Optimization of IMRT Fluence-Map Smoothing can Result in Improved IMRT Plans

Abstract

Commercially available IMRT treatment planning systems (TPS) typically include a fluence-map smoothing option designed to decrease irregularities/complexity in delivered beam fluence patterns, thus allowing for decreased monitor units and shorter treatment times. The cost of reducing the complexity of the inverse-planned optimal fluence pattern is typically a reduction in plan quality. Here we study the behavior of 3 different treatment planning systems when smoothing parameters are systematically adjusted. We evaluated a case of sinonasal carcinoma treated by IMRT in our clinic. RTOG recommendations for target coverage and object-at-risk (OAR) dose constraints were used for planning. Contours were transferred via DICOM to the Varian Eclipse v6.5, BrainLab Brainscan v5.31, and Nomos Corvus v6.2 TPS's. Each set of TPS specific plans were optimized identically to each other, except that the TPS-inherent “smoothing” or “complexity” values were varied. For each TPS, plans were optimized with a minimal smoothing level, vendor-defined default smoothing, and double the default smoothing. All plans were normalized for identical target coverage. Key metrics of plan quality and delivery efficiency were gathered for each plan. The various smoothing scenarios studied here can impact significantly on plan quality (see Fig. 1). Results varied widely, by vendor, with regard to both plan quality and delivery efficiency (see Table 1). Appropriate use of the fluence map smoothing functionality inherent to typical IMRT inverse planning systems can result in valuable improvements to plan quality, such as reduced critical structure maximum doses. The behavior of fluence-map smoothing functions varies widely by vendor. It appears important to understand the behavior of the specific TPS in use at each clinic.Tabled 1% Change Relative to Default Smoothing PlanCorvusCorvusEclipseEclipseBrainLabBrainLabSmoothing SettingMinimalMaximum/DoubleMinimalMaximum/DoubleMinimalMaximum/DoubleLeft Optic Nerve Max. Dose−15.1%12.6%0%1.8%0%0%Right Optic Nerve Max. Dose−17.7%5.5%1.5%0.8%−1.4%2.9%Optic Chiasm Max. Dose−20.3%−2.5%0.1%2.0%−1.5%0%Monitor Units−14.0%−83.6%37.8%−26.0%4.2%−6.0%Segments90.7%−96.2%13.7%−11.4%0.5%−1.5% Open table in a new tab