Encrypted login | home

Program Information

Replacing Pre-Treatment Phantom QA with 3D In-Vivo Portal Dosimetry for IMRT Breast Cancer

no image available
J Stroom

J Stroom1*, S Vieira2 , I Olaciregui-Ruiz3 , R Rozendaal4 , M van Herk5 , E Moser6 , C Greco7 , (1) ,,,(2) Champalimaud Foundation, Lisbon, ,(3) The Netherlands Cancer Institute, Amsterdam, North Holland, (4) The Netherlands Cancer Institute, Amsterdam, North Holland, (5) The Netherlands Cancer Institute, Amsterdam, North Holland, (6) Champalimaud Foundation, Lisbon, Lisbon, (7) Champalimaud Foundation, Lisbon,


SU-F-BRE-13 Sunday 4:00PM - 6:00PM Room: Ballroom E

Purpose:Pre-treatment QA of individual treatment plans requires costly linac time and physics effort. Starting with IMRT breast treatments, we aim to replace pre-treatment QA with in-vivo portal dosimetry.

Methods:Our IMRT breast cancer plans are routinely measured using the ArcCheck device (SunNuclear). 2D-Gamma analysis is performed with 3%/3mm criteria and the percentage of points with gamma<1 (nG1) is calculated within the 50% isodose surface. Following AAPM recommendations, plans with nG1>90% are approved; others need further inspection and might be rejected. For this study, we used invivo portal dosimetry (IPD) to measure the 3D back-projected dose of the first three fractions for IMRT breast plans. Patient setup was online corrected before for all measured fractions. To reduce patient related uncertainties, the three IPD results were averaged and 3D-gamma analysis was applied with abovementioned criteria . For a subset of patients, phantom portal dosimetry (PPD) was also performed on a slab phantom.

Results:Forty consecutive breast patients with plans that fitted the EPID were analysed. The average difference between planned and IPD dose in the reference point was -0.7+/-1.6% (1SD). Variation in nG1 between the 3 invivo fractions was about 6% (1SD). The average nG1 for IPD was 89+/-6%, worse than ArcCheck (95+/-3%). This can be explained by patient related factors such as changes in anatomy and/or model deficiencies due to e.g. inhomogeneities. For the 20 cases with PPD, mean nG1 was equal to ArcCheck values, which indicates that the two systems are equally accurate. These data therefore suggest that proper criteria for 3D in-vivo verification of breast treatments should be nG1>80% instead of nG1>90%, which, for our breast cases, would result in 5% (2/40) further inspections.

Conclusion:First-fraction in-vivo portal dosimetry using new gamma-evaluation criteria will replace phantom measurements in our institution, saving resources and yielding 3D dosimetry of the actual patient treatment.

Contact Email: