Optimization of normalized prescription isodose selection for stereotactic body radiation therapy: Conventional vs robotic linac

Purpose: Although modern technology has allowed for target dose escalation by minimizing normal tissue dose, the dose delivered to a tumor and surrounding tissues still depends largely on the inherent characteristics of the radiation delivery platform. This work aims to determine the optimal prescription isodose line that minimizes normal tissue irradiation for stereotactic body radiation therapy (SBRT) for a conventional linear accelerator and a robotic delivery platform. Methods: Spherical targets with diameters of 10, 20, and 30 mm were constructed in the lungs and liver of a computer based digital torso phantom which simulates respiratory and cardiac motion. Normal tissue contours included normal lung, normal liver, and a concentric 10 mm shell of normal tissue extending from the spherical target surface. For linac planning, noncoplanar, nonopposing three dimensional (3D) conformal beams were designed, and variable prescription isodose lines were achieved by varying the MLC block margin. For CyberKnife planning, variable prescription isodose lines were achieved by inverse planning. True 4D dose calculations were used for the moving target and surrounding tissue based on each of ten phases of a 4D CT dataset. Doses of 60 Gy in three fractions were prescribed to cover 95 of the target tumor. Commonly used conformality, dosimetric, and radiobiological indices for lung and liver SBRT were used to compare different plans and determine the optimally prescribed isodose line for each treatment platform. Results: For linac plans, the average optimal prescription isodose line based on all indices evaluated occurred between 59 and 69 for lung tumors and between 67 and 77 for liver tumors depending on the tumor size. CyberKnife plans had average optimal prescription isodose lines occurring between 40 and 48 for lung tumors and between 41 and 42 depending on the tumor size. However, prescription isodose lines under 50 are not advised to prevent large heterogeneous dose distributions within the target. Conclusions: The choice of prescription isodose line was shown to have a significant impact on parameters commonly used as constraints for lung and liver SBRT treatment planning for both linac-based and CyberKnife delivery platforms. By methodically choosing the prescription isodose line, normal tissue toxicities from SBRT may be reduced.