Dynamic evaluation of 4D robust optimisation for motion management in scanned proton therapy of hepatocellular carcinoma

Abstract

Pencil beam scanning (PBS) proton therapy is due to its steep dose gradients a promising treatment option for many cancer patients, particularly if the tumour is surrounded by critical organs. Currently, organ motion often poses an obstacle for the irradiation of tumours in the thorax or abdomen since the interference of pencil beam and tumour motion can lead to distortions of the intended dose distribution and requires dedicated motion mitigation techniques. A technique which has been recommended by recently published guidelines to reduce motion effects in PBS therapy of thoracic malignancies is 4D robust optimisation. 4D robust optimisation uses time-resolved CT images from different respiratory phases to incorporate uncertainties due to physiologic motion in treatment plan optimisation. Its effect has not yet been sufficiently examined and further research is needed to establish it in clinical routine. Within this thesis, the utility of 4D robust optimisation to mitigate motion effects was investigated for hepatocellular carcinoma, representative for mobile abdominal targets. For this purpose, a 4D dynamic dose calculation routine including an empirical beam delivery time model was developed based on a non-clinical, preliminary script from a global software provider for radiation therapy. The customised 4D tool enables prospective estimations before treatment, 4D dose reconstruction after treatment and comprehensive in silico examinations for research purposes. The validity of the routine and the time model was confirmed for typical beam settings in clinical end-to-end tests. In the treatment planning study on 4D robust optimisation, the routine revealed no significant improvement of the target coverage under motion for the investigated patients. A great advantage over conventional PBS plans was, however, that the same level of robustness against motion was reached for lower organ at risk doses. Since dose inhomogeneities averaged out for the applied fractionation scheme of 15 fractions, 4D robust optimisation was found to generate clinically acceptable sum plans. Due to the unknown biological effect of strong over- and underdosage in a single fraction, the method should be supplemented by further motion mitigation techniques in case of large tumour motion. The use of computer generated CTs for more comprehensive 4D evaluations was successfully tested in a proof of concept study and will allow to model a broad range of clinical cases in upcoming research projects. Thus, the benefit of 4D robust optimisation can be systematically investigated on a large scale for different subgroups. Although the elaboration of operating procedures for the irradiation of moving targets at the West German Proton Therapy Centre Essen (WPE) is still in process, the 4D dynamic dose calculation routine is already in clinical use to assess potential motion effects for current patients with minor tumour motion.