Towards Interventional Particle Therapy for Advanced Cancer Care

Introduction

Particle therapy (PT) has advantageous physical properties for radiotherapy when compared to conventional X-rays. Instead of depositing dose in tissue with approximately exponential decay, in PT the majority of the dose is deposited in the region defined by the Bragg Peak. By knowing the precise 3D position of the tumour over time, the beam can be shaped in a way that exposes the whole tumour, while minimizing the dose delivered to the surrounding normal tissue.

However, despite clear physical advantages, the clinical evidence for the benefits of PT is still lacking. Apart from some cases, such as in paediatrics, studies show either minor clinical benefit of PT or in some cases even no benefit at all. The higher cost of a PT treatment is thus put into question and the exponential growth of PT treatment rooms may come to a halt without some breakthrough in PT cancer care.

The RAPTOR Consortium

The Real-time adaptive PT of cancer (RAPTOR) initiative was started by Massachusetts General Hospital (USA), University of Ljubljana (Slovenia) and Cosylab (Slovenia) and consists of PT academic and research institutes, clinics and companies around the world. Currently, the members are:

• Aarhus University (Denmark)
• Centro Nazionale di Adroterapia Oncologica (Italy)
• GSI Helmholtz Centre for Heavy Ion Research (Germany)
• LMU Munich (Germany)
• MedAustron (Austria)
• MedPhoton (Austria)
• National Institute of Radiological Sciences (Japan)
• OncoRay (Germany)
• Paul Scherrer Institut (Switzerland)
• ProtonVDA (Cosylab)
• RaySearch (Sweden)
• University Medical Center Groningen (Netherlands)
• Tilburg University (Netherlands)

Founding committee:

• Massachusetts General Hospital (United States of America)
• University of Ljubljana (Slovenia)
• Cosylab (Slovenia)

The initial meeting was hosted by Cosylab in July 2018 in Ljubljana, Slovenia, where the participants discussed the current state of adaptive PT and the future steps necessary to realize interventional PT for advanced cancer care.

Towards interventional PT

The high precision of PT comes as a double-edged sword, since PT is normally less robust than X-ray radiotherapy. Several uncertainties, such as changes in anatomy, positioning, organ delineation and systematic uncertainties can have a significant impact on where the final dose is delivered. The reliability of PT has increased in recent years by the robust treatment planning, however, it still remains sensitive to larger uncertainties which have to be minimized to exploit the full benefit of PT.

The clinical workflow in PT has been adopted from conventional X-ray radiotherapy, where the treatment plan is based on the initial computed tomography (CT) scan of a patient. Since, the treatment usually lasts several weeks, it is likely that the initial treatment plan becomes less valid due to the changes of the patient anatomy as the treatment progresses.

With RAPTOR, we propose a shift in the paradigm by preparing a daily PT treatment plan for each patient, that is based on the current patient images. This would significantly increase treatment accuracy, since most of the uncertainties would be minimized. The advantages of PT could thus be better used in an interventional capacity.

Due to its flexibility and rapid adaptation, RAPTOR will enable patient treatments in the sitting and standing position, which will facilitate much simpler, more compact, and less costly beamline designs. RAPTOR will therefore make the benefits of proton therapy available to many more patients.

The proposed changes in the clinical workflow would require different levels of process automation, which would increase the cost-effectiveness of a PT treatment. Furthermore, the RAPTOR initiative is driven by institutes and clinics from all around the world, which would require harmonization and possibly even standardization of processes, making comparisons between PT clinics easier.

By customising the dose to the present patient image, the margins around the tumour could be reduced. The dose to the tumour could then be increased, without any additional dose delivered to the surrounding normal tissue. This should increase the control of the tumour and subsequent patient outcome.

To turn interventional PT into a clinical reality another innovation is needed – treatment verification. Since the treatment plans will be generated on a daily basis, a part of the quality assurance (QA) will have to shift from before the treatment to after or even during the treatment itself. Such verification would enable a more precise analysis, because an actual dose plan would be available to the clinician not only a simulated one.

Work Packages

The development of the RAPTOR consortium research proposal will be spread over three main work packages that all revolve around the central workflow (Figure 1):

• Adaptive intervention
• In-vivo treatment verification
• Integration and interfaces

Adaptive Intervention

While the normal clinical routine usually follows the same path (imaging, planning, QA, delivery), the time scales are significantly different in adaptive PT. Several changes will have to be introduced to enable interventional delivery, e.g. adjustable optimization of treatment plan, rapid QA and flexible beam delivery.

Adjustable Optimization of Delivery

Currently the treatment plan is based on a single, initial patient CT scan that is rarely adapted during the course of the treatment. Dedicated research will focus on optimizing the treatment plan based on the day-to-day patient images from different imaging modalities, such as cone-beam CT, dual-energy CT and magnetic resonance imaging. Furthermore, with the help of radiomic features of previously treated patient cohorts we aim to refine and personalize treatment adaptation schemes.

Rapid QA

Since the treatment plan will be optimized just prior to the patient irradiation, the treatment QA has to be adapted accordingly. Special automated QA techniques will have to be developed to ensure safe and precise treatment delivery in significantly less time than is currently done.

Flexible Beam Delivery

Similar to QA, the beam delivery is currently focused on delivering a specific plan with almost no room for adaptation. Dedicated research will focus on making beam delivery flexible enough to offer a variety of options for fast plan adaptation.

In-vivo Treatment Verification

For unambiguous treatment verification several techniques will have to be refined, further improved concerning data acquisition and data analysis and finally integrated in the clinical workflow.

PT Compatible Imaging Protocol

Any adaptive technique will rely heavily on knowledge of patient anatomy during treatment. There are two tasks related to imaging in adaptive PT:

• provide reference information for dose delivery estimates (either via range measurement techniques and/or log-file based algorithms) to update delivered dose log.
• provide anatomical details as input for daily plan adaptations

The target of RAPTOR is to determine specifications for both imaging tasks and optimized inclusion of such imaging in PT treatment protocol.

Range Measurement

One of the essential elements of treatment verification is an accurate measurement of particle range in a patient. Particle range is carefully simulated in the treatment planning process; however, it is subject to substantial uncertainties and in addition very sensitive to anatomical changes and a verification of particle range in the actual patient is crucial for treatment success. The most promising techniques for range verification are prompt gamma detection, gamma spectroscopy or a combination of the above. Furthermore, so-called range-probes, using the particle beam at high energies shooting through the patient and measuring its remaining range, is a further promising technique that has not yet translated into clinical application. These techniques will be investigated, further developed, validated and integrated in a clinical-realistic workflow.

Accumulated Dose

Radiation doses delivered to all parts of the patient will be recorded and accumulated at every time point during the course of the treatment, so that a complete, patient-specific profile of the entire treatment will be generated.

Integration and Interfaces

In addition to dedicated research into specific parts of the proposed workflow, a seamless integration between all parts is essential for a fast and responsive system.

Standardization of Interfaces and Documentation

All interfaces between different parts of all equipment should be standardized to enable easier integration, comparison and cost reduction. Similarly, documentation should be available in a standardized format, which will also enable easier inter-facility comparison.

Validation

The proposed changes will change the established clinical workflow and accelerate most of the workflow. Hence, a new procedure has to be developed, which will rigorously validate the entire system. Each component must be safe by itself as well when it is a part of the full system with all components integrated. Additionally, a realistic, anatomical 4D phantom for precise dynamic end-to-end verification will be developed.

Data Analysis

With the standardization of documents and availability of the precise dose accumulation, machine learning can be used for data analysis. Patient data, imaging data,  and patient outcomes together with machine, range verification and clinical workflow data can be analysed and the results can be used to further optimize the adaptive workflow.

Conclusion

The RAPTOR initiative will develop an adaptive PT approach that results in a dynamic treatment plan based on the individualistic changes in patient anatomy. Verification will be available after each treatment for inspection and analysis and feedback into the dynamic adaptation process. The impact of anatomical changes and other uncertainties on PT will thus be significantly reduced. The system will improve both the efficacy and efficiency of PT. The development of new workflows will allow extension of indications that can be treated with PT and thus enable PT for a wider range of patients.