Evaluation of the complexity of treatment plans from a national IMRT/VMAT audit - Towards a plan complexity score.

PURPOSE

This work aimed to characterize and compare the complexity of the plans created in the context of a national IMRT/VMAT audit. A plan complexity score is proposed to summarize all the evaluated complexity features.

MATERIALS AND METHODS

Nine complexity metrics have been computed for the audit plans, evaluating different complexity aspects. An approach based on Principal Component Analysis was followed to explore the correlation between the metrics and derive a smaller set of new uncorrelated variables (principal components, PCs). The resulting PCs were then used to calculate a plan complexity score. Plan quality was also assessed and the correlation between plan complexity, quality and deliverability investigated using the Spearman's rank correlation coefficient.

RESULTS

The first two PCs explained over 90% of the total variance in the original dataset. Their representation allowed to identify patterns in the data, namely a clear separation between plans created using different technologies/techniques. The calculated plan complexity score quantified these differences. Sliding window Eclipse plans were found to be the most complex and VMAT Eclipse group presented the highest complexity variability, for the evaluated parameters. Concerning plan quality, no differences between treatment technology/technique have been identified. However, plans with larger number of monitor units tended to be associated with higher deviations between calculated and measured doses.

CONCLUSIONS

The proposed plan complexity score allowed to summarize the differences not only inter- but also intra-groups of technologies/techniques, paving the way for improvement of the planning strategies at the national level through knowledge sharing.