肺癌立体定向放疗RapidPlan商业模型的本地改造可行性探究

Objective To explore the feasibility and optimization effect of using the commercial model of RapidPlan stereotactic body radiotherapy (SBRT) for lung cancer in the Henry Ford Health System (HFHS) after being modified according to local needs for radiotherapy planning. Methods The following adaptive modifications were made to the HFHS model built in Eclipse in combination with the latest clinical guideline evidence and the actual clinical situation of our unit: ① The internal gross tumor volume (IGTV) of the target area structure and the organ at risk structure lung (Lung) were added, and corresponding parameters were set. ② The upper limit (Upper) value of the planning target volume (PTV) structure was adjusted from 109% to 125%. ③ The original training set was replaced with 73 historical simultaneous integrated boost plans of our unit, and statistical confirmation and outlier handling were performed with the assistance of the Model Analytics software. Ten cases that did not participate in the modeling were selected for independent verification and comparison of the automatic optimization results of the model before and after modification under the same field arrangement conditions. After normalizing the target dose, the following were compared: the PTV conformity index, the mean dose, the maximum dose, and the dose - volume parameters of the organs at risk, etc. Results The "tail" of the PTV DVH and the "shoulder" and "tail" of the IGTV DVH in the modified model verification plan (M_local_ P) were higher than those in the original HFHS model verification plan (HFHS_ P). The PTV_ CI (1.07 ± 0.13) of M_local_ P was significantly smaller than that of HFHS - P (1.25 ± 0.24), and the difference was statistically significant (Z = - 2.497, P < 0.05). Except for heart_ D_(15cc) and heart_D_(max), most of the dosimetric parameters of the organs at risk in M_local_ P were lower than those in HFHS_ P, and the standard deviation was smaller. However, for heart_ D_(15cc) and heart_ D_(max), the difference between the two plans did not exceed 3.06%. There were 10 dosimetric parameters in HFHS_ P that did not meet the requirements, and the PTV_ CI values of two cases that did not pass were 1.52 and 1.74, far exceeding the clinically acceptable range. Conclusion The HFHS commercial model can be locally modified by methods such as replacing the planning training set and adjusting parameters. Moreover, the plan optimized by the modified model better meets the clinical requirements of our unit in terms of target conformity and high dose, and also performs better in terms of organ at risk protection and consistency.

目的探究以本地需求为目标改造亨利福特医院健康系统(HFHS)RapidPlan肺癌立体定向放疗(SBRT)商用模型后用于放疗计划的可行性和优化效果。方法结合最新临床指南证据和本单位临床实际对Eclipse内置HFHS模型做出如下适应性改造:①增加靶区结构内在肿瘤靶区体积(IGTV)、危及器官结构肺(Lung),并设置相应参数。②将计划靶区体积(PTV)结构上限(Upper)值由109%调整至125%。③用本单位73例历史同步推量计划替换其原训练集,并在模型分析(Model Analytics)软件辅助下进行统计学确认和离群值处置。选取10例未参与建模的病例用于独立验证和比较改造前后模型在相同布野条件下的自动优化结果。靶区剂量归一后比较:PTV适形指数、危及器官平均剂量、最大剂量和剂量-体积参数等。结果改造后的模型验证计划(M_(local)_ P)PTVDVH “尾部”与IGTVDVH “肩部”及“尾部”表现较原HFHS模型验证计划(HFHS_ P)高。M_(local)_ P的PTV_ CI(1.07±0.13)显著小于HFHS- P(1.25± 0.24),差异有统计学意义(Z =-2.497,P<0.05)。除心脏_ D_(15cc)、心脏_D_(max)外M_(local)_ P绝大部分危及器官剂量学参数均较HFHS_ P更低,且标准差更小。但对于心脏_ D_(15cc)和心脏_ D_(max),两种计划差异不超过3.06%。HFHS_ P不满足要求剂量学参数达10项次,其中两项未通过例数的PTV_ CI值为1.52与1.74,远超临床可接受范围。结论HFHS商用模型可通过计划训练集替换、参数调整等方法进行本地化改造。且改造后模型优化的计划在靶区适形性与高量等方面更满足本单位临床要求,在危及器官保护和一致性等方面也表现更优。