TA30-Independent Verification of IMRT Delivered With a Multileaf Collimator
Memorial Sloan-Kettering Cancer Center, New York City, New York.
Introduction: Independent verification is an important component in radiation therapy quality assurance. In this presentation, we describe computational and empirical methods for independent verification of intensity modulated radiation therapy (IMRT) delivered with a multileaf collimator (MLC). The methods presented are applicable to both dynamic and segmental delivery.
Methods: The computational algorithms consist of two parts. The first part calculates the delivered intensity distribution in-air, and the second part calculates the absolute dose distribution in a homogeneous phantom. The effects included in the intensity calculation are direct exposure, mid-leaf and interleaf transmissions, rounded leaf-end, tongue-and-groove, extra-focal source distribution, and scatter from the leaves. Dose calculation in phantom is based on measured data, including the output factors, tissue-maximum ratios, and off-center ratios. The effect of intensity modulation is accounted for by pencil beam convolution. Empirical measurements utilize ion chambers and film. Chamber measurement is considered as the standard, but is limited to point measurement only. Two dimensional distributions are measured with film. Special care is needed for film dosimetry to correct for film over response due to the low energy scattered photons.
Results: The accuracy of the computational algorithms has been verified with chamber and film measurements for a variety of IMRT fields ranging from simple patterns to clinical intensity-modulated fields. These fields cover a wide range of field sizes and varying degrees of intensity modulation. The majority of the dose comes from direct exposure. Significant contribution also comes from mid-leaf transmission due to the large amount of time spent by part of the field under the MLC. The effects of interleaf transmission and tongue-and-groove effects can be predicted by the algorithms. For very large fields, scatter from the MLC may contribute up to 5%. For film measurement, the effect of film over response was analysed by Monte Carlo methods and can be included in film dosimetry. The agreement between calculation and measurement in general was within 2-3% in dose, or within 1mm in distance in high dose gradient regions. Typical calculation time per beam is less than 30 seconds on a 266 MHz Alpha station.
Summary: Computational and empirical methods have been developed for independent verification of IMRT treatment. The agreement between calculated and measured results is typically within 2-3% in dose or 1mm in distance. Calculation is now used for routine independent check, while empirical verification is performed when new conditions arise, such as new disease sites, new treatment techniques, or new software release.