Room: AAPM ePoster Library
Purpose: The Eclipse v15.5 TPS has an auto-feathering feature that allows the co-optimization of two or more isocenters in the same plan. This feature benefits long targets, such a CSI, by lengthening the dose gradient region to be robust against longitudinal setup uncertainties. This work investigated the effects that relative isocenter locations and beam geometry had on the longitudinal auto-feathered dose gradient regions for VMAT CSI plans.
Methods: Patients who had previously received VMAT CSIs were retrospectively re-planned. Re-plans systematically varied the isocenter separation (in the craniocaudal direction) and jaw settings (jaw opening, as well as jaw tracking) in order to determine the effect each of these parameters had on the auto-feathered dose gradient region. Known longitudinal setup errors were simulated within the TPS to evaluate plan robustness in the presence of setup uncertainties.
Results: Optimizing isocenter location to allow for greater field overlap lengthened the gradient region (decreased gradient slope). In a representative case, the gradient region went from 4.5 cm to 15 cm by placing isocenters in such a way that the full allowable jaw opening was utilized, while decreasing jaw openings always lead to decreased gradient regions (steeper gradient slope). For each case investigated, removing jaw tracking increased the gradient region by 1-6 cm compared to plans with jaw tracking. Plans with gently sloping dose gradients in the overlap region were more robust to simulated patient shifts.
Conclusion: Increasing the area of the jaw overlap region allows the auto-feathering function to provide longer, more gently sloping dose gradients in VMAT CSI treatment plans which, subsequently, reduces sensitivity to hot/cold spots during treatment due to setup uncertainties. Optimizing plans without the use of jaw tracking provided more gently sloping, robust dose gradients versus plans optimized with jaw tracking enabled.
Not Applicable / None Entered.
Not Applicable / None Entered.