Objective: This study aimed to develop a fully-automated patient tailored beam-angle optimisation approach for intensity-modulated proton therapy (IMPT). For oropharynx cancer patients, the dosimetric impact of increasing the number of fields from 4 to 12 was systematically assessed.
Approach: A total-beam-space heuristic was developed to simultaneously select optimal patient specific candidate beam directions, according to a cost-function that penalises dose to OARs involved in clinically used NTCPs. The method was dosimetrically validated by comparisons with fixed 4- and 6-field clinical beam-angle templates and equiangular configurations, including 72-field equiangular. The latter served as dosimetric 'Utopia' benchmark for the other evaluated beam configurations.
Main result: Using 4 optimised patient-specific fields instead of the clinical 4-field beam-angle template resulted in (xerostomia NTCP + dysphagia NTCP)-reductions for all patients, with...
Show moreObjective: This study aimed to develop a fully-automated patient tailored beam-angle optimisation approach for intensity-modulated proton therapy (IMPT). For oropharynx cancer patients, the dosimetric impact of increasing the number of fields from 4 to 12 was systematically assessed.
Approach: A total-beam-space heuristic was developed to simultaneously select optimal patient specific candidate beam directions, according to a cost-function that penalises dose to OARs involved in clinically used NTCPs. The method was dosimetrically validated by comparisons with fixed 4- and 6-field clinical beam-angle templates and equiangular configurations, including 72-field equiangular. The latter served as dosimetric 'Utopia' benchmark for the other evaluated beam configurations.
Main result: Using 4 optimised patient-specific fields instead of the clinical 4-field beam-angle template resulted in (xerostomia NTCP + dysphagia NTCP)-reductions for all patients, with averages of 3.0 %-point (range: 1.1-5.8) for grade 2 toxicity and 1.2 %-point (range: 0.3-2.8) for grade 3. For 6 fields these reductions were 2.4 %-point (range: 0.0-5.0) and 0.8 %-point (range:- 0.1-2.1). Xerostomia NTCPs significantly reduced with increasing numbers of patient-specific fields with a levelling off at 10-12 fields with NTCP values that closely approached those for Utopia 72-field equiangular plans. Beam angle optimisation took 52 min.
Conclusion: Automated, patient-tailored beam-angle optimisation could enhance IMPT plans at acceptable optimisation times. Improvements compared to the clinical beam-angle templates were highly patient-specific.
Show less
