基于自适应遗传算法的质子调强放疗扫描路径优化

IAGA Based Spot Scanning Path Optimization for Intensity Modulated Proton Therapy

研究和开发了基于自适应遗传算法的质子调强放疗扫描路径优化方法,并在此基础上对质子调强放疗扫描时间进行初步临床评估。利用自适应遗传算法具有的较强容错性和全空间最优搜索能力开发点扫描质子调强路径优化模块,并将其集成到自主研发的放疗计划系统,选取AAPM TG-119头颈部肿瘤和前列腺肿瘤模拟例题及两例临床病例进行测试,对比扫描路径优化前后质子放疗计划扫描路径长度。对于AAPM TG-119头颈部肿瘤和前列腺肿瘤模拟例题,总扫描路径长度分别降低了27.17%和18.72%,临床头颈部肿瘤和前列腺肿瘤病例总扫描路径长度分别降低了25.36%和32.95%,优化前后路径长度降低比率与零权重扫描点及肿瘤解剖结构有关。基于自适应遗传算法的扫描路径优化方法可减少质子调强放疗计划扫描时间,从而缩短病人治疗时间,可广泛应用于质子重扫描临床技术。

The objective of this research is to study and develop a proton spot scanning path optimization method using an improved genetic algorithm for proton therapy and to evaluate the scanning time under clinical conditions. An Improved Adaptive Genetic Algorithm（IAGA） based scanning path optimization module was developed and integrated into the home-grown treatment planning system. Four cases, including two AAPM TG-119 standard cases and two clinical cases, were selected to compare their scanning path length before and after scanning path optimization. For the two AAPM TG-119 cases, the optimized scanning path length dropped by 27.17% and 18.72%, and for the corresponding clinical cases, the optimized scanning path length dropped by25.36% and 32.95% respectively. The performance of scanning path optimization was affected by the number of zero-weight spots and connected regions in the scanning map. IAGA based scanning path optimization can reduce the total scanning path length in intensity modulated proton therapy and, therefore, can be used in spot rescanning to accommodate organ motion.