Department of radiation oncology, Fudan University Shanghai Cancer Center, Shanghai, China.
Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China.
Med Phys. 2017 Jun;44(6):2515-2531. doi: 10.1002/mp.12232. Epub 2017 Apr 20.
To develop a new automated treatment planning solution for breast and rectal cancer radiotherapy.
The automated treatment planning solution developed in this study includes selection of the iterative optimized training dataset, dose volume histogram (DVH) prediction for the organs at risk (OARs), and automatic generation of clinically acceptable treatment plans. The iterative optimized training dataset is selected by an iterative optimization from 40 treatment plans for left-breast and rectal cancer patients who received radiation therapy. A two-dimensional kernel density estimation algorithm (noted as two parameters KDE) which incorporated two predictive features was implemented to produce the predicted DVHs. Finally, 10 additional new left-breast treatment plans are re-planned using the Pinnacle Auto-Planning (AP) module (version 9.10, Philips Medical Systems) with the objective functions derived from the predicted DVH curves. Automatically generated re-optimized treatment plans are compared with the original manually optimized plans.
By combining the iterative optimized training dataset methodology and two parameters KDE prediction algorithm, our proposed automated planning strategy improves the accuracy of the DVH prediction. The automatically generated treatment plans using the dose derived from the predicted DVHs can achieve better dose sparing for some OARs without compromising other metrics of plan quality.
The proposed new automated treatment planning solution can be used to efficiently evaluate and improve the quality and consistency of the treatment plans for intensity-modulated breast and rectal cancer radiation therapy.
开发一种新的用于乳腺癌和直肠癌放射治疗的自动化治疗计划解决方案。
本研究中开发的自动化治疗计划解决方案包括迭代优化训练数据集的选择、危及器官（OAR）的剂量体积直方图（DVH）预测以及临床可接受的治疗计划的自动生成。通过对接受放射治疗的左乳腺癌和直肠癌患者的 40 个治疗计划进行迭代优化，选择迭代优化的训练数据集。实现了一种二维核密度估计算法（记为两个参数 KDE），该算法结合了两个预测特征，以生成预测的 DVH。最后，使用来自预测的 DVH 曲线的目标函数，使用 Pinnacle Auto-Planning（AP）模块（版本 9.10，Philips Medical Systems）对另外 10 个新的左乳腺癌治疗计划进行重新规划。自动生成的重新优化治疗计划与原始手动优化计划进行比较。
通过结合迭代优化训练数据集方法和两个参数 KDE 预测算法，我们提出的自动化规划策略提高了 DVH 预测的准确性。使用预测的 DVHs 中得出的剂量自动生成的治疗计划可以在不影响其他计划质量指标的情况下，更好地为某些 OAR 提供剂量保护。
所提出的新的自动化治疗计划解决方案可用于有效地评估和改善强度调制乳腺癌和直肠癌放射治疗的治疗计划的质量和一致性。
