Multicriteria optimization informed VMAT planning

Abstract

We developed a patient-specific volumetric-modulated arc therapy (VMAT) optimization procedure using dose-volume histogram (DVH) information from multicriteria optimization (MCO) of intensity-modulated radiotherapy (IMRT) plans. The study included 10 patients with prostate cancer undergoing standard fractionation treatment, 10 patients with prostate cancer undergoing hypofractionation treatment, and 5 patients with head/neck cancer. MCO-IMRT plans using 20 and 7 treatment fields were generated for each patient on the RayStation treatment planning system (clinical version 2.5, RaySearch Laboratories, Stockholm, Sweden). The resulting DVH of the 20-field MCO-IMRT plan for each patient was used as the reference DVH, and the extracted point values of the resulting DVH of the MCO-IMRT plan were used as objectives and constraints for VMAT optimization. Weights of objectives or constraints of VMAT optimization or both were further tuned to generate the best match with the reference DVH of the MCO-IMRT plan. The final optimal VMAT plan quality was evaluated by comparison with MCO-IMRT plans based on homogeneity index, conformity number of planning target volume, and organ at risk sparing. The influence of gantry spacing, arc number, and delivery time on VMAT plan quality for different tumor sites was also evaluated. The resulting VMAT plan quality essentially matched the 20-field MCO-IMRT plan but with a shorter delivery time and less monitor units. VMAT plan quality of head/neck cancer cases improved using dual arcs whereas prostate cases did not. VMAT plan quality was improved by fine gantry spacing of 2 for the head/neck cancer cases and the hypofractionation-treated prostate cancer cases but not for the standard fractionation-treated prostate cancer cases. MCO-informed VMAT optimization is a useful and valuable way to generate patient-specific optimal VMAT plans, though modification of the weights of objectives or constraints extracted from resulting DVH of MCO-IMRT or both is necessary.

Keywords: Head/neck; MCO-IMRT; Prostate; VMAT.

Fig. 1

Basic work flow of MCO-informed VMAT planning

Fig. 2

(a) Dose distribution comparison among VMAT, 7field-MCO-IMRT, and 20field-MCO-IMRT plans for a typical standard fractionation prostate patient and (b,c) DVH comparisons for two typical standard fractionation prostate patients.

Fig. 3

(a) Dose distribution comparison among VMAT, 7field-MCO-IMRT and 20field-MCO-IMRT plans for a typical hyporactionation prostate patient and (b,c) DVH comparisons for two typical hyporactionation prostate patients.

Fig. 4

(a) Dose distribution comparison and (b) DVH comparison among VMAT, 7field-MCO-IMRT and 20field-MCO-IMRT plans for a typical head/neck patient

Fig. 5

(a) Dose distribution comparison and (b) DVH comparison between single arc and dual arc VMAT plans for a typical head/neck patient.

Fig. 6

(a) DVH comparisons among gantry spacings 2, 3, and 4 for VMAT plans of a typical standard fractionation prostate patient, (b) hypofractionation prostate patient, and (c) head/neck patient

Fig. 7

DVH comparisons of VMAT plans among using different delivery times. (a) delivery time of 1 min, 1.5 min, and 3 min of a typical standard fractionation prostate patient, (b) delivery time of 1.5 min, 3.5 min, and 5 min of a typical hypofractionation prostate patient, (c) delivery time of 1.5 min*2, 3.5 min*2, and 5 min*2 for the dual arc VMAT plans of a typical head/neck patient.