reference. Method c used QASC as an external digital image-acquisition platform in combination with the MLC QA tool, and method d used film on the same QASC-based setup as a control. All methods were tested with 10 fixed collimators (5-40 mm). Images were analyzed using ImageJ and CKQA. Mean deviation, standard deviation, root mean square error (RMSE), and pairwise two-sided exact Wilcoxon signed-rank tests were used for exploratory comparisons.

All evaluated workflows achieved submillimeter mean deviations. Across both analysis approaches, the mean deviation ranked consistently in descending order: method a > method b > method d > method c. Under the present experimental conditions, method c yielded the lowest mean deviation and RMSE (0.116 mm and 0.155 mm with CKQA), whereas method a showed the largest deviations. Dataset-level exact Wilcoxon signed-rank tests showed that several comparisons involving method a were statistically significant, but statistically significant differences were not uniform across both ImageJ- and CKQA-based comparisons. For comparisons between analysis methods within each workflow, CKQA and ImageJ were not significantly different for methods a and b, whereas CKQA yielded significantly lower deviations than ImageJ for methods c and d (p < 0.05).

Laser-radiation field coincidence testing is an established CyberKnife QA procedure. The present study implemented and preliminarily evaluated a practical digital acquisition-and-analysis workflow for this routine task. The workflow combining MLC QA tool-based laser reference definition, QASC-based external digital image acquisition, and automated CKQA analysis provided a standardized and less operator-dependent procedure under the present experimental conditions. Its main value lies in reducing manual marking, film handling, scanning, and subjective image-analysis steps rather than in introducing a new physical measurement principle. In centers where QASC is not available, the film-based workflow using the MLC QA tool (method b) may serve as a practical alternative.