Head and neck plan quality: where we are and does it matter?

Radiation treatment planning for the Head and Neck (HN) cancers is challenging and complex, and here we investigate various aspects of HN plan quality. We present the results of an international plan quality study, in which a web-based platform (ProKnow) was used to analyse a total of 238 plans submitted from 34 countries, all for the same dataset, anatomy, and quantified, objective plan scoring algorithm. All treatment planning systems (TPS) studied were able to produce high quality plans: 6 of 6 TPS models had scores in the top 25%, and 4 of those in the top 10%. However, all TPS and modalities also showed substantial variability in plan quality across all planners, suggesting needs for training and/or automation. Concerning current automation solutions, both commercially available approaches (AutoPlanning and Knowledge-Based Planning) are so far unable to outperform experienced human planners when tight target dose coverage and homogeneity are required. Target dose homogeneity is not to be discounted, as it may be linked with the rate of complications such as reactive gastrostomy tube placement after oropharynx chemoradiation.