Particle therapy first emerged in 1946 under physicist Dr. Robert Wilson who recognized the potential therapeutic advantage of protons.1 The clinical use of protons did not begin until the mid-1950s in nuclear physics research facilities with limited application and few areas of the body that were treated.2 Therapy using heavier ions, such as carbon, started in 1972.2 Since then, proton therapy has flourished across the United States with 42 proton centers to date that are currently operational and several others under different phases of implementation and/or construction.3 Carbon ion treatment centers are more limited with 12 centers across 5 countries. There are no centers currently operating within North America, however the first is under construction at Mayo Clinic in Jacksonville, Florida.1

Particle therapy has gained traction due to the development of modern treatment delivery technologies, and the unique physics and radiobiologic properties of these particles.1 Protons and carbon ions are charged particles with a defined mass wherein their range is proportional to their energy.2 Protons differ from photons in that they deliver low doses of radiation upon entering the body and up to shortly before the end of their range where they deliver their maximum dose with a sharp fall off (Bragg peak) beyond which there is no exit dose.2 This minimizes dose to surrounding normal tissues by reducing entrance dose and eliminating exit dose. Carbon ions also deposit most of the energy in the Bragg peak, but have some exit radiation dose created by nuclear fragmentation of the carbon ions. As the majority of the dose is deposited at the Bragg peak, the integral dose to the body is decreased by 50-60% compared to photon therapy.4 The Bragg peak can be shifted in depth via range modulation then spread out through passive scattering or delivered with pencil scanning (PBS) with magnets.5 PBS allows for decreased entry dose/skin dose, lower neutron scatter, reduced-need for beam-shaping, and intensity-modulated proton beam radiotherapy (PBRT). PBS has largely replaced scattered or 3D conformal proton therapy and though apertures can still be helpful to provide sharp dose fall off at the edge of fields, they are not routinely used or required.4 Protons and carbon ions have a high linear energy transfer (LET) yielding larger mean energy per unit compared to photon therapy (Fig. 1).2 This higher LET leads to increased clusters of DNA damage with a shorter range of radiation deposition, causing more damage at the targeted site and sparing normal surrounding tissues.2,6 While protons have a modestly higher LET compared to photons, carbon ions differ from protons as they have a much higher LET, thus they deposit energy distally due to nuclear fragmentation and have a sharper lateral penumbra at greater depths.6 In contrast to conventional radiotherapy, carbon ion radiation allows for little repair of potentially lethal damage, and there has been great interest in its use for radioresistant tumors, particularly radioresistant sarcomas. At this time, there are no phase III clinical trials that show superiority of carbon therapy over proton therapy.7

Although clinical benefits continue to be debated over the use of protons and carbon ions, and we await phase III clinical data from many studies, it is clear that these particles reduce integral dose. This is important for young patients who will be at greatest risk for late effects from the lower doses of radiation, and for those with large tumors or undergoing treatment with large fields (ie, craniospinal irradiation), where there will be the greatest absolute reduction in integral dose. For patients requiring very high doses of radiation, which is necessary for many sarcomas, the decreased higher integral dose may reduce adverse effects of radiation. It remains to be determined if the reduction in integral dose and other physical and biological effects of these particles confer a measurable clinical benefit compared to photons.

Some limitations of proton therapy and carbon ions are the difference in LET and the possibility that the impact of LET may be tissue-specific. Little data is available on the biological impact of these differences. Although the slightly increased LET of proton therapy is accounted in a prescription dose equivalent to photons, Gy (RBE), this estimate is based on limited data and it is known that the general radiobiological effectiveness of 1.1 used for this equation is variable at different locations (varies from approximately 0.9-1.27 and possibly higher). Software systems have been developed to estimate LET for each individual radiation plan, but these systems are early in use and require validation. The clinical data for protons is reassuring so far as the distal end of the proton beam is not placed in critical structures and multiple beams are used for plans going to a high dose. Much less clinical data is available for carbon ions, but it is known that the LET is much higher. This is potentially beneficial for tumor cell kill but requires careful attention to critical structures and monitoring for adverse effects in healthy tissues.

Chordomas are rare primary bone tumors with an incidence of 0.1 for every 100,000 individuals, and a prevalence of less than 1 for every 100,000. Chordomas account for only 1%-4% of bone cancers and 17% of all primary bone tumors of the spine.8, 9, 10 Predominantly affecting Caucasians, chordomas are thought to arise from remnants of the embryonic notochord in the axial skeleton, most frequently in the sacrum, followed by the skull base and the mobile spine.11,12 The peak incidence of chordomas is the 5th-6th decade of life, with a median age of diagnosis of 60.8 Chordomas are rare in patients under the age of 40, and fewer than 5% of cases present in children and adolescents.13 In this age group, chordomas most commonly originate from the skull base, followed by the mobile spine and the sacrum. Chordomas are typically slow growing, but locally invasive and destructive; historically reported chordoma survival rates are low, with a median survival of around 6 years, and 5-, 10-, and 20-year survival rates rapidly declining from 67.6% to 39.9% and 13.1% respectively based on SEER data. Inferior outcomes have been reported in younger patients, with children ≤ 5 years of age often experiencing particularly poor outcomes due to atypical histologic features (poorly-differentiated chordoma), which have a particularly aggressive clinical course.14,15 For these patients, chemotherapy can be effective and it is often administered prior to definitive surgery and radiation with the most common chemotherapeutic agents being vincristine, doxorubicin, and cyclophosphamide alternating with ifosfamide and etoposide. Tazemetostat, a selective EZH2 inhibitor, has also shown activity for poorly differentiated chordomas.16

As chordomas are particularly resistant to chemotherapy and radiation, treatment is challenging. While surgical resection greatly improves outcomes, it is often hampered because chordomas are frequently not diagnosed until they extend beyond the primary affected bone into neighboring tissues.17 Most studies have found that negative margin resection is the treatment of choice for chordomas, as it greatly improves local control. In cases of subtotal resection, recurrence rates have been recorded as high as 50-100% compared to 0-53% when en bloc resection is performed.18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28 However, due to chordoma's typical close proximity to critical structures such as the brainstem, nerve roots, and spinal cord, en bloc resection with clear surgical margins is often not feasible without causing the patient significant neurological morbidity. Reported incidence of metastasis in chordoma cases is between 3-48%. Metastases are more common in primary sacral chordomas, likely due to the increased size of tumors at this site.29,30

Patients with chordomas typically undergo postoperative radiotherapy following resection, although traditional x-ray radiotherapy only offers palliative benefits and does not improve median survival when a dose of 50-60 Gy is feasible. In this scenario, overall recurrence rates with x-ray radiotherapy are 80%-100%, and for this reason, higher doses of radiation are required. If this cannot be achieved with photons without unacceptable side effects, then particle therapy alone or in combination with photon therapy is the preferred radiation modality. Chordomas of the mobile and the axial skeleton are treated either with preoperative radiation therapy (19.8-50.4 Gy RBE) to reduce the risk of tumor seeding at surgery followed by maximal safe resection and adjuvant postoperative radiation (Fig. 2) or with postoperative radiation therapy alone (76 Gy RBE).31,32

To date, there are no large prospective studies or randomized trials for chordomas and radiation therapy (RT). Therefore, there is limited guidance for tumor management. Given that 30%-40% of chordomas originate at the base of the skull, numerous critical structures such as cranial nerves, brainstem, optic chiasm, and spinal cord render high dose delivery without toxicity difficult with conventional 3D RT.18,33,34 However, charged particle RT has marked a positive shift in treatment outcomes for chordomas.35 Proton radiation therapy (PRT), or mixed proton and photon therapies, results in improved local control (LC) rates, ranging from 67.4% to 87.5% at 3 years, 46% to 73% at 5 years, and 54% at 10 years.36, 37, 38, 39 Of note, our institution recently published a long-term experience with pediatric skull base chordomas demonstrating that 10-year overall survival (OS) and progression-free survival (PFS) rates were 78% and 69%, and 20-year OS and PFS rates were 64% and 64%, in 204 children and adolescents with chordoma, when treated with proton or mixed radiotherapy after surgical resection.40

At our institution, for base of skull chordomas, radiation to a cumulative dose of 76 Gy (RBE) is delivered to the GTV in 38 fractions, while the GTV should include the tumor bed (any bone in contact with the initial tumor prior to surgery) and any residual gross disease, whereas the CTV should contain this target volume and an additional margin for microscopic extension of disease of about 2 mm. For spinal chordomas, the initial CTV includes the gross tumor volume and tissues suspected of subclinical tumor invasion. For patients undergoing preoperative radiation, this includes the gross tumor with 1 cm or more of presacral pelvic soft tissue margin on extraosseous tumor, as well as grossly involved vertebrae plus one vertebra above and below to cover potential spinal canal or dorsal venous extension. Where this CTV would extend beyond a fascial barrier (ie, pleura or peritoneum), the volume is reduced to encompass but not extend beyond the fascia. Biopsy sites are also included in CTV. For sacral tumors, with a local failure pattern that includes infiltration along the paraspinal, gluteal and piriformis musculatures, sacral nerve roots, or sciatic nerves, more generous margins of up to 2 cm or more along high-risk planes of spread are recommended. For patients undergoing postoperative irradiation alone, CTV includes surgically manipulated tissues including scars, drain sites, and stabilization hardware because of high risk of tumor implant. Because of chordoma's high T2 signal, fusion with T2 sequence is highly recommended.

The dose to target volumes should be delivered in such a manner as to respect constraints on adjacent normal structures. These constraints are as follows: the maximum dose to the surface of the spinal cord or brain stem shall not exceed 67 Gy (RBE), and the dose to the center of the spinal cord or brain stem should not be higher than 55 Gy (RBE). The maximum dose to the optic chiasm and both optic nerves should not exceed 62 Gy (RBE). It is critical that the portions in contact with the brainstem and chiasm be surgically removed, as the dose required to control chordomas is higher than the tolerance of these structures. Despite advancements made in the past decade in our knowledge of chordoma biology, systemic therapies still offer a limited benefit. Thus, patients with progressive disease should be encouraged to participate in clinical trials when and where available.

Many patients are not good candidates for surgical resection due to tumor location, size, proximity to vital structures, or comorbid conditions. Approximately 35%-65% of sacral chordomas and 21% of mobile spine chordomas are amenable to en bloc surgical resection with adequate surgical margins, given the proximity of the tumor to adjacent critical structures.19 There is a limited but growing body of literature supporting the use of definitive high dose radiation therapy in cases where en bloc resection of chordoma is not medically feasible or declined by the patient.41,42 In a retrospective analysis of 188 patients with biopsy-proven chordoma treated with carbon ion radiation alone to a total dose of 67.2 to 73.6 Gy (RBE), Imai et al.42 found that overall survival and local control at 5 years were 81.1% and 77.2%, respectively. Our institution recently reported their experience with 67 patients with unresectable spinal or sacral chordoma treated with definitive proton radiotherapy.41 Our retrospective analysis showed that high-dose proton radiotherapy offers favorable survival and local control for appropriate chordoma patients compared to surgical and lower dose RT interventions.41 A trend towards improved disease-free survival with doses >78 Gy (RBE) was observed.41 The overall survival at 5 and 8 years were 83.5% and 65.9%, respectively, in this series, which compare favorably with other surgical and radiological series, as do the local control at 5 and 8 years of 81.2% and 62.3%, respectively.41

Chondrosarcomas are rare tumors of the cartilage that often arise in the long bones or pelvis, where they can be managed with surgery alone. However, chondrosarcomas at the base of the skull and the axial skeleton can be challenging to control with surgery alone.43 Although many chondrosarcomas are low-grade neoplasms, they are locally destructive and when located next to critical structures, such as at the base of the skull or spine, they can have a high rate of recurrence after surgery.44 They may occur at any age and most patients do not have pre-existing predispositions or conditions. However, the tumors arise more often in patients with Ollier's Disease and Maffucci Syndrome, which are both disorders characterized by multiple benign enchondromas.45 Chondrosarcomas are typically conventional (myxoid/hyaline) but may also be described as de-differentiated and mesenchymal. They are graded by a 3-tier grading system with grades 1 and 2 behaving similarly (more indolent), whereas grade 3 and mesenchymal tumors, while less common, behave aggressively with high rates of metastasis and resistance to chemotherapy. Interestingly, chondrosarcomas are more indolent when they originate in the base of the skull, but behave more aggressively when their origin is in other skeletal sites.46

Historically, chondrosarcomas of the base of the skull were treated with surgery alone. Complete resection alone can be curative, but is difficult to achieve without substantial morbidity in the skull base. Traditional radiation therapy was found to be ineffective due to the inability to use 3D-conformal radiation therapy to deliver high radiation doses safely, which exceeded adjacent normal tissue tolerances. However, higher doses of radiation, in the range of 70 Gy (RBE) that are possible to deliver safely with protons showed very high control rates when combined with surgical resection.47

With regards to radiation timing, chondrosarcomas of the base of the skull are treated with postoperative radiation alone. On the other hand, for spinal chondrosarcomas, except where immediate posterior decompression is undertaken for relief of spinal cord or cauda compression, it is recommended to deliver some preoperative irradiation (19.8 Gy-50.4 Gy RBE) to reduce the risk of tumor seeding at surgery. This allows a substantially smaller CTV than exclusive postoperative radiation, because the CTV will not include surgical scars, uninvolved vertebrae instrumented with stabilization hardware, and drain sites.

Similar to spinal chordomas, radiation is delivered to spinal chondrosarcomas as described above with doses to the elective CTV encompassing the tumor bed and adjacent tissues considered at risk for subclinical disease. More recently, series of chordomas and chondrosarcomas treated with carbon ions and advanced proton radiation that can deliver similar biologically effective doses have also demonstrated excellent disease control.48, 49, 50 Optimal care for chondrosarcoma of the skull-base now generally consists of maximal resection of the portions of tumor displacing the brainstem or optic structures, followed by high dose particle radiation to the portion of the tumor that is more difficult to resect. Patients with low-grade chondrosarcoma usually have superior long-term survival compared to those with chordoma. There is a paucity of studies with robust outcomes, as most publications are limited to case series with relatively short follow-up.

Ewing sarcoma is the second most common malignant bone tumor in children and adolescents. However, around a third of cases of Ewing sarcoma present as extra-osseous tumors. Approximately 15%-30% of patients with Ewing sarcoma have metastatic disease at presentation, which is the main determinant of clinical outcome. Common metastatic sites include lungs, bone, and bone marrow.51,52 Most patients presenting with localized Ewing sarcoma harbor micro-metastatic disease, therefore chemotherapy is routinely used in the management of all Ewing sarcoma patients.

When complete surgical resection of Ewing sarcoma can be achieved without unacceptable morbidity, it is the preferred treatment and radiation therapy is not utilized because of the risk of increased second malignancy. Radiation therapy is indicated for inoperable or partially resected tumors.53,54 Historical evidence suggests a superiority of surgical resection over RT for the local control of Ewing sarcoma, but surgery is more often utilized for tumors in the appendicular skeleton, which may be more amenable to local control than tumors in the axial skeleton (ie, spine), which are most often treated with radiation therapy. Local failure rates following primary RT range from 10% to 25%. Patients with metastatic disease receive additional local treatment to metastatic sites after completion of chemotherapy, which includes whole lung irradiation in patients with pulmonary metastases. A meta-analysis comparing the 2 modalities revealed that although a higher incidence of local failure was found with radiation therapy, there was no significant difference in overall survival.55 Radiotherapy in preoperative and postoperative settings has been shown to improve local control for Ewing sarcoma.56

The standard dose for definitive radiation for Ewing sarcoma is 55.8 Gy (RBE) at 1.8 Gy (RBE) per fraction, although higher doses have been used in adults. The initial extent of disease should be treated to a dose of approximately 45 Gy adjusting for the shifting of organs or tissues after regression of the soft tissue portion of the tumor following chemotherapy, but the target volume for the tumor involving bone should not change following chemotherapy. Proton radiation is routinely used in children to limit the integral dose and decrease late effects in normal tissues. It may also be beneficial in adults, particularly young adults. Common locations for definitive RT include the pelvis, sacrum, and mobile spine. Proton radiation limits radiation dose to healthy normal tissues in the chest, abdomen, and pelvis and often facilitates complete sparing of ovaries for female patients.57,58 For those patients requiring whole lung irradiation, the standard dose is 15 Gy (RBE) at 1.5 Gy (RBE) per fraction. This is more commonly delivered with photons, but protons have been investigated for breast tissue sparing.59

Osteosarcoma is the most common primary bone tumor in children, accounting for 3% of pediatric malignancies. En bloc resection of the tumor with negative margins is the gold standard of treatment. Preoperative RT may be recommended in patients with tumors at high risk for resection with microscopic positive margins, while adjuvant radiation is typically indicated in the setting of positive margins or for patients at high risk for local relapse.60 Definitive radiation up to 76 Gy is the preferred dose for inoperable tumors, while microscopic disease should be treated with a radiation dose > 64.8 Gy. Margins are dependent on location and anatomic constraints. With the development of RT techniques and the incorporation of particle-based treatment, including proton RT and carbon ion radiotherapy, unresectable or incompletely resected osteosarcomas can be safely and effectively treated (Fig. 3).61,62 In some high-risk situations, higher doses can be delivered in adult populations.63

Retroperitoneal sarcomas make up approximately 15% of all soft tissue sarcomas. Given their location, complex anatomy, proximity to critical normal tissues, and typically large size (>10 cm), they can be challenging to treat.64,65 Histopathologies vary with liposarcomas and leiomyosarcomas being the most common, but less common histologies with differing biology are also seen.66 Curative treatment requires surgical intervention.67 Wide surgical margins for retroperitoneal sarcomas are challenging to achieve and even when nominally negative margins are achieved (in ∼50% of cases), local recurrence (LR) is frequent, particularly in patients with liposarcomas where local recurrences occur in 30%-40% of patients and represent a leading cause of death.64, 65, 66 EORTC-62092 (STRASS) is a phase 3 randomized controlled trial including sites from Europe and North America, which randomized patients to preoperative photon radiation and surgery, compared to surgery alone.67 They recruited 266 patients with a median follow-up of 43.1 months.67 Median dose was 50.4 Gy due to dose constraints for normal tissue tolerance including small bowel, kidney, and liver.67 Median abdominal recurrence-free survival was 4.5 years in the radiotherapy and surgery arm, and 5 years in the surgery alone arm (HR 1.01, 95% CI 0.71-1.44).67 The clinical trial defined progression during radiation therapy as abdominal recurrence even when patients had a subsequent margin negative (R0) recurrence. On sensitivity analysis, when local recurrence was defined only when a tumor recurred after surgery, there were twice as many local relapses noted in the surgery alone group for liposarcomas.68 Additionally, serious adverse effects were found to be higher in the radiation arm (24% vs 10%). Based on these results, the authors did not recommend preoperative radiotherapy for retroperitoneal sarcomas.67 However, others interpret the results to support the use of neoadjuvant radiotherapy for retroperitoneal liposarcomas.68 Of note, Gronchi et al.69 have highlighted that disease recurrence patterns for retroperitoneal sarcomas are related to histology, where liposarcomas have a higher local recurrence rate as compared to other histologies, while dedifferentiated liposarcomas and high grade leiomyosarcomas have higher rates of distant metastases. These differences in patterns of failure by histology types were also observed in the STRASS trial, emphasizing the importance of histology in considering whether to deliver radiation therapy.

Yoon et al. described that over half of patients with RPS have macroscopically and microscopically positive margins after maximally safe resection.65 Therefore, given the potential benefits for local control and the unique physical properties of protons, there is a potential role for dose escalation with protons to meet dose constraints to surrounding organs at risk for retroperitoneal sarcomas. DeLaney et al.64 conducted a phase I/II study of protons for retroperitoneal sarcomas with dose escalation for positive margins with the aim of reducing local recurrence. In the phase I portion of the study, patients were treated to 50.4 Gy (RBE) to CTV1 (gross tumor volume and at risk of subclinical disease) with simultaneous integrate boost (SIB) to CTV2 (areas at risk for positive margins) of 60.2, 61.6, and 63 Gy (RBE).64 The purpose of the phase I study was to determine maximum tolerated dose for further study in phase II. Eleven patients received treatment with protons with dose escalation without experiencing dose limiting toxicity. At 18 months follow-up, there were no local recurrences.64 These results show promise for dose escalation through the use of protons or heavy ions.

Cardiac soft tissue sarcomas are a rare and aggressive malignancy with a poor median survival of less than one year.70 They originate from the atria (majority of cases), ventricles, interventricular septum, or heart valves.70 The literature concerning optimal care is limited, but the photon literature notes improved progression free survival (PFS) with the addition of radiation.70 Data is more limited for proton and heavy-ion therapies.

Radiation therapy can be challenging to deliver for cardiac treatment in the setting of cardiac contraction and respiratory motion, 2 motions that are not in-sync with one another. The need for motion management and cardiac target tracking are crucial to target the tumor correctly. Accounting for motion while irradiating the heart has been accomplished mainly using an expanded target volume called the internal target volume (ITV), with or without the use of respiratory gating through breath hold techniques. In the setting of proton and heavy-ion therapy, a more precise technique has been used through cardiac gating by delivering treatment at a specific point in the cardiac cycle that leads to further OAR sparing through not requiring an ITV. In one proton study, the investigators looked at the advantages of this technique by evaluating the dose-volume metrics of cardiac gating through ECG with respiratory gating, referred to as cardiac-respiratory double gating (CRDG).71 When comparing CRDG to respiratory single gating (RSG) plans, they found that the median treated volume reduction was 21.7 cm3 and 23.3 cm3 for diastole and systole, respectively.71 RSG plans were found to irradiate a larger cardiac volume due to the need for an ITV to compensate for cardiac movement.71 Additionally, there was a reduction in mean OAR dose with the use of CRDG, with the extent of dose sparing varying based on the location of the tumor.71 Of note, currently there are no studies that have investigated the optimal cardiac gating window with radiotherapy delivery. Other areas of interest include optimizing the number of fields to increase conformality and optimizing beam geometry for motion. Additionally, the authors express caution with regards to the timing of dose delivery, due to various timing components of the proton accelerator.71

Rhabdomyosarcomas (RMS) are soft tissue sarcomas that are present in children and adolescents, and account for approximately 5% of all pediatric malignancies. Treatment regimens include systemic therapy, and local therapy with surgery and/or radiotherapy.72 Definitive radiation ranging from 50.4 Gy (RBE) to 59.4 Gy (RBE) is standard of care for children with inoperable tumors. Slightly higher doses may be considered for aggressive tumors in adults, as a worse prognosis is observed in older patients. The omission of radiation for select patients with RMS leads to a poorer prognosis, especially in high risk disease, and radiation is recommended for children of all ages.72 Given the young patient age and the variety of locations where this tumor can arise, the use of PBRT can be utilized to limit normal tissue dose, which is expected to limit late radiation-associated complications. A recent meta-analysis reviewed the use of PBRT for rhabdomyosarcoma across 11 studies from the United States, Japan, and Switzerland.72 These studies included 392 patients with embryonal sub-type, 144 patients with alveolar sub-type, and 8 patients with other conditions.72 Sixty-four percent of patients had unfavorable sites and 36% had favorable sites. LC and PFS at 5 years were 84% and 76%, respectively.72 Acute and late toxicity were mainly grades 1 to 2. Acute grade 3 toxicity was observed in 6 studies with incidence ranging from 9% to 25%.72 Late grade 3 toxicity was documented in 7 studies with an incidence of 2.1%-26%.72 This meta-analysis confirmed that PBRT is a feasible, safe, and effective modality for RMS with promising rates of LC, OS, PFS, and toxicity levels, although superiority could not be established.72 Because of the clear reduction in integral dose, randomized studies of protons versus photons in pediatric patients are not likely to be conducted due to ethical concerns, and it is expected that an increasing number of pediatric patients undergoing curative radiotherapy will be treated with protons wherever they are available.