ALLEGATO 2. Hadrontherapy for the treatment of cancer: a systematic review of safety and effectiveness.

ALLEGATO 2 Hadrontherapy for the treatment of cancer: a systematic review of safety and effectiveness. 1

Contributions Authors: (in alphabetical order) Fedele Bonifazi, Laura Camilloni, Angelo Capizzi, Francesco Cardinale, Anna Cavazzana, Elisa Giani, Susanna Maltoni, Giovanni Mastrandrea, Horand Meier, Massimiliano Orso, Nicola Pace, Valeria Romano, Sergio Sassano RIHTA rete italiana HTA External reviewer: Leslie Levin, MB MD FRCP (Lon) FRCPC Head, medical Advisory Secretariat Health Quality Ontario 2

Contents Abstract... 4 Background... 4 Objectives... 4 Data Sources... 4 Review Methods... 4 Results... 5 Limitations... 5 Conclusions... 5 Background... 7 Objective... 8 Evidence-Based Analysis: Methods... 8 Team-work and roles... 8 Literature search and appraisal... 8 Evidence-Based Analysis: Results... 9 Secondary literature search... 9 Table 1. Secondary literature search, included studies:... 10 Primary literature search... 11 Safety and Effectiveness assessment... 12 A. Skull base tumours and others head and neck tumours... 12 Results... 12 Conclusions on skull base tumours and other head and neck tumours... 14 B. Ocular neoplasms... 15 Results... 15 Conclusions on ocular neoplasms... 20 C. Non-small cell lung cancer (NSCLC)... 22 Results... 22 Conclusions on NSCLC... 24 D. Gastrointestinal cancer... 24 Results... 25 Conclusions on gastrointestinal cancer... 26 E. Pelvic cancer... 26 Results... 26 Conclusions on pelvic cancer... 27 F. Prostate cancer... 27 Results... 28 Conclusions on prostate cancer... 31 G. Skeleton and soft tissues tumours... 31 Results... 31 Conclusions on skeleton and soft tissues tumours... 32 Economic aspects... 33 Results... 33 Conclusions on economic aspects... 34 Discussion... 35 Conclusions and recommendations... 36 Appendix 1. Secondary literature search, included studies... 37 Appendix 2. Secondary literature search, excluded studies... 38 Appendix 3. Quality assessment (through AMSTAR checklist) of included HTAs and SRs... 39 Appendix 4. Primary literature search, excluded studies... 40 Appendix 5. References... 41 Appendix 6. Epidemiological elements... 45 3

Abstract Background Cancer is one of the leading causes of death worldwide. Therapies for cancer include surgery, chemotherapy, and radiation therapy (RT). Hadrontherapy (HT) is a form of RT that relies on protons and/or carbon ions beam. Because of the nature of protons and carbon ions, accuracy of HT is theoretically much higher than that of traditional radiotherapy, representing a technology that might improve tumour control due to greater precision in radiation dose delivered to the tumour while reducing adverse effects due to reduced radiation dose to surrounding tissues. Tumours for which HT could yield to better results than other radiotherapies include tumours of head and neck, eye, liver, lung, prostate, pancreas, soft tissue, skeleton, gastrointestinal system and pelvis, which all together are responsible of approximately 80,000 new cases per year in Italy. Objectives Objective of the analysis is the assessment through a systematic review of available literature of HT for the treatment of cancer in terms of relative safety (decrease in adverse effects) and relative effectiveness (local control and overall survival) in relation to other RT techniques. Data Sources The following databases and web sites were searched for secondary literature: The Cochrane Database of Systematic Reviews; Database of Abstracts of Reviews of Effects (DARE); Health Technology Assessment (HTA) database; NHS Economic Evaluation Database (NHS EED); TripDatabase; INAHTA and AHRQ web sites. The following databases and web sites were searched for primary literature in order to update the secondary literature search through RCTs: Pubmed; Clinicaltrials.gov; Controlled-trials.com; Cochrane Central Register of Controlled Trials. Only Randomized Controlled Trials were included in the primary literature search. Review Methods Outcomes of interest regarded safety (decrease in adverse effects) and effectiveness (local control and overall survival). Relevant studies were shared in the working group in order to extract data and evaluate the methodological quality of documents using criteria from the AMSTAR checklist. At least two researchers read each document and reached a consensus. Limits were applied for publication year: 2007 - present. Duplicates were removed. Only HTA reports and Systematic Reviews were included in the secondary literature hits. Only, Italian, French and German languages were considered. In addition, the working group decided to update the evidence, performing a literature search for primary literature, limited to RCTs in order to evaluate if recent evidence could change the conclusions of secondary literature. A literature search was not undertaken for economic analyses since the latter was not an objective of this review. 4

Results From the secondary literature search, 33 publications were retrieved: 15 documents met the inclusion criteria and were included. From the primary literature search, 9 publications were retrieved: no documents met the inclusion criteria and were included. Published studies were limited to specific disease sites which appeared in the reports; these were: Skull base tumours and others head and neck tumours: there is no clear clinical evidence to assess the clinical superiority in efficacy between HT and standard therapies. In the absence of well-designed trials no definitive conclusions can be drawn. Ocular tumours: secondary studies differed in disease, intervention and the design of primary studies included. No conclusions could be drawn. There is limited evidence on the effectiveness and safety of HT for ocular tumours due to the lack of well-designed and well-reported studies. The limited studies done indicate a potential improvement in local control with Proton radiotherapy for uveal melanoma if close to optic disk or fovea or thickness >5 mm. Lung cancer: One of the studies with highest quality of evidence is a meta-analysis for the treatment of stage I NSCLC patients which included 8 studies of low quality. The meta-analysis demonstrates that survival rates for particle therapy are higher than those for conventional radiotherapy but similar to those achieved with stereotactic radiotherapy. No firm conclusions can be drawn concerning adverse events. In general, all documents underline that studies concerning efficacy are non-comparative and have low methodological quality, thus the efficacy of the use of this technology for these patients remains unclear. Gastrointestinal tumours: The small number of patients precludes firm conclusions so that the role of HT in GI cancers remains unclear. Pelvic tumours: No firm conclusions can be drawn due to the small number of patients. In general terms, there are no definitive conclusions on the relative efficacy of HT for pelvic tumours. Prostate cancer: all documents agree that currently available evidence is insufficient to demonstrate improved outcomes for HT compared to alternative treatments. Skeleton and soft tissues tumours: due to the lack of comparative data, there is no evidence of improved patient outcomes for HT. Limitations HT shows continuous hardware and software evolutions, leading to variations in HT practice. It is not clear how such variations could have affected clinical outcomes. The quality of the included systematic reviews was assessed by the Amstar checklist and discussed but was not reported formally. Conclusions The lack of ad-hoc designed comparative studies makes it impossible to draw firm conclusions about the effects of HT for cancer treatment. In specific circumstances, clinical studies suggested an increase of safety 5

and effectiveness by using HT instead of traditional RT for some type of tumours (uveal melanoma, skull and neck chordomas, and NSCLC). Nonetheless, there is uncertainty regarding these estimates, due to methodological and design biases. While waiting for and contributing to the availability of new high quality evidence on effectiveness and safety of HT, the currently operating/in development Italian centres should give priority to patients with the above mentioned pathologies (uveal melanoma, skull and neck chordomas, and NSCLC). No new HT centre should be planned in Italy, without new scientific evidence proving its need. 6

Background Cancer is one of the major causes of death worldwide: in 2008, it accounted for 13% of all deaths (WHO 2011). The incidence of cancer was 11.3 million in 2007 and is expected to increase to 15.5 millions by 2030 (WHO 2011). In Italy, cancer accounts for approx. 122.000 deaths per year, with 250.000 new cases per year (Airtum, 2009). Therapies for cancer include surgery, chemotherapy, radiation therapy (RT), or a combination of each, according to the tumour type, stage and location. RT is defined as the medical use of ionizing radiation (IR) for cancer treatment. Target tissue is exposed to a certain IR dose in order to kill malignant cells and increase local cancer control rate (LC). Particles traditionally used in RT include photons, electrons, and gamma rays: due to the nature of their interaction with matter, to deliver a certain amount of dose to the tumour, part of the radiation dose will be absorbed by the tissues that surround the target volume. Although advanced techniques are used for dose planning and delivering to minimize exposure of non-malignant tissues, adverse effects related to radiation exposure are common. According to the part of the body treated, acute side effects of RT include local necrosis, diarrhoea, fatigue, hair loss, nausea and vomiting, mucositis, changes in sexuality and fertility, and damage to skin, throat, urinary tract and bladder. Late side effects include infertility, lymphedema, changes in the brain causing cognitive dysfunction, and changes in the joints and mouth, in addition to increasing the risk of developing secondary cancer (NCI 2011). Such adverse effects may lead to dramatic consequences, especially when the damaged non-malignant tissue represents a critical part of the body such as nerves, brain, glands and eyes. In the last decades a new form of RT, named Hadrontherapy (HT), has been introduced. HT relies on the use of protons and carbon ions instead of other particles. The particular interaction of protons and carbon ions with matter makes it possible to concentrate the delivered dose to the target tissue with greater precision while decreasing the dose to surrounding tissues, therefore theoretically increasing effectiveness on malignant tissue and reducing adverse effects (including damages to surrounding organs). Such increase in precision is particularly important when target tumours are surrounded by important and radiation sensitive healthy tissues. This is why HT has claimed to potentially become the elective therapeutic strategy for some cases of ocular tumours, other head and neck tumours, and prostate cancer, especially in early stages. Lately, HT has been also suggested to be indicated for the treatment of some cases of cancers of lung, liver, pancreas, gastrointestinal system, and pelvis. All together, the previous pathologies account for almost 110,000 new cases of tumour per year (see details in Appendix 6). It is thus of primary importance to evaluate the safety, the effectiveness and the appropriateness of HT in order to drive its implementation and diffusion in Italy. Moreover, new promising forms of RT and advances in older techniques (such as for Intensity Modulated Radiation Therapy IMRT, Brachytherapy, Radiosurgery, etc.) associated with a higher therapeutic index are claiming to show increased advantages over traditional RT. Currently (December 2011), there are 38 HT facilities operating worldwide (Particle Therapy Co-Operative Group, 2011). In Italy, only one centre is operating (Catania, Sicily), treating only ocular tumours because of the low energy of the protons produced. Another centre is in the start-up phase (Pavia), and a third centre is under construction (Trento). The low uptake of HT worldwide may be a combination of factors including uncertainty of efficacy compared to more conventional approaches to RT, limited utility and high costs of HT facilities. The disproportionate increase in HT facilities across Italy relative to worldwide uptake is concerning and warrants this assessment of relative effectiveness and safety. 7

Objective The objective of this analysis is the assessment of HT (proton, ion and neutron beam therapy) in comparison with other RT techniques (conventional RT, IMRT, stereotactic surgery, brachytherapy) for the treatment of cancer in terms of relative safety (decrease in adverse effects) and effectiveness (local control and overall survival). An economic analysis was not included among the objectives of the present document. Evidence-Based Analysis: Methods The project was co-ordinated by the Italian National Agency for Regional Health Care Services (Agenzia Nazionale per i Servizi Sanitari Regionali Age.Na.S). Age.Na.S is developing a national network for HTA (RIHTA). The RIHTA Network identified a multi-regional team of contributors with different expertise that was involved in writing this report. The research question for this report was: what is the body of evidence about the use of HT for the treatment of cancer? The working team decided to perform first of all a search for secondary literature (HTA reports and Systematic Reviews) in order to collect available evidence on HT. In addition, the team decided to update the evidence, performing a literature search for primary literature, limited to RCTs to evaluate if recent evidence could change the conclusions derived from the secondary literature review. Team-work and roles Valeria Romano (VR) designed and performed the search strategies; Fedele Bonifazi (FB), Laura Camilloni (LC), Angelo Capizzi (ACZ), Francesco Cardinale (FC), Anna Cavazzana (AC), Elisa Giani (EG), Susanna Maltoni (SM), Giovanni Mastrandrea (GM), Massimiliano Orso (MO), Nicola Pace (NP), Sergio Sassano (SS) reviewed the efficacy of proton therapy. The project was managed by Horand Meier (HM). Literature search and appraisal The following databases and web sites were searched for secondary literature: The Cochrane Database of systematic reviews; Database of Abstracts of Reviews of Effects (DARE); Health Technology Assessment (HTA) database; NHS Economic Evaluation Database (NHS EED); TripDatabase; INAHTA and AHRQ web sites. Last search executed on 27 October 2011. The following search terms were included in all searches: "hadron therapy" OR hadrontherapy OR "proton beam" OR "charged particles" OR "proton therapy" OR "heavy ions therapy" OR "charged particles" OR protontherapy. Limits were applied for publication year: 2007 - present. Duplicates were removed. Only HTA reports, including a systematic review, and Systematic Reviews were included in the secondary literature search. Only, Italian, French and German languages were considered. Four researchers (AC, FB, NP and SM) independently screened abstracts and titles and included HTA reports and systematic reviews focusing on clinical effectiveness and/or adverse events for one or more diseases. 8

Relevant studies were shared in the working group in order to extract data and evaluate the methodological quality of documents using criteria from the AMSTAR checklist. Each document was read from at least two researchers and disagreements were discussed to reach a consensus. The following databases and web sites were searched for primary literature in order to update the secondary literature search through RCTs: Pubmed; Clinicaltrials.gov; Controlled-trials.com; Cochrane Central Register of Controlled Trials. It was not deemed necessary to investigate additional data sources. The search terms included were the same used for the secondary literature search. Last search executed on 29 November 2011. Limits were applied for publication year: 2007 to present. Duplicates were removed. Only Randomized Controlled Trials were included in the primary literature search. Only, Italian, French and German languages are considered. Two researchers (AC and MO) independently screened RCTs and included them on the basis of clinical effectiveness/efficacy and/or adverse events for one or more diseases. The quantity and quality of the evidence (secondary and primary) available to answer the questions on the performance of HT for cancer was summarised. Evidence-Based Analysis: Results Secondary literature search From the secondary literature search, 33 publications were retrieved: 15 documents met the inclusion criteria and were included. Results for secondary literature search (SRs, HTA reports including a SR): 33 publications retrieved 18 did not met the inclusion criteria: Intervention (7) Date of publication (6) SR not included (2) 15 publications included Duplicate (2) Language (1) Table 1 lists the characteristics of included studies (complete references can be found in Appendix 1); the list of excluded studies are reported in Appendix 2 along with the reasons for their exclusion. 9

Table 1. Secondary literature search, included studies: 1 st Author Year Agency Title Content Language Bannuru R.R. 2011 Samson D.J. 2011 Samson D.J. 2011 BLANCHARD S. 2010 HAS BlueCross Blueshield BlueCross Blueshield Comparative Evaluation of Radiation Treatments for Clinically Localized Prostate Cancer: An Updated Systematic Review Proton beam therapy for prostate cancer Proton beam therapy for non-smallcell lung cancer Carbon ion radiotherapy (Hadrontherapy) Flynn K. 2010 VATAP Proton beam therapy for cancer Grutters J.P. 2010 Amichetti M. 2009 Bekkering G.E. 2009 Trikalinos T.A. 2009 AHRQ Ollendorf D.A. 2008 Wilt T.J. 2008 Comparison of the effectiveness of radiotherapy with photons, protons and carbon-ions for non-small cell lung cancer: a meta-analysis Proton therapy in chordoma of the base of the skull: a systematic review The effectiveness and safety of proton radiation therapy for indications of the eye: a systematic review Particle Beam Radiation Therapies for Cancer Brachytherapy/proton beam therapy for clinically localized, low-risk prostate cancer Comparative effectiveness of therapies for clinically localized prostate cancer Cancers: - Prostate Cancers: - Prostate Cancers: - Lung Cancers: - Ocular - Head and neck - Skull base - Prostate - GI - Lung - Pelvic - Soft tissue sarcoma Cancers: - Ocular - Head and neck - Skull base - Prostate - GI - Lung Cancers: - Lung Cancer: - Skull base Cancers: - Ocular Cancers: - Ocular - Head and neck - Skull base - Prostate - GI - Pelvic Cancers: - Prostate Cancers: - Prostate French 10

1 st Author Year Agency Title Content Language Lee I. 2007 ANZHSN Lee I. 2007 ANZHSN Proton beam therapy for the treatment of neoplasms involving (or adjacent to) cranial structures Proton beam therapy for the treatment of uveal melanoma Huybrechts M. 2007 KCE Hadrontherapy Lodge M. 2007 A systematic literature review of the clinical and cost-effectiveness of hadron therapy in cancer Cancers: - Head and neck - Skull base Cancers: - Ocular Cancers: - Ocular - Head and neck - Skull base - Prostate - GI - Lung - Pelvic - Soft tissue sarcoma Cancers: - Ocular - Head and neck - Skull base - Prostate - GI - Lung - Pelvic Quality assessment of included HTAs and/or SRs was assessed through AMSTAR checklist and results are reported in Appendix 3. Primary literature search From the primary literature search, 9 publications were retrieved but none met the inclusion criteria and were therefore excluded. The list of excluded studies and the reasons for exclusion are reported in Appendix 4. Search for primary studies: 9 trials retrieved 9 did not met the inclusion criteria: On-going trial (5) Study protocol (2) Study design (1) Intervention (1) 0 trials included 11

Safety and Effectiveness assessment Relative safety (decrease in adverse effects) and effectiveness (local control and overall survival) of HT in comparison to other RT techniques (conventional RT, IMRT, stereotactic surgery, Brachytherapy) have been assessed in the following anatomic sites: A. Skull base, head and neck; B. Eye; C. Lung; D. Gastrointestinal system; E. Pelvis; F. Prostate; G. Skeleton and soft tissues. A. Skull base tumours and others head and neck tumours Seven studies from the secondary literature search included treatment of skull base tumours and six include head and neck tumours. Of the studies that included skull base tumours, four are HTA-reports (HTA-VATAP: Flynn Karen et al.; 2010, HTA-HAS: HAS; 2010, Blue Cross and Blue Shield Association TEC: Samson et al.; 2011, KCE-REORT: Huybrechts et al.; 2007), two are systematic reviews (Lodge et al., 2007; Amichetti et al, 2009) and one is a horizon scanning analysis (ANZHSN, 2007). The study of Amichetti et al. focused exclusively on skull base tumours. Four documents (AHRQ report, KCE report, ANZHSN report and Amichetti et al.) also include paediatric patients. Results Flynn K, 2010 Brief overview: proton therapy for cancer (VATAP - Veterans Health Administration Office of Patient Care Services Technology Assessment Program) The VATAP report shows results and conclusions of three systematic reviews, one horizon scanning report, one cross-sectional study and one case series study but doesn t provide a summary of the results. Blanchard S, 2010 Hadrontherapie par ions carbone - rapport préliminaire (Haute Autorité de Santé - HAS) The HAS report focuses on carbon-ion HT only and includes 18 documents: 3 recommendations, 1 descriptive systematic review, and 14 other publications (including 1 meta-analysis and representing 22 retrospective and prospective studies). Conclusions drawn from these studies suggest that carbon-ion HT: has a better performance than conventional radiotherapy as for adenoid cystic carcinoma, base of skull chordomas and chondrosarcomas, salivary glands tumours with incomplete resection; is equivalent to other radiotherapy techniques for base of skull chordomas and chondrosarcomas and other head and neck tumours; 12

can lead to late adverse events especially reported for skull base chordomas and chondrosarcomas. However these conclusions are based on low quality evidence reflecting uncertainty in the estimates of effectiveness. Trikalinos TA, 2009 Particle Beam Radiation Therapies for Cancer. Technical Brief No. 1. (Agency for Healthcare Research and Quality - AHRQ) The AHRQ report includes 243 papers; 56 studies describe treatment of cancers of the head and neck: 53 single-harm studies, 1 non-randomized comparative studies and 2 RCT. The report does not summarize the results. Huybrechts M. et al., 2007 Hadrontherapie (Federaal Kenniscentrum voor de Gezondheidszorg - KCE) The KCE report examines 45 documents about the use of proton or carbon ion therapy, with the following results: Skull base chordomas. Case series with proton beam therapy show high rates of local progressionfree survival and overall survival. The safety seems acceptable, even if there are insufficient data to compare toxicity induced by photon or proton beam therapy. Case series with carbon ion irradiation also show good results in chordomas of the skull base without serious toxic reactions. There are however no comparative studies between carbon ion therapy and (current) standard treatments. Skull base chondrosarcomas. In chondrosarcomas of the skull base, there are no differences between photon and proton irradiation in local control and overall survival rates. Currently, carbon ion therapy seems less effective than existing treatments. The included studies are small, heterogeneous, non-comparative case series reflecting low quality of evidence. Other intracranial tumours. There is no evidence for the use of proton or carbon ion therapy in the other malignant or not malignant intracranial tumours. Paediatric cranial tumours. Proton radiation therapy seems to be safe and well tolerated by children suffering from CNS tumours (no RCTs available, sparse retrospective evidence). Salivary gland tumours. One randomized controlled trial showed a significantly better local control rate - but no difference in survival rates - with neutrons (vs conventional photons therapy) in inoperable, unresectable or recurrent malignant salivary gland tumours. A comparative case series also found better local control (without any significant difference in survival rates) with carbon ion and photons than with photons alone in patients with cystic adenoid carcinoma. Lodge M, 2007 A systematic literature review of the clinical and cost-effectiveness of hadron therapy in cancer The systematic review by Lodge identifies two small retrospective low quality studies on protons for head and neck cancer, and two prospective studies and nine retrospective studies for central nervous system tumours. They found that in base of skull chordomas, both protons and C-ions appear to be superior to photons. In central nerve system tumours other than base of skull chordomas, the results of photons are similar to those of protons and ions. 13

Amichetti M, 2009 Proton therapy in chordoma of the base of the skull: a systematic review Amichetti conducted a systematic review specifically on the treatment of base of the skull chordoma with proton therapy compared to other radiation techniques (conventional radiation therapy, ion therapy, fractionated stereotactic radiation therapy, and radiosurgery). No prospective trials (randomized or nonrandomized) were found and the review consisted of seven uncontrolled single-arm studies, providing clinical outcomes for 416 patients in total: five studies consisted of adult and paediatric patients and two consisted of paediatric patients only. These reports were mainly based on advanced inoperable or incompletely resected tumours. The use of protons showed improved results compared to the reported use of conventional photon irradiation, resulting in the best long-term (10 years) outcome for this tumour with relatively few significant complications considering the high doses delivered with this therapeutic modality. Lee I, 2007 Proton beam therapy for the treatment of neoplasms involving (or adjacent to) cranial structures (Australian and New Zealand Horizon Scanning Network ANZHSN) The ANZHSN report reviews a total of 24 case series reporting safety and efficacy outcomes on the use of proton beam therapy for the treatment of patients with various head and neck tumours. Four comparative treatment planning (modelling) studies are also included. The report includes paediatric and adult head and neck tumours. Most included studies report local tumour control rates which are similar to conventional photon radiotherapy; the prevalence of proton radiation-induced side effects in the included studies appears to be within the range expected for conventional photon therapy, with some studies inferring that proton therapy is substantially safer. However the lack of consistency across studies and the lack of direct comparative studies severely limit the validity of these conclusions. Conclusions on skull base tumours and other head and neck tumours Some of the retrieved reports provide a conclusion on the assessment of efficacy and safety of HT specifically in skull-base and head and neck cancers (HAS, Lodge, KCE, Amichetti, ANZHSN), others (VATAP, AHRQ) only a general conclusion on its benefit-risk profile for all possible indications in cancer. All the documents highlight the need of well-designed randomised controlled trials to allow more reliable conclusions on efficacy and safety of HT in base of skull and head and neck cancers. VATAP report. For all the considered cancer indications (thus including also skull-base and head and neck neoplasm), authors judge that the literature reflects the early clinical investigation status for effectiveness of proton therapy and conclude that there are no indications for which proton therapy has been shown to be more effective than its alternatives. HAS report. Keeping in mind all the limitations of available literature, results suggest that: - for adenoid cystic carcinoma of the head and neck, for neoplasms of salivary glands not completely resectable and for skull-base chordomas and chondrosarcomas, carbon-ion hadrontherapy seems to perform better than conventional radiotherapy; - for skull-base chordomas and chondrosarcomas and for head and neck cancers, carbon-ion hadrontherapy seems to be equivalent to other high-technology radiotherapy techniques (for example stereotactic radiotherapy); - late toxicity is reported especially for skull-base chordomas and chondrosarcomas. The report goes on to state that, in general, there are no sufficient data to draw definitive conclusions on the relative merits of HT. It seems it is a promising technique for the treatment of inoperable tumours, surrounding by healthy tissue. 14

AHRQ report concluded There are many publications on particle (mainly proton) beam therapy for the treatment of cancer, but these studies don t document the circumstances in contemporary treatment strategies in which radiotherapy with charged particles is superior to other modalities. Therefore, comparative studies in general, and randomized trials in particular, are likely needed to document the theorized incremental advantages of particle beam therapy over other radiotherapies. KCE report concluded that for skull base chordomas, there is currently no clear clinical evidence from comparative studies to assess the clinical superiority in efficacy between proton or carbon ion and classical photon therapies (or their combination, or high precision photons therapy). * + In chondrosarcomas of the skull base, there are no differences between photon and proton irradiation in the results of case series for both local control rate and for overall survival whilst carbon ion therapy seems currently less effective than existing treatment. * + A better local control (but no significant difference in survival) with carbon ion and photons than with photons alone in patients with cystic adenoid carcinoma was reported. * + There is currently no evidence to support the use of proton therapy as first line treatment in CNS tumours in children. Lodge et al concluded In the absence of well-designed trials clearly demonstrating the superiority of protons and C-ions over the best photon techniques, no definitive conclusions on the relative merits of photons, protons and ions can be drawn for head and neck cancer. Amichetti et al concluded On the basis of the available literature, superiority in terms of local control of protons has been demonstrated only versus conventional photon radiotherapy and in terms of treatment planning and delivery with respect to other modalities mainly in larger lesions in proximity to normal critical structures and in irregularly shaped tumors, Data on stereotactic radiotherapy are scarce and data on IMRT are still lacking. There is a need to compare proton therapy with photon-based treatments delivered at the best of conformal technologies available. ANZHSN report concluded The evidence for proton beam therapy in neoplasms involving, or adjacent to, cranial structures remains inconclusive. Further studies are required to determine if proton therapy is indeed substantially better compared to conventional radiotherapy, as inferred by numerous model/treatment-planning studies. B. Ocular neoplasms Seven analyses on Particle Beam Therapy for the treatment of ocular diseases were found by secondary literature search. Four of these are HTA-reports (HTA-VATAP: Flynn K, 2010; HTA-HAS: Blanchard S, 2010; AHRQ: Trikalinos TA, 2009; KCE-Report: Huybrechts, 2007), two are systematic reviews (Bekkering GE, 2008; Lodge M, 2007) and one is a horizon scanning report (ANZHSN: Lee I, 2007). The systematic review of Bekkering and the horizon scanning report of Lee considered only the efficacy of HT in ocular diseases while the HTA reports studied the application of this technology in more types of tumours. Results Trikalinos TA, 2009 Particle Beam Radiation Therapies for Cancer (Tufts Medical Center Evidence-based Practice Center - AHRQ) For uveal melanoma particle beam therapy was used for a wide range of melanoma locations (i.e., choroid plexus, ciliary body, or iris) and sizes; the outcomes of interest are overall survival, cancer specific survival (rates of eye retention, vision retention, visual acuity and changes in tumours size), other surrogate outcomes of overall survival, adverse events, outcomes that are relevant to the quality of life. 15

Three RCTs were reported comparing lower vs. higher doses of particle beam therapy, particle beam therapy vs. other radiotherapy (brachytherapy) or vs. combination with additional therapy (laser thermotherapy for uveal melanoma). These studies are shown in Table 2. No trial reported significant differences in overall or cancer-specific survival or in total serious adverse events. Table 2. Comparators assessed in the randomized controlled trials (Uveal melanoma) Centre, reference Comparison N Survival (overall/ specific) MGH (USA), Gragoudas 2000 Higher vs. Lower dose proton RT 188 No/No UCSF (USA), Char 1993 Helium RT vs. I 125 Brachytherapy 320 Yes/Yes CPO (France), Desjardins 2006 Proton RT vs. Proton RT + laser TTT 151 Yes/Yes CPO=Centre de protonthérapie d'orsay; MGH=Massachusetts General Hospital; N=number of enrolled patients; RT=radiotherapy; TTT=transpupillary thermotherapy; UCSF=University of California San Francisco. Yes=overall/specific survival reported; No=overall/specific survival not reported. The same authors found seven nonrandomized studies comparing particle beam radiotherapy with brachytherapy, conventional photon radiation or surgery. These are shown in Table 3 Table 3. Nonrandomized controlled trials (uveal melanoma) Centre, reference Comparison N CPO (France), Desjardins_2003 Proton RT vs. I 125 brachytherapy Survival (overall/ specific) 1272 Yes/No UCSF (USA), Harbour_1997 Helium RT vs. I 125 brachytherapy 766 No/No MGH (USA), Seddon_1990 Proton RT vs. enucleation 556 Yes/Yes UCSF (USA), Char_1996 CCO (UK), Wilson_1999 Helium RT vs. I 125 brachytherapy Proton RT vs. I 125 brachytherapy vs. Ru-106 brachytherapy 426 No/No 267 Yes/No MGH (USA), Sedonn_1985 Proton RT vs. enucleation 120 Yes/Yes UCSF (USA), Char _2003 Proton RT vs. proton RT + laser TTT 56 No/No CCO=Clatterbridge Centre for Oncology; CPO=Centre de protonthérapie d'orsay; GI=gastrointestinal; GSI=Gesellschaft fuer; IMRT=intensity-modulated radiotherapy; LLU=Loma Linda University; MGH=Massachusetts General Hospital; N=number of included patients; NIRS=National Institute of Radiological Sciences; RT=radiotherapy; SFRT=stereotactic fractionated radiotherapy; TTT=transpupillary thermotherapy; UCSF=University of California San Francisco. Yes=overall/specific survival reported; No=overall/specific survival not reported. Lee I, 2007 Proton beam therapy for the treatment of uveal melanoma (Australian and New Zealand Horizon Scanning Network ANZHSN) 2 small RCTs (Level II evidence), 1 comparative observational study (Level III-B evidence) and 22 peer reviewed case series studies (Level IV evidence) reported safety and effectiveness outcomes on the use of proton beam therapy for the treatment of patients with uveal melanoma. One RCT (Gragoudas et al. 2000) showed that patients treated with a higher dose of proton beam therapy [70 Cobalt Gray Equivalent (CGE)] had similar rates of rubeosis/neovascular glaucoma compared to patients receiving a lower radiation dose (50 CGE), 7% (7/94 patients) and 10% (9/94 patients) respectively. These 16

results indicate that radiation dosage may play a lesser role compared to the size and location of the tumour in causing neovascular glaucoma or rubeosis. Comparisons between the randomised patient groups who received 70 CGE or 50 CGE revealed that the incidence of posterior subcapsular opacity was twice as high in patients who received the lower dose of 50 CGE, but this difference was not statistically significant. In patients with tumours less than 4 DD to the optic disc and/or the macula, despite a 30% reduction in radiation dose (70 CGE to 50 CGE), there were no statistically significant differences in retention of visual acuity between patients treated with 70 CGE or 50 CGE proton beam therapy at 12 months and 60 months post-treatment. There was no difference in local tumour control rates in patients treated with 70 CGE or 50 CGE proton beam therapy. Given the low event rate, the lack of statistical significance for these observations may have been attributed to a type 1 error and do not necessarily exclude real differences. In an effort to decrease the risk of exudates and glaucoma, the RCT by Desjardins et al. (2006) examined if the systematic addition of transpupillary thermotherapy (TTT) after proton beam therapy had a beneficial effect on these ocular complications. No significant difference in the reduction in the incidence of glaucoma was noted between the two treatment groups. Patients who received TTT had less severe glaucoma, with intraocular pressure that tended to return to normal after several months of treatment. In addition, patients treated with TTT had significantly lower secondary enucleation rates (p = 0.02). The majority of studies included were low quality (level IV evidence). The key measures of the effectiveness of proton beam therapy for the treatment of uveal melanoma were local tumour control, survival and retention of the treated eye which differed from the key complications assessed following proton beam therapy, these being the incidence of rubeosis, neovascular glaucoma, cataracts and vision loss. The utilisation of proton beam therapy for the treatment of uveal melanomas appears to be associated with substantial complications that are highly dependent on tumour characteristics (size, location) as well as pre-treatment patient characteristics (e.g. pre-treatment visual acuity, glaucoma etc.). The lack of comparative studies severely limits the conclusions that can be drawn from the available evidence. Local tumour control rates achieved with proton beam therapy was approximately 95%, indicating that this treatment is capable of preventing recurrence in most patients. The incidence of metastases appears to be comparable to other treatment modalities, with 5-year metastases-free survival ranging from 73% to 80%. Overall patient survival after proton beam therapy ranged from 78% to 85% at 5-years post-treatment, which appears to be similar to patients treated with brachytherapy. Bekkering GE, 2008 The Effectiveness and Safety of Proton Radiation Therapy for Indications of the Eye This systematic review includes 4 RCTs: 1 for uveal melanoma and 3 for Age-related macular degeneration (AMD). Methodological quality of the included studies was low. None of the controlled trials reported concealment of allocation and only two blinded their patients. Uveal Melanoma: 1 RCT and 12 case series were found. The RCT including 188 patients, was designed to determine whether a reduction in proton radiation dose would decrease radiation- induced complications for patients with uveal melanoma at high risk of these complications. All tumours were located within four disk diameters of the optic disk and/or the macula. The study showed no reduction in visual loss when reducing the radiation dose from 70 to 50 CGE. Choroidal Hemangioma: no RCTs were identified: there was one observational study with historical controls and two case series. The observational study showed that proton as well as photon radiation were effective 17

in resolving retinal detachment. Proton therapy, however, appeared to be associated with more side effects. Adverse effects reported (experimental vs. comparison intervention): grade 4 retinopathy 1 vs. 0; grade 3 adverse effects on lens 0 vs. 1; grade 3 lacrimation 1 vs. 1; any grade retinopathy 40% vs. 15.7%. The results of the 2 case series suggested that vision improved in the majority of patients. Subfoveal Choroidal Neovascular Membranes (CNVM) Associated with Age-Related Macular Degeneration (AMD): four clinical trials and one case series were found. 2 RCTs compared proton versus sham radiation or observation. The first trial reported no differences between the two groups with regard to visual acuity. Adverse effects observed: retinal detachment unrelated to treatment, non-impairing optic neuropathy, no case of radiation retinopathy. The second reported a reduction of vision loss in the proton group compared to controls at 1 year. However, this difference was not significant at 2 years; none adverse effects reported. Two trials compared two doses of proton radiation. The first RCT evaluated 166 patients who were randomized to lower-dose (16 CGE) or to higher-dose (24 CGE) proton radiotherapy. No differences between the two groups were demonstrated; radiation complications were observed. The nonrandomized study suggests more favourable results on visual acuity for the high-dose compared to the low-dose groups, however, the rate of adverse events (retinopathy, lash loss, uveitis, conjunctival hyperemia, cataracts or optic nerve margin swelling) was also higher in the high-dose group. One small prospective case series was found, that reported stable or enhanced visual acuity for 86% at 3 months and 61% at 18 months. No secondary effects related to the treatment were observed. No radiation cataracts or retinopathies; 4 patients with keratitis, which resolved. Flynn K, 2010 Brief overview: proton therapy for cancer (VATAP - Veterans Health Administration Office of Patient Care Services Technology Assessment Program) The HTA report included 4 studies on ocular diseases: Brada 2009, Trikalinos 2009, Bekkering 2008, ANZHSN (2007b). Brada (2009): This was a quantitative summary of low quality highly variable (diagnoses, interventions, follow up) observational primary studies and was not used in this report. Trikalinos (2009): This report was based on single-arm studies, case series or cohort studies, (220, of which 80 on ocular cancers), RCTs (10, 4 on ocular cancers), CCTs (13, 7 on ocular cancers). RCTs and CCTs compared treatment with and without charged particles; no reports of statistically significant or important differences in survival or adverse events were found. Bekkering (2008): Methodological quality was low and there were large variations in intervention techniques and patient characteristics within and between studies. Results for uveal and choroidal melanoma suggest survival benefit but with significant side effects including choroidal hemangioma and AMD. No benefit for HT was reported. ANZHSN (2007b): Proton beam, like other radiation therapy for uveal melanoma, is associated with considerable complication rates: rubeosis, neovascular glaucoma, cataract, vision loss; approximately comparable tumours control rates and patient survival ( 95%) to brachytherapy. One comparative study showed that enucleation rates for proton were comparable to brachytherapy. 18

Blanchard S, 2010 Hadrontherapie par ions carbone - rapport préliminaire (Haute Autorité de Santé - HAS) Results are based on the systematic review of HT reported by Lodge (2007). For results, please refer to this review described later. Huybrechts M. et al., 2007 Hadrontherapie (Federaal Kenniscentrum voor de Gezondheidszorg - KCE) Proton Therapy: 2 cohort studies followed patients (440 eyes) with uveal melanomas treated with enucleation or with proton beam irradiation. Cox regression analysis adjusted for prognostic variables found no difference in disease free survival (RR 1.0 [95% CI 0.7-1.4]) or overall survival (RR 1.2 [95% CI 0.9-1.2]). One retrospective comparative case series studied 597 patients with choroidal melanomas treated either with proton beam irradiation or with episcleral ( 106 Ru and 125 I) brachytherapy. Local recurrence rates were better with proton therapy but with a higher mortality rate. The model was not appropriately adjusted for possible confounders. The RCT of Gragoudas assessed the clinical effects of two doses of either 50 or 70 CGE in 188 patients with small or medium size melanomas (< 15 mm in diameter and < 5 mm in height) near the optic disc or macula. Local and distant tumour recurrence and cancer specific death rates were similar. Visual acuity loss was similar. The lower-dose group did experience significantly less visual field loss. The trial was underpowered to conclude whether there were differences in cancer control rates. A systematic search found an additional study not included in the systematic reviews: Char 2002 compared late recurrence between 3 different treatments reported as retrospective cases series for uveal melanomas ( 125 I brachytherapy; proton therapy; helium therapy) after long term follow-up and concluded that more late recurrences (5 to 15 years) occur with extended follow-up after brachytherapy but not after proton or helium therapy. Many potential biases are possible in such a retrospective design study and the quality of evidence is low to very low. Neovascular age-related macular degeneration: A systematic search found 3 clinical trials on this topic (2 RCTs and 1 prospective case series). The 3 studies are of low quality. Biases are not accounted for. The validity of the results is uncertain. The RCT of Ciulla compared proton beam therapy with sham therapy in neovascular age-related macular degeneration. There is no statistical difference between the 2 groups. The two other studies that compared the effect of 2 different doses of proton therapy concluded that there is no significant difference in rates of visual loss between the 2 doses. There is therefore currently no evidence for the use of proton therapy in neovascular age-related macular degeneration. Carbon and Helium Ion Therapy: This topic is considered in one systematic review. Our search did not find additional clinical studies not included in the systematic reviews (SRs) or HTA reports. In the SR of Lodge, two studies dealt with Helium ion therapy. The two other studies (Tsuji, 2007 and Hirasawa, 2007) are prospective phase I/II clinical trials in which eye retention rates and severe side effects such as neovascular glaucoma were analyzed. The rates with carbon ion (84% eye retention and neovascular glaucoma > 40%) seems less impressive than proton or photon therapy rates. Lodge M, 2007 A systematic literature review of the clinical and cost effectiveness of hadron therapy in cancer 19

For proton therapy, two prospective phase I/II dose escalation studies and eight retrospective studies were identified. Weighted means for local tumour control and 5 year overall/cause specific survival were 97% and 85% / 85%, respectively. The weighted mean for eye retention was 90%, and neovascular glaucoma occurred in 12% of patients. For ion therapy, six studies were identified which were performed in two institutions. In Berkeley, two retrospective case studies (He-ions, n = 1343) and one phase III study were reported (helium ions vs. brachytherapy, n = 184; ion arm n = 86). In the only phase III trial that met the inclusion criteria, 184 patients with uveal melanoma were randomised between helium ions and 125 I brachytherapy. There were significantly more tumour recurrences in the brachytherapy than in the He arm (13.3% vs. 0%, p < 0.001). Reported complication rates of dry eye, epiphora and neovascular glaucoma were more frequent in the helium-ion treated than in the brachytherapy group (27.9% vs. 8.2%, 20.9% vs. 3.1%, 29.1% vs. 11.2%; p values not given in the article, but according to the 95% confidence intervals all differences were statistically significantly different). There was no significant difference in the rate of enucleation between both groups (He: 9.3%, brachytherapy ( 125 I): 17.3%) and similarly the incidence of distant recurrences and mortality rates were not different between both groups (He: 9.3% vs. 125 I: 8.2% and He: 16.2% and vs. 125 I 19.4%, respectively). In both groups, approximately 66% of the patients retained at least 20/50 visual acuity 3 years after treatment. Two prospective phase I/II trials for carbon ion therapy were performed at Chiba. Eye retention rates and severe side effects such as neovascular glaucoma were specifically investigated. Both outcomes, i.e. eye retention rate 84% and neovascular glaucoma >40%, appeared to be inferior to proton or photon therapy. The results of stereotactic radiotherapy (STR) with photons were similar to those of proton therapy. However, most series with SRT photon therapy only report 2 3 year results compared to 5 year results reported in the proton therapy studies. Moreover, the incidence of neovascular glaucoma seems to be lower with proton therapy than with SRT photons (12% vs. 16%). Conclusions on ocular neoplasms Trikalinos TA, 2009 concluded Overall, no study found that charged particle radiotherapy is significantly better than alternative treatments with respect to patient-relevant clinical outcomes. However, this Technical Brief did not intend to assess outcomes or evaluate the validity of claims on the safety and effectiveness of particle beam radiotherapy. Lee I, 2007 concluded The evidence on the safety and effectiveness of proton beam therapy for the treatment of uveal melanomas is mixed. The local tumour control rates achieved are remarkably consistent across studies despite the heterogeneity across study cohorts and methodology, and at the very least appears to be comparable to brachytherapy. Meanwhile, the incidence of metastasis and overall patient survival is comparable to brachytherapy as well. However, it is unclear if proton beam therapy results in a substantial improvement of eye preservation rates and the ocular complications observed post-treatment are of concern. Further studies are required to address the limitations of previous studies and to compare proton beam therapy to existing techniques. It is important to note that due to the heterogeneity between the included studies, it may be impossible to compare results in a meaningful manner. A large proportion of existing studies on proton beam therapy for uveal melanoma are quasi-anecdotal due to the fact that treatment parameters and patient selection are so variable. Bekkering GE, 2008 concluded There is limited evidence on the effectiveness and safety of proton radiation due to the lack of well-designed and well-reported studies. There is a need to lift evidence on 20