2010 AAPM Annual Meeting
Dr. Thomas R. Mackie, PhD, University of Wisconsin - ADCL, Madison, WI, 53706
For more information about the American Association of Physicists in Medicine, visit aapm.org/
Commercial clinical treatment planning systems(TPS) employ various algorithmsfor
calculating dose distributions and determining optimal beam fluences and apertures. We
present a review of commonly-used computational techniques, as well as considerations
relevant to clinical physicists. These include TPS commissioning, verification,
heterogeneity corrections,small-field accuracy, and QA. In addition, we discuss newer
methods which are becoming more widely adopted such as MonteCarlo dose algorithms
and multicore/distributed computational techniques.
Radiotherapy dose calculation algorithms are now almost exclusivelymodel-based
instead of correction-based. Model-based algorithms construct from first principlesthe
dose in the patient using models ofradiation interaction with matter. The most commonly
used model-based algorithm isthe convolution/superposition approach but MonteCarlo
algorithms are now more generally available in commercial planning systems. The
accuracy ofthe predicted dose distributionsis only as good asthe performance ofthe
algorithm and the fidelity ofthe input data. Using data measured using a water phantom,
the energy spectrum can be determined from measured depth-dose data while the fluence
profile can be determined from profile data. With some IMRT orstereotactic systems not
using field flattening systems, the intensity profile can be quite non-uniform. The effect
ofleaves, collimators and accessories are taken into account with special measurements
that depend on the delivery system. New methods of obtaining the necessary data are
being established that promise more convenience and accuracy.
Dose calculation algorithms are also being used for other purposes. Independent dose
calculations are being used assecondary checksto verify the results oftreatment
planning systemsso that patient-specific measurements may not need to be performed.
Dose reconstruction can be used to determine the dose actually received by the patientso
that the effect of changesin anatomy or machine output can ascertained. The ICRU will
be soon releasing new guidelinesfor quality assurance of dose calculation systems
including specificationsfor accuracy.
In orderto take full advantage ofthe capabilities of commercial inverse planning
solutions, it is critical to have an understanding of the underlying optimization algorithms
and optimization tools. The firststep in the inverse planning processisto define your
treatment goals. The objectives and constraints commonly used in commercial planning
systems will be discussed along with the advantages and pitfalls of using biology-based
functionssuch as equivalent uniform dose. Aftertreatment goals have been defined, the
treatment plan optimization isthen performed. Comparisons will be provided ofthe key
optimization algorithmsin routine clinical use along with techniquesforimproving
delivery efficiency such as direct aperture optimization. The final portion ofthe talk will
explore future developmentsin IMRT planning. For example, we will discuss
multicriteria optimization (MCO), and how this enhanced planning capability holdsthe
promise of dramatically reducing treatment planning times by allowing usersto rapidly
explore tradeoffsin target dose coverage and criticalstructure sparing.
1. To understand the physics of model-based dose calculations including commissioning, verification, heterogeneity corrections,small-field accuracy, and QA.
2. To understand the use of dose calculation algorithms in patient QA such as secondary monitor checks and dose reconstruction.
3. To understand the differences between commercial inverse planning solutions and how to best utilize the algorithms to maximize IMRT plan quality.