Consensus Report From the Stockholm Pediatric Proton Therapy Conference

Abstract

According to the American Society for Radiation Oncology's Model Policy published in 2014 (1American Society for Radiation Oncology. Model policies: Proton beam therapy. Available at: www.astro.org/uploadedFiles/Main_Site/Practice_Management/Reimbursement/ASTRO%20PBT%20Model%20Policy%20FINAL.pdf. Accessed April 26, 2016.Google Scholar), solid tumors in children are considered among the highest priority for proton therapy. Worldwide, there are currently 54 facilities offering proton therapy and 61 more under construction (2Particle Therapy Cooperative Group. Home page. Available at: www.ptcog.ch/. Accessed March 11, 2016.Google Scholar). As the number of institutions proliferates, expert opinion is important in guiding safe and rational adoption and use of this technology in young patients. In June 2015, 24 international leaders in pediatric radiation oncology, pediatric oncology, medical physics, and radiobiology convened in Stockholm to exchange ideas and perspectives on treating children with proton therapy. To encompass the spectrum of current opinion, participants represented a wide cross-section of public and private health care settings, including institutions with access and those without access to proton therapy. The meeting had 3 predefined aims: (1) To better define the role of proton therapy in the management of pediatric cancers; (2) to reach a consensus on the clinical scenarios for which proton therapy provides the most benefit in pediatric cancers; and (3) to prioritize the innovation and development necessary to advance the subspecialty. These aims were accomplished via interactive presentations, anonymous surveys, facilitated dialogue, group breakout sessions, and dosimetry plan comparisons. Early in the meeting, full consensus was established on fundamental aspects of pediatric radiation oncology. Using existing technology, radiation therapy is very effective in curing many pediatric tumors. However, participants uniformly agreed that delivery of radiation in pediatric patients requires extra caution owing to the radiosensitive nature of developing tissue, which has a low dose threshold for late effects, and the long natural life expectancy of survivors. The intent of pediatric radiation oncologists is consistent: deliver as low a dose of radiation to as small a volume necessary to cure the child. Although this paradigm is not necessarily unique to children, pediatric radiation oncologists place a heightened emphasis on late effects. With modern imaging and radiation therapy techniques, there are many diverse paths that converge at this goal of reducing radiation exposure to healthy tissue; participants concurred that proton therapy represented but one avenue of advancement toward this end. Common agreement was identified on broad topics related to proton therapy indications. All participants considered “access to a proton therapy system a complement to my pediatric program.” The majority of participants (58%) thought that most (50%-90%) of the pediatric tumors requiring radiation should be treated with proton therapy. However, no participant thought that proton therapy should be systematically substituted for photon radiation in every pediatric tumor, and 10% of participants believed that less than 25% of pediatric tumors should be treated with proton therapy. Participants approached this question from the perspective of their local, current health care environment in terms of accessibility and cost of proton therapy. Although this difference accounts for some of the diversity in viewpoint, the main source of heterogeneity was attributed to mixed opinions related to tumor type and site, which is a surrogate for radiation dose and sensitive normal tissues adjacent to the malignant target. Many participants felt that disease-specific indications are often vague and an oversimplification. Although general recommendations for specific disease sites can be made by consensus, physicians and payers must veer from rigid guidelines to address individual patient and disease characteristics (such as age, comorbidities, anatomy, tumor location, and tumor distance from critical structures). When the group was presented with various clinical scenarios and asked, “What is the treatment of choice for the following pediatric tumors?,” clear trends emerged (Fig. 1). Participants could answer “proton therapy,” “photon therapy,” or “both, depending on circumstances.” The group felt that proton therapy was a preferred indication for the majority of pediatric central nervous system tumors, as well as skull base tumors and retinoblastoma. The exception was high-grade glioma of the brain and brainstem, for which participants favored photon therapy owing to the low cure rate and subsequent improbability a child survives long-term to derive the relative reduction in late complications. As a corollary, these tumors are highly infiltrative and often recur in areas of the brain just beyond the initially defined volume. The precision of proton therapy in this setting may undermine its applicability. The group also indicated that photon therapy was generally appropriate for Wilms tumor, where the dosimetric precision of proton therapy is unnecessary given the low dose and broad targets encompassing the entire abdomen or lungs. There was a diversity of opinion on Hodgkin lymphoma and neuroblastoma, common midline tumors that are curable with a lower radiation dose. There were also mixed opinions on rhabdomyosarcoma and Ewing sarcoma, for which the relative value of proton therapy depends on the anatomic tumor site. To further advance the discussion regarding disease-specific indications, comparison plans were reviewed for ependymoma, craniopharyngioma, medulloblastoma, and rhabdomyosarcoma. Each respective plan was generated independently by a pediatric radiation oncologist from a predominantly photon institution and a pediatric radiation oncologist from a predominantly proton institution. The specific photon and proton technique (such as intensity modulated radiation therapy, volumetric arc therapy, tomotherapy, double-scattering proton delivery, or pencil-beam proton delivery) was not constrained. Instead, experienced pediatric radiation oncologists and their staff selected their best available option to simulate “real-world” circumstances, which in turn introduced variability and bias owing to different planning goals and priorities. Across all plan comparisons, the integral dose was lower for proton therapy, although the magnitude of relevance was debated. For posterior fossa ependymoma, the most clinically meaningful difference was a lower dose to the cochlea, temporal lobes, and hypothalamic–pituitary axis with the proton plan (Fig. 2). For the craniopharyngioma, the most clinically relevant difference observed was the reduced low–moderate dose to the supratentorial brain (Fig. 3). The impact of this difference on cognitive development is theorized and age-dependent. For the medulloblastoma case, proton therapy reduced the dose to the lung, heart, face, abdomen, and pelvis but delivered a higher dose to the scalp and paraspinal soft tissue.Fig. 3Craniopharyngioma photon (IMRT, top) and proton (bottom) comparison plans in 3 planes.View Large Image Figure ViewerDownload Hi-res image Download (PPT) There were mixed differences with regard to brain exposure, with the photon plan delivering a higher dose to the frontal and temporal lobes, whereas the proton plan delivered a higher dose to the medial parietal and occipital lobes (Fig. 4). For the parameningeal rhabdomyosarcoma case, proton therapy delivered a lower dose to the cochlea, oral cavity, face, and infratentorial brain but a higher dose to the frontal lobe and larynx (Fig. 5). Although valuable for discussion, this dosimetric exercise was not intended to be exhaustive nor reach a definitive conclusion, and the details of each plan comparison are beyond the scope of this report. Instead, readers are referred to more rigorous and standardized dosimetric comparisons of proton and photon therapy for ependymoma, craniopharyngioma, medulloblastoma, and rhabdomyosarcoma published in high-impact journals over the last decade, including articles by Macdonald et al (3MacDonald S.M. Safai S. Trofimov A. et al.Proton radiotherapy for childhood ependymoma: Initial clinical outcomes and dose comparisons.Int J Radiat Oncol Biol Phys. 2008; 71: 979-986Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar), Boehling et al (4Boehling N.S. Grosshans D.R. Bluett J.B. et al.Dosimetric comparison of three-dimensional conformal proton radiotherapy, intensity-modulated proton therapy, and intensity-modulated radiotherapy for treatment of pediatric craniopharyngiomas.Int J Radiat Oncol Biol Phys. 2012; 82: 643-652Abstract Full Text Full Text PDF PubMed Scopus (134) Google Scholar), Lee et al (5Lee C.T. Bilton S.D. Famiglietti R.M. et al.Treatment planning with protons for pediatric retinoblastoma, medulloblastoma, and pelvic sarcoma: How do protons compare with other conformal techniques?.Int J Radiat Oncol Biol Phys. 2005; 63: 362-372Abstract Full Text Full Text PDF PubMed Scopus (206) Google Scholar), and Ladra et al (6Ladra M.M. Edgington S.K. Mahajan A. et al.A dosimetric comparison of proton and intensity modulated radiation therapy in pediatric rhabdomyosarcoma patients enrolled on a prospective phase II proton study.Radiother Oncol. 2014; 113: 77-83Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar).Fig. 5Parameningeal rhabdomyosarcoma proton and photon comparison plans. Left: Proton pencil beam (PBS) plan. Middle: Photon tomotherapy. Right: A dose subtraction image to highlight differences between the 2 modalities.View Large Image Figure ViewerDownload Hi-res image Download (PPT) The conference participants spent considerable time exploring the magnitude of benefit and pediatric patient prioritization. Compared with other current pediatric indications, the group felt that the relative amount of benefit for total body irradiation, whole-abdomen irradiation, whole-lung irradiation, and whole-brain irradiation was questionable. In fact, the value of proton delivery may be so low under these conditions that some participants asked whether it was outweighed by the uncertainties surrounding aspects of particle therapy dose deposition. For example, are the benefits of “cardiac- and breast-sparing” whole-lung radiation outweighed by the technical uncertainty of proton delivery? Are the benefits of “lens-sparing” whole-brain radiation outweighed by the ill-defined, end-of-range radiobiologic effect on the optic nerves and retina when delivered via a posterior beam? In cases with a consensus view of probable benefit, group members posed questions of prioritization. In a child with metastatic incurable cancer, do we use protons to simply reduce the acute effects of treatment? Should protons be applied to reduce a hypothetical risk of unilateral, high-frequency hearing loss from 7% to 5%? In this context, it is important to recognize that in many cases, pediatric radiation oncologists often do not possess models to accurately quantify side-effect risks and therefore must be cautious not to overstate precision when claiming the comparative estimates of benefit. Even when data are available to accurately characterize risk, decisions of patient prioritization are linked to the supply of available proton treatment slots, which in turn depend on health care resources and cultural expectations. Finally, given the limited distribution of proton centers worldwide, we should be cognizant of the financial and emotional burden placed on families that must relocate to a distant treatment facility. When families feel compelled to provide the “very best” for their children, the dialogue needs to encompass all aspects of medical care, including psychosocial elements. The group reached a consensus on long-term goals to aid in patient prioritization. With respect to side-effect reduction, pediatric dose–effect models need to be developed and refined across all radiation modalities. In the ideal setting, accurate modeling will permit detailed, patient-specific “virtual trials,” which can then be placed into the context of local health care environments. With respect to improving tumor control rates, the use of proton therapy to increase tumor dose (and possibly cure rates) is promising but should be explored within formal clinical trials. Until these long-term goals are achieved, the group settled on the broad position that the best proton candidate is a young child with a curable tumor requiring the delivery of moderate to high focal radiation dose in proximity to critical tissue. Though lacking objective quantification, this approach of patient prioritization will optimize both the magnitude of risk reduction and the duration of benefit. Beyond these concerns, conference participants discussed existing barriers to proton therapy. Participants felt strongly that more research was needed on the radiobiology of protons in children, specifically investigations involving the end of range radiobiologic and physical dose uncertainty. One-third of the group felt that radiobiologic uncertainty represented a “significant obstacle for broad adoption of proton therapy.” In addition, the group largely agreed that ongoing lack of pediatric treatment capacity worldwide, particularly at proton centers committed to research and those with pediatric expertise, represented an important barrier. Finally, the participants vocalized concerns about persistent referral barriers even when there was an unambiguous medical indication for proton therapy. This apprehension branched into diverse themes. For example, one participant shared the perspective, “These doctors bought a proton machine and suddenly think they are pediatric radiation oncologists.” In reality, the complexity of delivering quality pediatric care extends far beyond radiation modality. It commonly involves unique aspects of anesthesia/immobilization, setup imaging, multiagent concurrent chemotherapy, family dynamics, decision-making capacity, and devoted research resources for participation in cooperative protocols. Another type of referral barrier was characterized by the statement, “It is too expensive for my state/country to refer every child for proton therapy. We have higher health care priorities and less resources.” In fact, this is a reality for most of the world's population. What happens when one's role as an individual patient fiduciary conflicts with one's role as societal fiduciary? There is no easy answer to such a question; it requires deliberate ethical introspection and pragmatism. A third theme identified as a referral barrier is reflected in sentiments such as, “If I refer this child for proton therapy, I will not meet my salary revenue goals,” or “I have been treating children with radiation for decades and I am very skilled at what I do. There is no need for me to refer a patient elsewhere.” Issues of personal reimbursement and professional ego are uncomfortable topics for each of us to confront. Perhaps equally revealing were issues that conference participants did not view as significant barriers to broad proton therapy implementation in children. One example is the issue of neutron contamination generated by proton scatter (7Hall E.J. Intensity-modulated radiation therapy, protons, and the risk of second cancers.Int J Radiat Oncol Biol Phys. 2006; 65: 1-7Abstract Full Text Full Text PDF PubMed Scopus (896) Google Scholar, 8Gerweck L.E. Huang P. Lu H.M. et al.Lifetime increased cancer risk in mice following exposure to clinical proton beam-generated neutrons.Int J Radiat Oncol Biol Phys. 2014; 89: 161-166Abstract Full Text Full Text PDF PubMed Scopus (3) Google Scholar). Generally speaking, the attendees did not feel this out-of-field exposure represented an undue risk of radiation carcinogenesis when considered in light of the reduced dose along the beam path afforded by proton therapy. Nevertheless, participants agreed that newer techniques of pencil beam delivery, which further reduce neutron scatter, will provide extra reassurance. In addition, participants did not consider randomized studies or data from a “model-based approach” to patient selection as mandatory before broad adoption of pediatric proton therapy. Although each holds value, the subsequent delay in proton therapy implementation is not justified given the current level of evidence available for most common pediatric tumors. Furthermore, the feasibility of such trials is highly questionable given the current accessibility of proton therapy and biases in patient selection related to patient and physician preference. The conference concluded with a consensus on the most desirable areas for advancement within pediatric proton therapy delivery. The overwhelming highest priority was reduction in cost of proton therapy systems, specifically the enormous capital investment required to purchase hardware. Although protons may result in long-term cost savings when utilized in children (9Mailhot Vega R. Kim J. Hollander A. et al.Cost effectiveness of proton versus photon radiation therapy with respect to the risk of growth hormone deficiency in children.Cancer. 2015; 121: 1694-1702Crossref PubMed Scopus (40) Google Scholar), short-term economic viability cannot be ignored. Given their complexity, treating pediatric cancers with proton therapy is—at best—a cost-neutral activity over the short term, and efforts to recover the high cost of proton therapy equipment result in pressure to treat more profitable diseases in adults (10Johnstone P.A. Kerstiens J. Doing poorly by doing good: The bottom line of proton therapy for children.J Am Coll Radiol. 2014; 11: 995-997Abstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar). Another high-priority concern was the need for a formalized national or international registry of children treated with proton therapy. Such an undertaking would be expensive and complicated but may represent the quickest way to gather data on disease control and empirically refine our models of pediatric normal-tissue dose effect. Other important needs included the development of safe, volumetric image guidance (such as low-dose cone-beam computed tomography with a collimated, variable field of view). Most proton facilities currently rely on less advanced orthogonal kilovoltage image guidance. Additionally, participants desire treatment rooms that comfortably accommodate anesthesia units and are equipped with age-appropriate entertainment/distraction devices. The conference was successful in that a diverse group of international leaders identified common themes to shape global discussion around the use of proton therapy in the treatment of children with cancer. The broad consensus was that the technology of proton therapy represents an incremental and nonexclusive advancement in the field of pediatric oncology, consistent with long-standing goals of the specialty. Where available, it has an accepted role in young children with curable tumors who require moderate to high radiation doses. Yet important questions remain surrounding the magnitude of benefit, patient prioritization, and radiobiology. With the hope of moving the technology forward in a way that best serves our pediatric patients, the group identified specific technological and research initiatives that require attention. As more evidence is developed and more pediatric patients are treated worldwide, a similar meeting should reconvene to provide an updated perspective.