RapidPlan模型跨病种应用的可行性探索研究

Feasibility study on the application of RapidPlan model across disease species

摘要:

目的 探索RapidPlan自动计划模型跨病种应用的可能性。方法 抽取10例直肠癌手动优化临床应用（CMO）计划作为测试组，使用已被证实在宫颈癌自动计划中表现优秀的RapidPlan模型对其进行优化，并对比直肠癌手动计划和自动计划的差异，判断模型的跨病种应用的可行性。结果 模型生成的自动计划与手动计划均满足临床需求，相较于手动计划，RapidPlan自动计划模型优化计划的危及器官（Organ at Risk, OAR）膀胱V₃₀、V₄₀、V₅₀分别降低了19.0%、17.6%、20.7%（P<0.05）。计划靶区（PTV）的D₂降低了1.6%，D₉₈提高了0.9%，适形度指数（CI）值降低了5.6%（P<0.05），均匀度指数（HI）则差异无统计学意义（P>0.05）。结论 利用宫颈癌RapidPlan自动计划模型优化直肠癌计划得到的结果相较于手动计划更优，模型的跨病种使用在临床上有一定的可行性。

Abstract:

Objective To explore the possibility of applying RapidPlan automatic planning model across different diseases.Methods Ten cases of clinical manual optimization (CMO) for rectal cancer were selected as the test group, and the RapidPlan model, which has been proven to perform well in cervical cancer automatic planning, was used to optimize it. The differences between manual and automatic plans were compared to determine the feasibility of the model's cross disease application.Results Both the automatic and manual plans generated by the model met clinical needs. Compared with manual plans, the RapidPlan automatic plan model optimized plans for organ at risk (OAR), bladder V30, V40, and V50 decreased by 19.0% (P=0.035), 17.6% (P=0.042), and 20.7% (P=0.001), respectively. The D2 of planning target volume (PTV) decreased by 1.6% (P=0.016), D98 increased by 0.9% (P=0.002), and the conformance index (CI) value decreased by 5.6% (P=0.019). However, there was no statistically significant difference in the homogeneity index (HI).Conclusion The optimization of rectal cancer planning using the RapidPlan automatic planning model for cervical cancer yields better results compared with manual planning, and the cross disease use of the model has certain feasibility in clinical practice.