Click here to


Are you sure ?

Yes, do it No, cancel

Implementation of a Novel Dose and Anatomy Features-Driven Automated IMRT/VMAT Inverse Planning

Y Yang*, N Kovalchuk, L Xing, Stanford Univ School of Medicine, Stanford, CA


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: implement and evaluate an in-house-developed automated treatment planning tool with a novel dose-anatomy guided voxelization scheme in a commercial treatment planning system (TPS).

Methods: sparse voxelization scheme, characterized by partitioning the voxels into subgroups according to their geometric, anatomical and dosimetric features, is implemented in an in-house-developed automated planning solution to represent prior knowledge and to facilitate the plan optimization. A scripting application programming interface is employed to interact with a commercial TPS to implement automatic plan evaluation and update optimization parameters. The automated planning performs in an iterative fashion until reaching an acceptable tradeoff among target coverage/dose homogeneity and sparing of critical organs at risk. In each iteration, the featured voxel subgroups are updated based on current dose distributions. To evaluate the auto-planning solution performance, 20 previously treated prostate and H&N cases were randomly selected and their differences in target and organ-at-risk metrics from the clinical plans were analyzed to evaluate clinical acceptability of the proposed automated planning approach.

Results: H&N plans, with same target coverage and similar dose homogeneity (average global Dmax 107.3% vs. 107.4% for manual plans), our approach reduced maximum doses to brainstem and cord by 5.1±2.6Gy and 2.9±1.4Gy (p<0.03), respectively. The mean contralateral parotid, ipsilateral parotid, contralateral submandibular gland, pharynx, esophagus, cochlea doses were also reduced by 2.2 ± 2.9Gy, 4.8±4.7Gy, 3.6?±5.2Gy, 2.0±7.1Gy, 3.9±2.6Gy, 3.8±5.0Gy (p<0.045), respectively. Similar results were obtained for the prostate cases. With the similar PTV global Dmax (106.5% vs. 106.8% for manual plans), our approach achieved better sparing for both bladder and rectum.

Conclusion: automated planning solution is capable of efficiently generating treatment plans with plan quality not inferior to manual plans. It is integrated within a commercial TPS platform and would be feasible to adopt it into standard workflow to improve plan quality and treatment planning efficiency.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email