Proton therapy for the treatment of children with CNS malignancies

SUMMARY: 

Proton therapy is a novel technique for treating pediatric malignancies. As a tool to reduce normal-tissue dose, it has the potential to decrease late toxicity. Although proton therapy has been used for over five decades, most pediatric dosimetry studies and clinical series have been published over the last 10 years. The purpose of this article is to review the physical, radiobiological and economic rationales for proton therapy in pediatric CNS malignancies, and provide an overview of the current challenges and future direction of research and utilization of this approach.

PRACTICE POINTS.

• Proton therapy allows for more sparing of normal tissue (decreased integral dose) in comparison with photon therapy, decreasing the risk of late toxicities and secondary malignancies.

• The decrease in integral dose is important in pediatric neuro-oncology cases where the overall survival is 80–90% and the brain is still developing. In recent economic and risk models, the reduction in late toxicities increases quality-adjusted life years and cost–effectiveness.

• Compared with historical photon cohorts, current clinical studies of proton therapy for pediatric CNS malignancies show equivalence or improvements in local control, progression-free survival and overall survival. The Children’s Oncology Group protocols are now incorporating proton therapy.

• Delivery of proton therapy requires the building of a proton center with cyclotrons or synchrotrons, which necessitates a large initial investment and specialized expertise.

• Future directions of research and development include improvement of proton delivery techniques, quality assurance, appropriate patient selection, radiobiologic studies and cost–effectiveness analyses.

Background

Pediatric CNS and cranial tumors comprise the majority of pediatric solid tumors. Children with these malignancies are treated with either surgery, chemotherapy, radiation therapy or a combination of these modalities. Given that the 5-year overall survival (OS) for these children is above 70%, the goal in the treatment of pediatric CNS tumors has been to reduce toxicity, while maintaining the efficacy of treatment. Proton therapy is a type of external-beam radiation therapy, which was initially pioneered in 1946, when Robert Wilson published an article regarding the unique physical properties of protons and their potential application in oncology [1]. The first proton beam centers were built in the 1950s and patients first received proton radiation therapy in 1958 [2]. These centers demonstrated the potential for normal tissue sparing, with more accurate tumor targeting. In the past 10 years, proton therapy has gained favor for pediatric malignancies, where there is an expectation of long-term survival for the growing number of patients who are cured of their disease. According to a recent patterns-of-care study [3], the total number of children treated at US proton centers increased by 33% between 2010 and 2012 (from 465 patients in 2010 to 694 patients in 2012). In 2012, the three most commonly treated pediatric tumors were brain tumors (ependymoma: 106 patients; medulloblastoma: 89 patients; and low-grade glioma [LGG]: 78 patients).

Discussion

• Physics

Proton therapy capitalizes on the physical attributes of protons. In comparison with photons, protons have a mass of 1 atomic mass unit and are positively charged. Because of these two characteristics, the dose distribution of a proton differs from that of a photon. A high-energy proton travels in a straight line through tissue with minimal loss of velocity. The linear energy transfer (LET) is defined as the rate of loss of energy [4] per unit of length. For protons, the LET along its track is low until it reaches its maximal depth, where it deposits the majority of its energy. The high-energy region is described as the Bragg Peak and the proton delivers no further energy beyond this point, allowing for a narrow range of dose deposition.

High-energy protons are accelerated using cyclotrons or synchrotrons. In a cyclotron, a static magnetic field is used to accelerate the protons in an increasing spiral pattern. They are extracted at a set energy level and exit in a beamline. They are kept under a vacuum to prevent proton–air interactions and steered using magnets. A synchrotron is a tube-ring vacuum with magnets. Linear accelerators insert a beam into the synchrotron, which travels through and is contained within the tube. The strength of the magnetic field increases proportionally with the energy level [5]. Passive-scattering systems insert flattening filters to scatter the protons over a set area. Brass apertures are customized for each patient to target the area of the tumor while blocking surrounding areas and compensators are designed to accommodate normal tissue and tumor heterogeneity. Most proton centers use passive beam scattering to deliver treatment; thus, most published studies involve passive beam scattering.

Active beam scanning is a developing modality wherein magnets, rather than flattening filters, are programmed to spread the dose. While passive-scattering systems are used to deliver radiation using homogenous fields, intensity-modulated proton therapy (IMPT) treatment applies active beam scanning to deliver radiation using multiple superimposed inhomogeneous fields.

To target tumors effectively, the proton energy range is modulated, creating a spectrum of proton energy and a plateau described as the spread-out Bragg Peak (SOBP). The entrance dose is 20–25% of the prescribed dose. The tissue preceding the Bragg peak receives approximately the entrance dose, while the tissue within the SOBP receives the dose prescribed, and the tissue beyond the SOBP receives almost no dose.

A proton deposits energy through the ionization and excitation of atoms by interacting with orbital electrons. Approximately 20% of a proton’s initial energy is lost through inelastic nuclear interactions. By contrast, a high-energy photon deposits energy entirely through the generation of high-energy electrons in tissue. Its energy decreases exponentially with depth and it delivers some dose beyond the intended planning target volume. While advances in photon-based radiation therapy, such as intensity-modulated radiation therapy (IMRT), have allowed for more target conformity, the surrounding tissue still receives up to 20% of the prescription dose.

The mass of the protons also allows for less scattering in air and within tissue, creating a sharper dose decrease at the lateral edges of the beam over a smaller distance (called the penumbra). With a sharper penumbra than photons, proton therapy can target tumors, while sparing the surrounding organs at risk.

• Radiation biology

There are similarities between the biological effect of protons and photons. High-energy protons and Compton electrons produced by photons have similar track patterns and LET at the nanometer and micrometer scales. Researchers believe that both conventional radiation and proton therapy kill cells through DNA breaks, activating the apoptotic pathway, generating free radicals and inducing cell necrosis [5]. The relative biological effectiveness (RBE) is the ratio of the dose of photons to that of protons necessary to produce the same biologic effect. For proton therapy, the RBE of protons extrapolated from the end points of cell killing generally range from 1.1 to 1.2. Whereas the photon-based radiation dose is expressed as Gy, the proton dose is expressed as Gy (RBE). Recent studies of proton versus photon radiation discussed below have shown newly discovered differences in the radiobiology of the two modalities.

Subcellular studies, wherein tissue cell lines are irradiated using protons and photons of similar LET, have shown that protons more potently produce dsDNA breaks (DSBs) than photons, and the size of the repair foci of DSBs are larger with protons than photons. The number of ssDNA breaks and DSBs, and amount of complex DNA damage increased with protons along all segments of the SOBP. When Goetz et al. compared the epigenetics of proton and photon irradiation, they observed hypermethylation with proton therapy and hypomethylation with photon therapy [4,6]. Animal studies comparing whole-body photon and proton irradiation have revealed dose-dependent differences in gene expression profiles, with upregulation of both proangiogenic [7] and inflammatory [8] genes, and downregulation with proton therapy. Proton therapy was noted to significantly modulate 22 of 84 core genes involved in stem cell differentiation of extracellular matrix regulators [9]. The clinical implications of these subcellular differences are unknown.

Proton and photon therapy may also differ in cellular mechanism and cell cycle modulation. At a cellular level, while the entrance dose with proton therapy mimicked that of photon therapy, the dose increased at the Bragg Peak and distal shoulder region, reflecting a variance in RBE. In addition, clonogenic cell killing may depend on the pathology of the irradiated tissue. Proton therapy has been studied in different types of cell. Animal studies have demonstrated that proton therapy results in longer G2 phase accumulation and cell cycle arrest in human glioblastoma and melanoma cells [10,11], but not in thyroid cells [12]. The radiobiology of proton therapy in terms of angiogenesis, microenvironment modulation and metastasis continues to be elucidated.

Studies have estimated the risk of secondary malignancies and late toxicities based on dosimetric plans with conventional photon radiation, photon-based IMRT and IMPT. Monte Carlo simulations were utilized to estimate stray proton therapy and IMRT dose. Dose-risk models were based on the biological effects of ionizing radiation [13]. While the assumptions in each analysis varied slightly, the total lifetime attributable risk (LAR) of all secondary malignancies after craniospinal irradiation (CSI) in pediatric patients was estimated to be lower in proton therapy than IMRT [14–16]. Doses from secondary neutrons generated in patients were added to proton therapy plans. While the Athar analysis showed slightly higher LAR with 3D proton therapy than IMRT for secondary solid tumors of the lung and bronchi, as well as gynecologic and genitourinary malignancies, the other studies uniformly demonstrated a lower LAR with all proton modalities. The secondary neutron dose from IMPT is lower than with 3D proton therapy and, as such, the LAR for all malignancies was lowest with IMPT. Monte Carlo simulation has shown that the risk of developing a second malignancy from secondary neutron generation is associated with the field characteristics, the organ of second malignancy and younger patient age [17]. The LAR of lung and thyroid tumors, and leukemia, was greatest in younger males. For females younger than 14 years of age, the LAR of secondary hematologic and breast cancer was greatest. The risks from the secondary neutron dose from 3D proton therapy was comparable with that from IMRT [17].

A retrospective matched-pair analysis of 558 patients treated with proton therapy and 558 from the Surveillance, Epidemiology and End Results database was recently performed [18]. Of these, 44 matched patients were treated for pediatric malignancies; 31 unmatched pediatric patients were excluded from the analysis due to the rarity of their malignancy. At a median follow-up of 6.7 years, when adjusted for confounding variables, the hazard ratio of developing a second malignancy for the matched patients was 0.52 with proton therapy, and no pediatric patients treated with proton therapy developed a secondary malignancy in the treatment field [18].

The risk of late toxicities, including pneumonitis, heart failure, blindness, ototoxicity, hypothyroidism, xerostomia and premature ovarian failure after CSI were estimated to be lower with IMPT [14,19] when compared with conventional radiation and IMRT [14]. IMPT was found to be associated with a statistically significant lower potential years of life lost due to fatal complications [20].

Studies are ongoing to determine whether the increase in volume of tissue receiving low-dose radiation and the consequent theoretical risk translates into a difference in clinical outcome. These studies include many recent clinical series published on pediatric treatment outcomes.

• CSI for medulloblastoma/primitive neuroectodermal tumors

Medulloblastoma is the second-most common intracranial tumor in children, comprising 20% of pediatric CNS malignancies [21]. Standard-of-care treatment involves a multimodality approach with surgery, chemotherapy and radiation of the craniospinal axis (CSI). OS rates are greater than 70%. Given the large volume of radiation received to the surrounding normal tissues, patients who receive CSI are at risk of developing cardiovascular, pulmonary, endocrine, neurocognitive, auditory, ophthalmic, hematologic and reproductive late toxicities, growth impairment, and secondary malignancies.

Proton therapy has been most studied in medulloblastoma as a method of reducing normal tissue doses resulting from SOBP dose distribution (Figure 1). CSI dosimetric plans generated using photon radiation with a single posterior–anterior beam and proton therapy were compared in a 2-year-old patient. The prescription dose in both plans was 30 Gy and both plans delivered a homogenous dose to the target volume. Nevertheless, the predicted dose received by the heart and liver was 0 Gy (RBE) with proton therapy compared with a mean dose of 10.5 Gy to the heart and a mean dose of 6 Gy to the liver. The volume of the liver that received at least 50% of the dose was 33% and the volume of the heart that received at least 60% of the dose was 44%. Proton therapy delivered a decreased mean dose to the vertebral bodies, but there was hetereogeneity in the amount of dose received by the posterior and anterior elements as compared with the photon plan [22]. More recent dosimetric studies have compared modern photon techniques, including IMRT, with proton therapy. In a comparison of conventional radiation, IMRT and proton therapy, the dose to the cochlea was reduced from 101.2% of the prescribed dose with conventional radiation to 33.4% with IMRT and 2.4% with protons. The dose to 50% of the heart was 72.2% for conventional radiation, 29.5% for IMRT and 0.5% for proton therapy [23].

Figure 1. . Dose distribution for a pediatric patient.

(A) Proton and (B) photon dose distribution for craniospinal irradiation for a child with a medulloblastoma.

Further studies explored the possibility of sparing developing structures involved in normal neurocognitive growth and development. IQ scores in pediatric patients receiving radiation therapy are decreased when compared with age-matched normal and baseline values prior to radiation therapy. Armstrong et al. found that the dose received to the temporal lobe is correlated with a decrease in IQ [24]. The subventricular zone and hippocampus are known sites of neurogenesis involved in neurocognition. In this study, conventional photon therapy, IMRT, 3D conformal proton therapy and IMPT plans were generated, and the potential to decrease the dose to the subventricular zone was recorded. The dose to the clinical tumor volume was kept at 95% of the prescribed dose. The risk of impairment in task efficiency, organization and memory was calculated. Compared with the photon plans, both 3D proton therapy and IMPT were found to have a decreased risk of impairment for the same volume covered by 95% of the prescribed dose. Proton plans were also more sensitive to changes in clinical target volume coverage constraints [25].

The first clinical series on proton-based CSI for pediatric medulloblastoma was published in 2004 [26]. Yuh et al. reported acute toxicities for three children younger than 5 years of age with gross seeding in the posterior fossa (M2) and/or craniospinal axis (M3) who received concurrent chemotherapy [26]. The CSI dose was 36 Gy (RBE) and the posterior fossa boost was 18 Gy (RBE). Cardiac, thyroid, cochlea and bowel dose were less than 1 Gy (RBE), while the maximum esophagus dose was 15 Gy (RBE). There were no reported incidences of acute odynophagia, cough, dysphagia, nausea, vomiting, peptic pain or diarrhea reported in these three patients, and there were no grade 3–5 toxicities. There were also no treatment interruptions for neutropenia, anemia or thrombocytopenia [26].

A 2013 series reported the long-term outcomes for 15 children younger than 5 years of age with medulloblastoma and supratentorial primitive neuroectodermal tumors [27]. All patients were treated with surgery, chemotherapy and postoperative proton therapy. Eleven of 15 patients were treated with CSI and involved-field radiation (IFRT), while four were treated with IFRT alone. At a median follow-up of 39 months (range: 3–102 months), 13 of 15 patients were alive without evidence of disease. The 3-year local failure rate was 7.7% and the 3-year OS was 85.6% [27]. Late toxicities included measurable hearing loss on audiometric studies in nine of 13 patients. All patients received cisplatin before beginning proton therapy and five of these patients were found to have bilateral sensorineural hearing loss before the start of radiation. Three of 13 patients developed grade 2 endocrine dysfunction, requiring replacement with growth hormone, adrenocorticotropic hormone and thyroid-stimulating hormone. When these three patients with growth hormone deficiency were excluded from the analysis, there was no difference in age-adjusted height when compared with baseline. Neurocognitive testing in the 13 living patients demonstrated no difference from baseline [27].

• Craniopharyngioma

Craniopharyngiomas are locally aggressive benign tumors and represent 3–5% of pediatric intracranial malignancies. Because of the midline suprasellar location of the tumors, patients experience endocrine, visual, hematologic, neurological and neurocognitive effects of the tumor and long-term toxicities of treatment. Radiation has been used as the primary treatment or in conjunction with maximal safe resection, and tumor control rates approach 80–85%. Dosimetric analysis of IMRT, 3D proton therapy and IMPT for 14 children with a mean age of 5.1 years treated to 54 Gy demonstrated the potential of proton therapy to reduce the integral dose to the whole brain, whole body and critical structures when compared with IMRT (Figure 2) [28]. When 3D proton therapy and IMPT were compared across various studies, the IMPT plan delivered a significantly lower integral dose and dose to critical structures due to the decrease in secondary neutron dose [28–30]. Early clinical results of proton therapy have shown that it can be effective at reducing acute toxicities in craniopharyngioma. Of 13 patients treated with proton therapy, at a median follow-up of 8.4 months, progression-free survival (PFS) and OS rates were 85 and 100%, respectively. There were no grade 3 acute toxicities and one grade 2 toxicity. No late effects were noted [31].

Figure 2. . Proton therapy treatment plan for a child with a craniopharyngioma.

• Low-grade glioma

LGGs are the most common pediatric CNS tumors [32]. For low-grade brainstem gliomas in particular, radiation therapy is the standard treatment because of their unresectable nature and proximity to critical structures. Dosimetric plans using IMRT and proton therapy were generated for three patients with low-grade brainstem gliomas located in the mid-brain, pons and medulla (Figure 3). Proton plans used three coplanar beams and IMRT plans used 9–11 6-MV photon beams. Organ site and dose range were in accordance with Quantitative Analysis of Normal Tissue Effects in the Clinic guidelines. Both the IMRT and proton therapy plans delivered doses below the acceptable constraints. For the rostal tumors, the mean dose and volume of tissue treated in the pituitary and bilateral temporal lobes were significantly lower with proton therapy than IMRT. In a study by Brower et al., the proton therapy plan for the pontine LGG delivered a significantly lower dose to the right hippocampi, posterior nasopharynx, pituitary, hypothalamus and bilateral temporal lobes when compared with the IMRT plan [33].

Figure 3. . Proton therapy treatment plan for a child with a hypothalamic low-grade glioma.

Mature outcomes for proton therapy for nonbrainstem LGG were recently reported [34]. Thirty-six patients with a median age of 10.9 years (range: 2.5–21.5 years) and a median interval between diagnosis and radiotherapy of 2.3 years (range: 0.7–17.2 years) were irradiated to between 48.6 and 54 Gy (RBE). At a median follow-up of 6.8 years (range: 0.7–17.2 years), the 6- and 8-year PFS rates were 89.3 and 81.8%, respectively, and the 8-year OS rate was 100%. Subset analysis of patients younger than 7 years of age and those with high dose to the left temporal lobe showed a decline in IQ. Endocrinopathy at 6 years was 41.4%. Proton therapy was effective in LGG and allowed for some sparing of the hypothalamic–pituitary axis and left temporal lobes. Disease outcomes and late toxicities were favorable when compared with age-matched photon cohorts [34].

• Ependymoma

Ependymomas make up 8–10% of CNS tumors and the standard of care includes maximal safe resection followed by radiation therapy to the tumor cavity. The first series investigating proton therapy for ependymomas was published in 2008 [35]. Seventeen patients with grade II–III ependymoma ranging in age from 13 months to 12.8 years were treated with 3D proton therapy to a median dose of 55.8 Gy (RBE). When compared with IMRT, 3D proton therapy and IMPT plans decreased the mean dose to the temporal lobes (16, 4 and 2 Gy, respectively), pituitary (12, 0 and 0 Gy, respectively), hypothalamus (10.7, 0.2 and 0 Gy, respectively) and whole brain (Figure 4). The median follow-up was 26 months (range: 43 days to 78 months) from the start of proton therapy. Local control, PFS and OS rates were 86, 80 and 89%, respectively. There were no reported late toxicities [35]. An update of this study was presented in abstract form [36]. Seventy patients with a median age of 3.1 years (range: 3 months to 20 years) were treated with proton therapy. The median follow-up was 46 months and the 3-year local control, PFS and OS rates were 83, 76 and 95%, respectively. In a subset of 14 patients who were tested, mean IQ was 108.5 at baseline and 111.3 at the 2-year follow-up. A few patients experienced endocrinopathies and ototoxicities [36].

Figure 4. . Proton therapy treatment plan for a child with a posterior fossa ependymoma.

• Germ cell tumors

Germinomas and nongerminomatous intracranial germ cell tumors comprise 2–3% of pediatric CNS malignancies. In the modern era, radiation to the whole ventricular system and IFRT have been used in conjunction with platinum-based chemotherapy, with recently reported cure rates approaching 70–80%. MacDonald et al. reported the use of pencil-scanning proton therapy in 35 patients treated for intracranial germ cell tumors to a total mean dose of 23.4 Gy (RBE) [37]. IMRT, 3D proton therapy and IMPT plans were generated. Similar to prior studies, the proton therapy techniques demonstrated improvement in dose to the temporal lobes, normal brain and ocular structures with IMPT, delivering a slightly more conformal dose than 3D proton therapy (Figure 5). The median follow-up was 28 months (range: 13–97 months). The PFS rate was 95% and the OS rate was 100%. Two patients developed hypothyroidism and two developed growth hormone deficiency, requiring replacement. There were no neurocognitive deficits reported and the conclusion of the paper noted the improved OS, decrease in late toxicity and need for longer follow-up.

Figure 5. . Whole-ventricular-proton irradiation plus a tumor boost in a child with pineal germinoma.

• Atypical tertoid/rhabdoid

Atypical teratoid/rhabdoid tumors comprise 1–2% of intracranial pediatric brain tumors but 20% of tumors in patients younger than 1 year old who historically have a poor prognosis. A recent study by Chi et al. has shown an improvement in OS with multimodal therapy, including surgery, chemotherapy and radiation therapy [38]. Their encouraging results have led to studies using proton therapy for these tumors. De Amorim et al. treated ten patients ranging in age from 15 days to 19 years (median: 1.8 years), with proton therapy doses ranging from 50.4 Gy (RBE) to 55.8 Gy (RBE) delivered to the tumor cavity [39]. IMRT, 3D proton therapy and IMPT were compared, and proton therapy demonstrated an improvement in the dose to structures outside of the clinical target volume when compared with conventional photon radiation (Figure 6). The median follow-up was 27.3 months (range: 11.3–99.4 months) after the start of proton therapy. One patient died of disease and there were nine living patients. Acute toxicities included nausea and vomiting. Two of nine living patients developed central hypothyroidism, three had growth hormone deficiency and one had preradiation high-frequency sensorineural hearing loss [39].

Figure 6. . Proton therapy treatment plan for a child with an atypical teratoid/rhabdoid tumor.

• Glioblastoma

Glioblastoma comprises less than 5% of pediatric intracranial malignancies and is an aggressive tumor that confers a poor prognosis. There are little data for its treatment with proton therapy in the pediatric population. However, a study by Matsuda et al. reported the prognostic factors in 67 newly diagnosed adult glioblastoma patients who received either conventional photon radiation, high-dose particle radiotherapy with proton beam therapy or boron neutron capture therapy [40]. Patients who received proton beam therapy, boron neutron capture therapy and conventional photon radiotherapy received 96.6, 60 and 60–61.2 Gy, respectively. Surgical resection included gross total resection in 19%, subtotal resection in 70% and biopsy in 10% of patients. The OS rate with proton beam therapy or BCRT was 24.4 months (range: 18.2–30.5 months) compared with 14.2 months (range: 10.0–18.3 months) with conformal radiotherapy. A multivariate analysis of prognostic factors demonstrated that treatment with proton beam therapy and BCNT was significant. Extrapolating from this study in adults, proton beam therapy may have a role in pediatric patients, where preservation of the adjacent normal brain parenchyma is even more critical in children who will become long-term survivors.

• Quality of life & late toxicities

Because proton therapy delivers a decreased integral dose, the potential to improve health-related quality of life (QOL) continues to be assessed in the study setting, both during and after radiation therapy. Radiological studies evaluating changes in normal white matter in patients treated with proton therapy for craniopharyngiomas suggest minimal acute changes at 3 and 6 months [41].

Modern, conformal radiation therapy techniques for pediatric LGG and craniopharyngioma have resulted in improved neurocognitive patient outcomes, with no significant impairment in learning or memory [42]. Ototoxicity, a common side effect for pediatric CNS patients treated with chemoradiation, was evaluated in a prospective observational trial of patients treated with CSI for medulloblastoma using proton therapy. The rates of high-grade ototoxicity ranged from 5 to 8% [43,44]. When compared with similar cohorts treated with IMRT, grade 3–4 ototoxicity was similar. Nevertheless, more than 50% of patients experienced no ototoxicity with proton therapy compared with less than 25% of IMRT patients [45]. Hearing sensitivity significantly declined following proton therapy for all frequencies and, surprisingly, there was no correlation with predicted dose to the cochlea [43]. These studies are ongoing to determine possible improvements in QOL as well as the cost benefit from decreased ototoxicity.

A retrospective analysis of QOL and mood disorders in 23 patients treated with maximal safe resection followed by proton therapy for craniopharyngioma and their families found that 24–38% of patients had executive dysfunction and 47.8% of patients reported depression [46]. Lower QOL and executive dysfunction was found to be associated with depression and lower family socioeconomic status.

In a recent prospective observational study of 142 pediatric CNS patients, both patient- and parent-proxy QOL outcomes were collected. QOL was evaluated at baseline, during treatment, at the end of treatment and for up to 5 years of follow-up. QOL improved at 3 years when compared with QOL at the end of treatment and approached pretreatment QOL. When compared with QOL of children with noncancer chronic illnesses, QOL for these patients was higher. CSI was associated with decreased QOL when compared with focal radiation. Patient socioeconomic status and neurocognitive results were not evaluated and the patients were required to live near the proton center, thus introducing a bias [47].

• The economics of proton therapy

The improvement in QOL and late effects with proton therapy may be more cost-effective than photon therapy in pediatric patients who are expected to be cured of their disease, particularly when the management of potential second malignancies and late toxicities are evaluated [48]. In a Monte Carlo simulation model of survivors of pediatric medulloblastoma treated at 5 years of age with proton and photon therapy, LAR was estimated from previous studies, and model-valued health effects and costs were estimated per the US Panel on Cost Effectiveness in Health and Medicine. Proton therapy was found to be associated with a lower incremental cost–effectiveness ratio to society per quality-adjusted life years when compared with photon therapy [49]. A similar study in Sweden in 2005 demonstrated cost savings of €23,600 per patient and an additional 0.68 quality-adjusted life year per patient [50].

A proton center with a single-room gantry (the most common model currently under design) costs approximately US$25 million to build. Extensive time and training is required for all staff. With regards to gantry time, a simple pediatrics case requires 30 min [51], a complex case or a case with anesthesia requires 1 h [51] and CSI can require up to 1 h 25 min [52]. Assuming a model of 15-year financing with 5% interest, 14 h of gantry operation Monday through Friday with all allotted appointments filled and Medicaid/private reimbursement Indiana rates, a one-room facility would have to treat 12 patients a day for 15 years to pay off the debt [51]. While the principal impediments to constructing a large-scale proton facility include the cost of investment, declining payments and competition, the main barrier to entry for a single-room facility is developing efficient economies of scale. In the future, we may see fewer rotational gantries and more fixed-beam delivery systems. To reduce planning and treatment times, more efficient proton treatment planning software is being developed and more centers are using hypofractionated regimens. Given the availability of proton facilities, operating costs and cost of complex pediatric cases, it is important to determine who would benefit from proton therapy. The Agency for Healthcare Research and Quality recognizes pediatric solid tumors and tumors near critical structures, including the brain, eye and spinal cord, as commonly accepted indications for proton therapy [53].

Conclusion & future perspective

New proton technologies continue to be developed. Recent dosimetric studies have investigated field optimization in IMPT [29] and methods to minimize the effect of variations in daily set-up [54]. Preliminary clinical outcomes of ten pediatric CNS tumor patients with various pathologies treated with IMPT have been promising with no grade 2 acute toxicities [55].

The increasing evidence demonstrating normal tissue sparing and the potential for a decrease in late toxicities with proton radiation have sparked a debate on whether proton beam radiation should be the gold standard of care, in particular, for pediatric CSI [51]. The cost of a proton facility, limited pediatric study follow-up thus far and the question of the clinical effects of secondary neutron dose have been the main concerns with widespread adoption of proton therapy [56].

The Children’s Oncology Group Blueprint for Radiotherapy in 2012 outlines the goal of achieving reduction in tumor volumes, sparing dose to normal structures and enrolling eligible children into protocols with proton therapy [57]. As more pediatric patients are treated with proton therapy and followed, clinical outcomes and late toxicities will be further assessed. Advancements are needed in proton therapy quality assurance and standardization, and in the development of models for patient selection [58].

Acknowledgements

The authors would like to thank Jessica Kirwan for her editorial assistance.