Point dose measurements in VMAT : an investigation of detector choice and plan complexity.
Thesis DisciplineMedical Physics
Degree GrantorUniversity of Canterbury
Degree NameMaster of Science
Purpose: Volumetric Modulated Arc Therapy (VMAT) is an intensity modulated radiation therapy technique which can achieve highly conformal dose delivery through dynamic variation of dose rate, gantry speed, and multileaf collimator positions. Due to the complexity of treatment delivery, patient specific quality assurance (QA) is required to ensure agreement between calculated and delivered dose. Point dose measurements are a well established patient specific QA technique for VMAT. The aim of this study was to investigate the relationship between plan complexity and the agreement between measured and calculated point doses. The suitability of five different detectors for VMAT point dose measurements was also evaluated.
Methods: 45 previously treated prostate VMAT plans were selected for the study. Isocentre point dose measurements were carried out on a Varian iX linear accelerator using five commercial detectors in the CIRS Model 009 Cube 20 phantom. Measurements were made with IBA CC01 and CC04 compact ionisation chambers, IBA EFD3G and PFD3G diodes, and a PTW 60019 microDiamiond detector. Detector measurement repeatability was investigated and quantified by repeat measurements over three measurement sessions. The calculated dose was computed in both Pinnacle, using both 4 degree and 2 degree per control point gantry spacing (GS), and RayStation treatment planning systems. The agreement between measured and calculated dose was evaluated for each detector and calculation algorithm. A selection of established and novel aperture complexity metrics were calculated for the plan cohort. Correlations between complexity metrics and point dose discrepancy results were investigated.
Results: A statistically significant difference in mean measured dose between the CC04 chamber and all other detectors was found at the 95% confidence level. The between measurement sessions standard deviation was less than 0.5% of mean measured dose for all detectors excluding the PFD3G. The CC01 achieved the greatest repeatability followed by the CC04, EFD3G, microDiamond, and PFD3G. A statistically significant difference in mean calculated dose was found for Pinnacle (both 4 and 2 degree GS) and RayStation calculation algorithms. For both 2 degree and 4 degree GS the mean point dose discrepancy is less than 0.55% for all detectors.
Statistically significant linear relationships were found, with weak to moderate strength Pearson correlation coefficients, for the following established complexity metrics MCSv, PI, PM, CAS, CLS, and MAD metrics. The strongest Pearson correlation coefficient, r = 0.407, was found for the PI metric with CC04 measured and Pinnacle 2 degree GS calculated dose. Evaluation of plan complexity in progressively smaller ROI centred on the dose measurement and calculation point increases the correlation strength for some complexity metrics.
Conclusion: Both the choice of a dose calculation algorithm and detector have a significant influence on point dose discrepancy results. Consequently, the strength of correlations between complexity metrics and point dose discrepancy is algorithm and detector specific. Therefore, the utility of individual complexity metrics to identify plans likely to fail QA will be department specific. The poor correlation strength of complexity metrics with point dose discrepancy results limits their clinical usefulness in identification of plans likely to fail QA.