Comparative analysis of Pareto surfaces in multi-criteria IMRT planning.

In the multi-criteria optimization approach to IMRT planning, a given dose distribution is evaluated by a number of convex objective functions that measure tumor coverage and sparing of the different organs at risk. Within this context optimizing the intensity profiles for any fixed set of beams yields a convex Pareto set in the objective space. However, if the number of beam directions and irradiation angles are included as free parameters in the formulation of the optimization problem, the resulting Pareto set becomes more intricate. In this work, a method is presented that allows for the comparison of two convex Pareto sets emerging from two distinct beam configuration choices. For the two competing beam settings, the non-dominated and the dominated points of the corresponding Pareto sets are identified and the distance between the two sets in the objective space is calculated and subsequently plotted. The obtained information enables the planner to decide if, for a given compromise, the current beam setup is optimal. He may then re-adjust his choice accordingly during navigation. The method is applied to an artificial case and two clinical head neck cases. In all cases no configuration is dominating its competitor over the whole Pareto set. For example, in one of the head neck cases a seven-beam configuration turns out to be superior to a nine-beam configuration if the highest priority is the sparing of the spinal cord. The presented method of comparing Pareto sets is not restricted to comparing different beam angle configurations, but will allow for more comprehensive comparisons of competing treatment techniques (e.g., photons versus protons) than with the classical method of comparing single treatment plans.

Trade-off bounds for the Pareto surface approximation in multi-criteria IMRT planning.

One approach to multi-criteria IMRT planning is to automatically calculate a data set of Pareto-optimal plans for a given planning problem in a first phase, and then interactively explore the solution space and decide on the clinically best treatment plan in a second phase. The challenge of computing the plan data set is to ensure that all clinically meaningful plans are covered and that as many clinically irrelevant plans as possible are excluded to keep computation times within reasonable limits. In this work, we focus on the approximation of the clinically relevant part of the Pareto surface, the process that constitutes the first phase. It is possible that two plans on the Pareto surface have a small, clinically insignificant difference in one criterion and a significant difference in another criterion. For such cases, only the plan that is clinically clearly superior should be included into the data set. To achieve this during the Pareto surface approximation, we propose to introduce bounds that restrict the relative quality between plans, the so-called trade-off bounds. We show how to integrate these trade-off bounds into the approximation scheme and study their effects. The proposed scheme is applied to two artificial cases and one clinical case of a paraspinal tumor. For all cases, the quality of the Pareto surface approximation is measured with respect to the number of computed plans, and the range of values occurring in the approximation for different criteria is compared. Through enforcing trade-off bounds, the scheme disregards clinically irrelevant plans during the approximation. Thereby, the number of plans necessary to achieve a good approximation quality can be significantly reduced. Thus, trade-off bounds are an effective tool to focus the planning and to reduce computation time.

Pareto navigation: algorithmic foundation of interactive multi-criteria IMRT planning.

Inherently, IMRT treatment planning involves compromising between different planning goals. Multi-criteria IMRT planning directly addresses this compromising and thus makes it more systematic. Usually, several plans are computed from which the planner selects the most promising following a certain procedure. Applying Pareto navigation for this selection step simultaneously increases the variety of planning options and eases the identification of the most promising plan. Pareto navigation is an interactive multi-criteria optimization method that consists of the two navigation mechanisms 'selection' and 'restriction'. The former allows the formulation of wishes whereas the latter allows the exclusion of unwanted plans. They are realized as optimization problems on the so-called plan bundle -- a set constructed from pre-computed plans. They can be approximately reformulated so that their solution time is a small fraction of a second. Thus, the user can be provided with immediate feedback regarding his or her decisions. Pareto navigation was implemented in the MIRA navigator software and allows real-time manipulation of the current plan and the set of considered plans. The changes are triggered by simple mouse operations on the so-called navigation star and lead to real-time updates of the navigation star and the dose visualizations. Since any Pareto-optimal plan in the plan bundle can be found with just a few navigation operations the MIRA navigator allows a fast and directed plan determination. Besides, the concept allows for a refinement of the plan bundle, thus offering a middle course between single plan computation and multi-criteria optimization. Pareto navigation offers so far unmatched real-time interactions, ease of use and plan variety, setting it apart from the multi-criteria IMRT planning methods proposed so far.

Improved stratification algorithms for step-and-shoot MLC delivery in intensity-modulated radiation therapy.

In inverse planning for intensity-modulated radiotherapy (IMRT), the fluence distribution of each treatment beam is usually calculated in an optimization process. The delivery of the resulting treatment plan using multileaf collimators (MLCs) is performed either in the step-and-shoot or sliding window technique. For step-and-shoot delivery, the arbitrary beam fluence distributions have to be transformed into an applicable sequence of subsegments. In a stratification step the complexity of the fluence maps is reduced by assigning each beamlet to discrete intensity values, followed by the sequencing step that generates the subsegments. In this work, we concentrate on the stratification for step-and-shoot delivery. Different concepts of stratification are formally introduced. In addition to already used strategies that minimize the difference between original and stratified beam intensities, we propose an original stratification principle that minimizes the error of the resulting dose distribution. It could be shown that for a comparable total number of subsegments the dose-oriented stratification results in a better approximation of the original, unsequenced plan. The presented algorithm can replace the stratification routine in existing sequencer programs and can also be applied to interpolated plans that are generated in an interactive decision making process of multicriteria inverse planning programs.