ESTRO 2021 Abstract Book

S211

ESTRO 2021

Figure. Example of how SPMs could be used/displayed in clinical practice. a.) Parotid gland image with overlaid SPM. Red, orange, yellow and green are contours for which >10%, 35%, 60% and 80%, respectively, of the generated models agree that the voxels are part of the gland. b.) The same image with the parotid gland contour generated by the model (green, dice=0.86; no overlaid SPM) and red dotted circles indicating the most uncertain areas based on the SPM in (a). Conclusion We propose the integration of uncertainty information into DL-generated contours. This would help to highlight where a contour may be less reliable and reduce the possibility that they would be overlooked. SPMs are one way to achieve this and could be integrated into the on-line ART workflow. More work needs to be done to determine the optimal DL methodology for generating the uncertainty data. OC-0304 A novel direct machine-specific parameter Spot-scanning proton arc (SPArc) optimization algorithm X. Ding 1 , G. Liu 1 , L. Zhao 1 , D. Yan 1 , X. Li 1 1 Beaumont Health, Radiation Oncology, Royal Oak, USA Purpose or Objective It has been a challenge to implement proton arc therapy with the existing clinical system due to the mechanical limitation from the hundreds of tons of gantry and the complicated beamline configurations which need to switch numerous energy layers and spots precisely during the gantry rotation. To address these engineering challenges and generate a deliverable and efficient spot-scanning proton arc(SPArc) plan for a proton therapy system, we developed a novel SPArc optimization algorithm(SPArc DMSP ) by directly incorporating the machine-specific parameters such as mechanical constraints and delivery sequence. ) was built based on the machine-specific parameters of the prototype arc delivery system, IBA ProteusONE®, including mechanical constraint(maximum gantry speed, acceleration, and deceleration speed) and delivery sequence(energy and spot delivery sequence and time). SPArc DMSP resamples and adjusts each control point's delivery speed based on the DSM arc calculation through the iterative approach(Figure 1). In SPArc DMSP, clinical users could set the expected arc delivery time and gantry max acceleration as a mechanical constraint during the plan optimization. Four representative cases(brain, head neck, liver, and lung cancer patients) were selected to test SPArc DMSP . Two kinds of SPArc plans were generated using the same planning objective functions:(1)SPArc DMSP plan meeting the maximum allowable gantry acceleration speed(0.6deg/s 2 ); (2)SPArc DMSP-user-speed plan with a user pre-defined delivery time and constraint the acceleration speed<0.1deg/s 2 . Additionally, the arc delivery sequence was simulated based on the DSM arc and was compared. Materials and Methods A SPArc delivery sequence model(DSM arc