Background/Objectives: To evaluate an end-to-end pipeline for normo-fractionated prostate-only and whole-pelvic cancer treatments that requires minimal human input and generates a machine-deliverable plan as an output.Methods: In collaboration with TheraPanacea, a treatment planning pipeline was developed that takes as its input a planning CT with organs-at-risk (OARs) and planning target volume (PTV) contours, the targeted linac machine, and the prescription dose. The primary components are (i) dose prediction by a single deep learning model for both localizations and (ii) a direct aperture VMAT plan optimization that seeks to mimic the predicted dose. The deep learning model was trained on 238 cases, and a held-out set of 86 cases was used for model validation. An end-to-end clinical evaluation study was performed on another 40 cases (20 prostate-only, 20 whole-pelvic). First, a quantitative evaluation was performed based on dose–volume histogram (DVH) points and plan parameter metrics. Then, the plan deliverability was assessed via portal dosimetry using the global gamma index. Additionally, the reference clinical manual plans were compared with the automated plans in terms of monitor unit (MU) numbers and modulation complexity scores (MCSv).Results: The automated plans provided adequate treatment plans (or minor deviations) with respect to the dose constraints, and the quality of the plans was similar to the manual plans for both localizations. Moreover, the automated plans showed successful deliverability and passed the portal dose verification. Despite higher median total MUs, no statistically significant correlation was observed between any of the gamma criteria tested and the number of MUs or MCSv.Conclusions: This study shows the feasibility of a deep learning-based fully automated treatment planning pipeline that generates high-quality plans that are competitive with manually made plans and are clinically approved in terms of dosimetry and machine deliverability.
背景/目的：评估一种用于常规分割前列腺癌局部及全盆腔放疗的端到端计划流程，该流程需极少人工干预，并能生成可直接用于机器执行的治疗计划。 方法：与TheraPanacea公司合作开发了一套治疗计划流程。该流程以包含危及器官（OAR）和计划靶区（PTV）轮廓的定位CT图像、目标直线加速器型号及处方剂量作为输入。其核心组件包括：（i）针对两种放疗范围（局部与全盆腔）的单一深度学习剂量预测模型；（ii）旨在复现预测剂量的直接孔径容积旋转调强放疗（VMAT）计划优化算法。深度学习模型基于238例病例进行训练，并使用86例独立病例进行验证。随后，对另外40例病例（20例前列腺局部，20例全盆腔）进行了端到端临床评估研究。首先，基于剂量体积直方图（DVH）参数和计划质量指标进行定量评估；其次，通过射野剂量验证采用全局伽马指数评估计划的可执行性；此外，还将自动计划与临床参考人工计划在机器跳数（MU）和调制复杂度评分（MCSv）方面进行了比较。 结果：自动计划在剂量约束方面均达到可接受水平（或仅存在微小偏差），两种放疗范围下的计划质量均与人工计划相当。此外，自动计划展现出良好的可执行性，均通过了射野剂量验证。尽管自动计划的中位总机器跳数较高，但在所测试的伽马标准与机器跳数或MCSv之间未观察到统计学显著相关性。 结论：本研究证明了基于深度学习的全自动治疗计划流程的可行性。该流程能够生成高质量的治疗计划，其质量与人工计划相当，且在剂量学与机器可执行性方面均达到临床批准标准。
肿瘤(癌症)患者之家