• Haibo Lin, Xuanfeng Ding, Maura Kirk, Haoyang Liu, Huifang Zhai, Christine E Hill-Kayser, Robert A Lustig, Zelig Tochner, Stefan Both, and James McDonough.
• Department of Radiation Oncology, University of Pennsylvania, Philadelphia, Pennsylvania. Electronic address: LinH@uphs.upenn.edu.
• Int. J. Radiat. Oncol. Biol. Phys. 2014 Sep 1; 90 (1): 71-8.

PurposeTo propose and validate a craniospinal irradiation approach using a proton pencil beam scanning technique that overcomes the complexity of the planning associated with feathering match lines.Methods And MaterialsTen craniospinal irradiation patients had treatment planned with gradient dose optimization using the proton pencil beam scanning technique. The robustness of the plans was evaluated by shifting the isocenter of each treatment field by ±3 mm in the longitudinal direction and was compared with the original nonshifted plan with metrics of conformity number, homogeneity index, and maximal cord doses. An anthropomorphic phantom study using film measurements was carried out on a plan with 5-cm junction length. To mimic setup errors in the phantom study, fields were recalculated with isocenter shifts of 1, 3, 5, and 10 mm longitudinally, and compared with the original plans and measurements.ResultsUniform dose coverage to the entire target volumes was achieved using the gradient optimization approach with averaged junction lengths of 6.7 ± 0.5 cm. The average conformity number and homogeneity index equaled 0.78 ± 0.03 and 1.09 ± 0.01, respectively. Setup errors of 3 mm per field (6 mm in worst-case scenario) caused on average 4.6% lower conformity number 2.5% higher homogeneity index and maximal cord dose of 4216.1 ± 98.2 cGy. When the junction length was 5 cm or longer, setup errors of 6 mm resulted in up to 12% dosimetric deviation. Consistent results were reached between film measurements and planned dose profiles in the junction area.ConclusionsLongitudinal setup errors directly reduce the dosimetric accuracy of the proton craniospinal irradiation treatment with matched proton pencil beam scanning fields. The reported technique creates a slow dose gradient in the junction area, which makes the treatment more robust to longitudinal setup errors compared to conventional feathering methods.Copyright © 2014 Elsevier Inc. All rights reserved.

16 keywords...

hide…