Automated Instead of Manual Treatment Planning? A Plan Comparison Based on Dose-Volume Statistics and Clinical Preference

Purpose: Automated planning aims to speed up treatment planning and improve plan quality. We compared manual planning with automated planning for lung stereotactic body radiation therapy based on dose-volume histogram statistics and clinical preference. Methods and Materials: Manual and automated intensity modulated radiation therapy plans were generated for 56 patients by use of software developed in-house and Pinnacle 9.10 Auto-Planning, respectively. Optimization times were measured in 10 patients, and the impact of the automated plan (AP) on the total treatment cost was estimated. For the remaining 46 patients, each plan was checked against our clinical objectives, and a pair-wise dose-volume histogram comparison was performed. Three experienced radiation oncologists evaluated each plan and indicated their preference. Results: APs reduced the average optimization time by 77.3% but only affected the total treatment cost by 3.6%. Three APs and 0 manual plans failed our clinical objectives, and 13 APs and 9 manual plans showed a minor deviation. APs significantly reduced D2% (2% of the volume receives a dose of at least D2%) for the spinal cord, esophagus, heart, aorta, and main stem bronchus (P <.05) while preserving target coverage. The radiation oncologists found > 75% of the APs clinically acceptable without any further fine-tuning. Conclusions: APs may help to create satisfactory treatment plans quickly and effectively. Because critical appraisal by qualified professionals remains necessary, there is no such thing as fully automated planning yet. (C) 2018 Elsevier Inc. All rights reserved.