144. Metric for assessment of mechanical reproducibility of leaves positions during VMAT treatment. (December 2018)

• Record Type: Journal Article
• Authors: Fortunato, M.; Berta, L.; Colleoni, P.; Sonzogni, M.; Moretti, G.
• Is Part Of: Physica medica. Volume 56(2018)Supplement 2
• Page Start: 153
• Page End: 154
• Publication Date: 2018-12
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.04.155
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Purpose: To assess mechanical reproducibility of Multi Leaf Collimator (MLC) leaves positions during VMAT delivery and to formulate a suitable metric for daily quality assurance (QA) measurements. Methods: A plan consisting of 2 dynamic arcs (CW and CCW) has been delivered on a Varian linac for 12 months (444 total arcs) as daily QA to assess MLC Long Term Reproducibility (LTR) and 10 times consecutively (20 total arcs) for Short Term Reproducibility (STR). Dynalog files were generated from the linac and analyzed in MatLAB environment. The distribution of the positioning errors of MLC weighted for leaves aperture was calculated for every delivery. An Index Error (IE) was calculated as the percentiles of these distributions that minimize the CV among the sampled data (Fig. 1): average values ( X m ) and standard deviations (SD) were calculated in order to define the thresholds for 95% of statistic confidence of the sample (TH = X m + 2SD). IE is calculated with daily frequency and the percentage of events out of TH is evaluated including a window data of 1 month. The Attention Level (AL) is reached when the number of the events (Fig. 2) out of threshold is the 5% of the sample. The events out of the statistic confidence are related to mechanical mismatch of MLC during delivery. Results: The minimization CV value is obtained for the 49th percentile of MLC weighted positioning error distribution. The trends of average, SD and CV for the LTR and STR (Table) samples are similar this means the absence of perturbative boundary conditions. During monitoring of MLC performance the method highlighted the beginning of the malfunction (blue-line in Fig. 2) and the subsequent return into the limits following the maintenance service. Conclusions: The IE calculated as 49th percentile of MLC positioning error distributions can be used to define an AL for MLC performance monitoring for VMAT delivery and daily QA. The proposed metric provides a strong method for monitoring mechanical errors during VMAT plans. Furthermore, a mechanical equivalence between linacs could be assessed to ensure a safe machine switching when service is occurring.