Related Experiment Video
Updated: Jan 7, 2026
Radiation Planning Assistant - A Streamlined, Fully Automated Radiotherapy Treatment Planning System
Published on: April 11, 2018
Automated Planning for Bias-Free Validation of Discrete Proton Arc Therapy for Oropharyngeal Cancer
Wens Kong1, Merle Huiskes2, Steven J M Habraken2,3
1Department of Radiotherapy, Erasmus MC Cancer Institute, Erasmus University Medical Center, Rotterdam, the Netherlands.
Purpose:
To systematically compare fully automated planning for 36-field discrete proton arc therapy (36PAT) with: (1) intensity-modulated proton therapy (IMPT) with 4- and 6-field clinical beam-angle class solutions (4CS, 6CS), (2) IMPT with patient-specific, computer-optimized beam-angle configurations with 6, 8 or 10 fields (6BAO, 8BAO, 10BAO), and (3) 36-field equiangular ''Utopia'' IMPT (36Utopia; total focus on dosimetric quality, no delivery time considerations) for oropharyngeal cancer patients.
Materials And Methods:
All automated plan generations were performed with Erasmus-iCycle. An energy layer (EL) reduction algorithm was developed in Erasmus-iCycle to balance plan quality with delivery time in discrete PAT planning. Patient-tailored beam-angle configurations for BAO plans were obtained with the earlier published iCycle-pBAO. In 36Utopia plans, no EL reductions were applied. While beam configurations varied among approaches, all final plans were automatically generated using the published wish-list driven SISS-MCO optimizer with dosimetry-based sparsity-induced spot selection (SISS), followed by multi-criterial spot weight optimization (MCO) and resulting in Pareto-optimality in spot weights. The same wish-list for all delivery approaches prevented planning bias across approaches.
Results:
36PAT plans achieved organs at risk (OAR) doses and normal tissue complication probability (NTCP) approaching 36Utopia plans. Relative to CS plans, 36PAT plans reduced normal tissue dose, decreasing xerostomia and dysphagia NTCPs. Compared to 4CS, 36PAT reduced summed NTCPs for grade 2 toxicity by 6.1%-point (P = .002) and grade 3 by 2.1%-point (P = .002). For 6CS, reductions were 4.7%-point (P = .002) and 1.2%-point (P = .01), respectively. 36PAT plans also outperformed BAO plans with 6 and 8 fields but were comparable to 10BAO in OAR doses and NTCPs for similar EL numbers and Monitor Units (MU).
Conclusion:
PAT demonstrated superior dosimetric quality over clinical class solutions for oropharyngeal cancer and approached Utopia. Ten-field IMPT with personalized beam angles could be an alternative to 36-field PAT with similar expected toxicity, ELs and MUs, but a lower number of fields.
