Research project Multi-factorial patient selection for proton radiotherapy
Multi-factorial patient selection for proton radiotherapy – modelling, software development and clinical applications.
Proton radiotherapy is available for Swedish patients at the national proton radiotherapy centre, Skandion Clinic, operated together by the seven university hospitals in Sweden. Skandion Clinic has opened its doors for patients in August 2015. More than 1400 patients remitted to Skandion Clinic from all over Sweden have received proton therapy. Currently, the selection of the patients for proton therapy is based on the assumed better physical conformity of the proton plans in comparison to photons. Proton treatments are therefore mostly intended for paediatric patients and CNS cancer patients. There is lack of standards for patient selection for proton treatments and, in addition, there are no standardised criteria for comparison between proton and photon treatment plans. The selection of patients for proton therapy is therefore based on clinical and empirical considerations building up on the previous experience on proton therapy from the The Svedberg Laboratory.
The present research application aims at developing a testbed for patient selection for proton therapy beyond the current clinical practice based on empirical knowledge and testing it for feasibility.
The specific aims of the project listed as work packages are:
WP1. To develop a model for accounting for the clinical benefit of using protons versus photons in terms of complication free tumour control probability
WP2. To implement the model into a research version of a treatment planning system and perform in silico testing
WP2. To test the clinical feasibility of the testbed on a group of head and neck (H&N) carcinoma patients
WP3. To design a pilot study for the validation of the testbed
Project description
The most commonly quoted advantage of proton therapy over photon therapy is the potential of reducing the toxicity in normal tissue and the organs at risk (OARs). This is based on the fact that proton treatment optimisation succeeds at producing plans that ensure good coverage of the target and steep gradients of the dose outside the target thus leading to a better sparing of the normal tissue and the organs at risk and therefore lower deterministic toxicity. Based on the physical properties of protons, it is therefore expected that the radiation-induced side effects will be reduced. Less discussed, however, is the potential advantage of using protons to escalate the dose in some cases that might benefit from it such as targets close to the organs at risk or generally expected to be radioresistant. Nevertheless, proton therapy appears to be a beneficial for both increasing the control of the tumour and decreasing the toxicity of the normal tissue. The ultimate proof that this is the case, would be expected to result from conducting randomised controlled trials (RCT) in order to investigate if indeed the dose escalation allowed with protons for the same or reduced dose to the critical structures and thus decreased toxicity would also result in better local control, increased overall survival, lower radiation-induced side effects and better quality of life for patients. There are, however, very few, if any, clinical trials comparing photons and protons. One of the reasons for the limited number of direct comparisons in the form of clinical trials between proton and photon radiotherapy is fact that the ethical character of such an RCT is questionable. Thus, the in silico planning studies corroborated with the known physical properties of protons make questionable the choice of randomising patients in two arms of a trial not fulfilling the prerequisite of being equipoised, thus knowing a priori with considerable certainty that is biased in favour of one arm, namely the proton arm. Another frequently quoted reason for the low number of protons versus photons RCT is the fact that these two techniques are notoriously fast developing and therefore an RCT having as primary endpoint the late toxicity might not reach the expected relevance because of the long duration. Clinical trials might therefore not be the optimal choice for answering the questions regarding the selection of the patients that might benefit most of proton therapy. An alternative to clinical trials was therefore sought and proposed in the Netherlands and later embraced by proton centres in other countries. The Dutch model-based approach has been adopted by the Health Council of the Netherlands and it aims at introducing a method for selecting the patients that will benefit from protons in terms of prevention of side effects. The model-based approach consists of two consecutive phases. The first phase aims at the selection of patients who may benefit from protons compared to photons in terms of reduction of the NTCP, ΔNTCP. The second phase aims at the clinical validation of proton therapy by so-called sequential prospective observational cohort (SPOC) studies using appropriate historical comparisons as a reference or by RCT’s in selected situations.
Thus, the state-of-art models for patient selection for proton therapy are restricted to the frame of NTCP reduction assuming that the coverage of the target and hence the probability of controlling the tumour is the same for photons and protons.
The present projects aims at developing the framework for the selection of the patients and the treatment plans not only in terms of normal tissue complication probability reduction but also in terms of increased probability of tumour control by using the concept of complication free tumour control previously introduced.
Furthermore, all the previous studies employing the Dutch ΔNTCP model, assumed the standard constant relative biological effectiveness (RBE) of protons of 1.1 as recommended by the International Committee for Radiation Units - ICRU) (ICRU 2007). This, however, might be a problem, because these is increasing evidence that the RBE of protons is variable with the linear energy transfer (LET) of the protons, the dose per fraction and the type of tissue. The present project will therefore consider the potential variation of the RBE for protons in addition to the standard constant 1.1 value.
Project members
Project managers
Iuliana Livia Dasu
Professor
Members
Suryakant Kaushik
PhD student