19. Impact of modulation metrics in VMAT on pre-treatment quality assurance results: A preliminary retrospective prostate study

Abstract

Introduction In order to quantify the modulation degree of a plan, some metrics have been described in the literature [1] . For a Volumetric Modulated Arc Therapy (VMAT) arc, Modulation Index (MI) and Modulation Complexity Score (MCS) can be calculated by knowing the position of each leaf of the MLC at each control point. The aim of this study is to evaluate the correlation between metrics, respect of dosimetric constraints on DVH and QA results for the treatment of prostate cancer with VMAT. Methods Our retrospective study was focused on 34 consecutive patient treated with VMAT plans using two arcs to deliver 46 Gy (RA46) on pelvis followed by one arc delivering 20 Gy on prostate lodge (RA66). Plans were calculated on Eclipse (v13.7) and delivered on a Clinac21Ex equipped with a 120 leaves MLC. MI and MCS metrics were calculated for each of the 3 arcs for each patient with a home-made Python program, using the DICOM-RT plan. If the MI of an arc is upper than 0.18, the entire VMAT plan is re-optimized with the aim to reach the previous dosimetric goals as the MI stays below 0.18 [2] . D98%, D95% D2% for Planning Target Volume (PTV), D50% D40%, D18% for rectum and D50% for bladder were compared. QA tests are performed with Electronic Portal Imaging Devices (as1000 EPID). Gamma analysis using 3%/3 mm, 2%/2 mm and 1%/1 mm criteria are performed. QA results between initial plan and re-planed plan are compared. Results For RA46 and RA66 respectively, 4 and 12 of 37 plans were recalculated. MI and MCS relative differences are respectively (−9.7 ± 3.8)% and (10.5 ± 3.5)% for RA46, (−19.4 ± 7.6)% and (8.1 ± 10.8)% for RA66. Dosimetric parameters obtained for new plans have relative difference on PTV DVH between (−0.62 ± 0.71)% and (0.32 ± 0.31)% for both RA46 and RA66. Concerning rectum, the difference is between (−0.52 ± 2.47)% and (0.65 ± 4.35)% for both plans. On the bladder DVH, the dosimetric parameters differences are less than (−0.50 ± 0.91)% for RA46 but not for RA66 a moderate increase of the exposure is noticed (2.25 ± 4.33)% on D50%. QA results variation are between −0.68% and 0.75% for 3%/3 mm, 2%/2 mm and 1%/1 mm for RA46. Concerning RA66 plan, the relative difference on the gamma pass is −0.10% for 3%/3 mm, −1.14% for 2%/2 mm and −3.10% for 1%/1 mm. Conclusions Results indicate that plans can be recalculated with a less important modulation with conservation of dosimetric constraints. According to QA results, the decrease of the modulation doesn’t improve pre-treatment QA. The deterioration of QA results with a lower MI and higher MCS scores is an unexpected conclusion. To confirm these results, three hypotheses should be investigated: include more patients to increase their statistical confidence, evaluate the movement of EPID during arc course and investigate the relevance of MI and MCS for quantifying general complexity of plans.