Author: Zhitao Dai, Riao Dao, Xuanfeng Ding, Qingkun Fan, Yajun Jia, Gang Liu, Gang Peng, Yujia Qian, Hong Quan, Kunyu Yang, Sheng Zhang, Lewei Zhao π¨βπ¬
Affiliation: Department of Radiation Oncology, Stanford University, Corewellhealth William Beaumont University Hospital, School of Mathematics and Statistics, Wuhan University, Cancer Center, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Chinese Academy of Medical Sciences Cancer Hospital Shenzhen Hospital, School of Physics and Technology,Wuhan University, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen HospitalοΌ Chinese Academy of Medical Sciences and Peking Union Medical College π
Purpose: Lattice radiation therapy (LRT) is an innovative three-dimensional(3D) implementation type of spatially fractionated radiation therapy (SFRT). This study aims to develop a novel spot-scanning proton arc (SPArc) based lattice technique to deliver highly modulated peak-valley spatial dose distribution and modulated linear energy transfer (LET) distribution to targets.
Methods: We introduce SPArc based lattice algorithm which optimizes three key parameters sequentially including the energy layer selection, spot sparsity optimization and LET optimization (SPArcLESL). Simulated annealing (SA) algorithm was used to select energy layers. Meanwhile, spot sparsity optimization algorithm primal dual active set (PDAS) was developed to generate optimal spot distribution. With minimum monitor unit (MMU) constraint, use the alternating direction method of multipliers (ADMM) to optimize the LET distribution. Nine disease sites were selected to evaluate the plan quality and the delivery efficiency compared with the previously SPArc energy sequence optimization algorithm (SPArc_seq). The lattice Tumor Volume (LTV) is generated by 7-19 small spheres in the Gross Tumor Volume (GTV) via the in-house developed automatic contouring software based on the densest stacking algorithm. The beam delivery time (BDT) were simulated based on the published proton delivery sequence model.
Results: Compared to the SPArc_seq, the SPArcLESL plan exhibits superior spot distribution and higher delivery efficiency while improving the plan quality. More specifically, SPArcLESL effectively achieve superior dose gradients which improved the peak-to-valley dose ratio (PVDR) from 17.50Β±9.98 to 44.36Β±55.02(p < 0.01), shortened the BDT by 30.71% and increased the target spot coverage by 89.66% with reducing the spot number by 45.73%, while increasing the mean LET at the LTV from 4.23Β±0.53 to 4.97Β±0.73(p < 0.01), in average.
Conclusion: A novel SPArc lattice technique could deliver superior peak-valley spatial dose distribution which paves the roadmap for the clinical use of proton LATTICE.