Proton beam radiation has distinct tissue-sparing advantages over conventional photon-based radiation, making it especially beneficial for treating localized head and neck cancers, which typically have complex anatomy and are located near vital organs.
We recently published a comprehensive review of its use in head and neck cancer in the journal Oral Oncology. (1) Our review summarizes recent developments in treatment planning and delivery, existing clinical evidence, and ongoing prospective trials in major head and neck cancers that may further increase the therapeutic window of proton therapy for patients with head and neck cancer.
At Memorial Sloan Kettering Cancer Center (MSK), we are dedicated to improving oncologic outcomes while minimizing the negative effects of cancer treatment. The New York Proton Center is a partnership between MSK, Montefiore Health System, and Mount Sinai Health System. Our alliance allows us to improve cancer care and advance clinical evidence for proton therapy.
Proton Therapy for Head and Neck Cancer
The advantages of proton therapy over photon-based radiation, such as intensity-modulated radiation therapy (IMRT), trace to its unique physical characteristics. Most of the dose is applied across a very narrow range of depth, called the “Bragg peak,” allowing for dose escalation for better tumor control and improved survival while reducing exposure to normal tissues. It is particularly beneficial when treating head and neck cancers since major structures such as the salivary glands, oral cavity structures, pharyngeal mucosa, larynx, spinal cord, and brain tissue lie close to tumors. Sparing normal tissue leads to significantly reduced toxicities and improved quality of life for patients.
Recent Developments in Treatment Planning and Delivery
Planning for proton therapy is distinct from photon therapy. Protons have a finite and defined depth of penetration with virtually no exit dose, forming a spread-out Bragg peak to cover the range of the target. Proton beam penetration is more affected by tissue type, and its delivery affects dose distribution more than those with photons. (2)
Optimizing beam direction is essential since fewer proton beams are required than with IMRT. The goal is to achieve the shortest paths to the distal tumor edge while minimizing passage through complex tissues. Critical structures must be avoided as biological effectiveness is higher at the end of the beam, and range can be uncertain. Planning for proton beam therapy takes heterogeneities of tissues into account, with consideration for how varying tissue densities may impede power. Potential setup errors and any uncertainties related to changes in internal and external anatomy must also be considered, as they may shift the Bragg peak location. (3)
There are two main modes of proton delivery. With passive scatter, the beam’s spread can be shaped by customizing the field aperture, beamline, and range compensator designs to ensure tumor coverage with appropriate three-dimensional margins. In pencil beam scanning, multiple beams from different directions produce the desired dose pattern. This mode is used in intensity-modulated proton therapy (IMPT), which is especially useful for treating head and neck cancers with complex or irregular shapes (4) because it enables depositing a dose with higher biological effectiveness in the target and away from normal tissues. (5)
Clinical Evidence in Head and Neck Cancers
Clinical evidence shows that proton therapy provides excellent organ sparing with minimal exit dose compared to IMRT when treating unilateral head and neck tumors as follows: (6)
- Dagan et al. found very low rates of mucosal toxicity and well-preserved weight and nutritional status throughout treatment in a prospective evaluation of a series of patients who received ipsilateral head and neck proton therapy for parotid tumors. (7)
- Holliday and colleagues reported promising local control of 93.8 percent at a median follow-up of two years in patients with head and neck adenoid cystic carcinoma treated with proton therapy after surgery. Twenty-five percent of patients had acute toxicities, and 6.3 percent had chronic grade 3 or 4 toxicities. (8)
- Romesser and colleagues compared toxic effects associated with unilateral proton therapy vs. IMRT in patients with major salivary gland tumors or cutaneous squamous cell carcinomas. Proton therapy afforded similar tumor target coverage but with negligible radiation to the oral cavity, contralateral major salivary glands, and brainstem compared with IMRT. Reduced dose to these organs at risk resulted in fewer acute toxicities, including mucositis, dysgeusia, nausea or vomiting, and fatigue. (9)
- Chuong et al. from the Proton Collaborative Group found lower rates of acute toxicities than historical rates with IMRT in the largest series of patients with salivary gland tumors. Among 105 patients (90 with parotid and 15 with submandibular tumors), the rates of acute side effects were as follows: nausea (1.5 percent), dysgeusia (4.8 percent), xerostomia (7.6 percent), mucositis (10.5 percent), and dysphagia (10.5 percent). (10)
Trials in progress: MSK is leading the NCT02923570 trial, in collaboration with the Mayo Clinic and Mount Sinai Hospital, comparing proton therapy with IMRT for unilateral radiation in head and neck cancer. The primary endpoint is acute grade 2 or higher mucositis. The NRG/Oncology/RTOG 1008 study (NCT01220583) is a phase II randomized trial of postoperative radiation with or without weekly cisplatin for salivary gland tumors. The protocol allows for proton therapy.
Radiation is essential for managing oropharyngeal cancer, in both definitive and adjuvant settings, as an addition to chemotherapy and surgery. IMRT has been used to mitigate late toxicity such as xerostomia, but adverse effects remain quite high, especially for patients undergoing concurrent chemoradiation. (11) Further reductions in toxicities are needed given the increasing numbers of patients diagnosed with human papillomavirus (HPV)-positive head and neck cancers, which are typically associated with long survival times post-treatment. (11) Oropharyngeal cancer is now the most common HPV-positive malignancy in the United States.
IMPT is generally recommended for oropharyngeal cancer to enhance conformality and homogeneity of the radiation plan while minimizing radiation to major organs at risk. Early studies on dosimetry have found that IMPT spared organs at risk better than IMRT, (12), (13) an advantage that likely applies to oropharyngeal cancer. Clinical evidence of note is as follows:
- Gunn et al. reported acute toxicities in 50 patients with oropharyngeal cancers (98 percent HPV-positive) treated with IMPT as follows: grade 3 mucositis (58 percent), grade 3 dysphagia (24 percent), grade 2 or worse xerostomia (25 percent). No grade 4 or 5 toxicities were observed. (14) A separate report comparing the same 50 patients with 100 case-matched patients with oropharyngeal cancer treated with IMRT by Blanchard and colleagues found no statistical difference in overall survival (OS) or progression-free survival (PFS). Survival rates in the proton therapy cohort were excellent, with a two-year OS of 94.5 percent and PFS of 88.6 percent. (12)
- Sio and colleagues compared prospectively collected outcomes for 35 patients treated with IMPT with 46 treated with IMRT. The objective symptom burden from the top 5 scores from the MD Anderson Symptom Inventory for Head and Neck Cancer (MDASI-HN) survey was worse with IMRT only in the subacute phase. (15)
- Sharma and colleagues found that patients treated with proton therapy received a significantly lower dose to many normal structures than those treated with IMRT. This finding was reflected in higher scores in quality of life measures and less xerostomia at six and 12 months after radiation. (16)
Trials in progress: The ongoing randomized phase III trial NCT01893307 aims to demonstrate at least non-inferiority of IMPT vs. IMRT for oncologic outcomes. It also aims to clarify IMPT’s ability to reduce toxicities through secondary endpoints, including patient-reported outcomes, physician-rated toxicity outcomes, and cost-benefit analyses. Globally, two other prospective randomized trials have started recruiting: The TORPEdO (TOxicity Reduction using Proton bEam therapy for Oropharyngeal cancer) trial by the UK National Health Service, and the ARTSCAN V trial led by investigators at the Lund University Hospital in Sweden.
IMRT is often used to treat nasopharyngeal cancer to provide adequate tumor coverage close to multiple organs at risk. It has demonstrated excellent locoregional control with minimal grade 3 or higher xerostomia. (17) However, IMRT results have been suboptimal for more advanced T4 tumors, radioresistant Epstein-Barr virus-negative tumors, and locoregional recurrence after irradiation. (18), (19), (20) Proton therapy has shown excellent locoregional control with less acute toxicities as follows:
- Chan and colleagues first reported clinical outcomes from a phase 2 trial using a combined proton/photon approach in 23 patients with locally advanced nasopharyngeal cancer. There was no local or regional relapse at a median follow-up of 28 months, and two patients developed distant metastases. Two-year disease-free survival (DFS) and OS were 90 percent and 100 percent, respectively, with no grade 4 or 5 toxicities. The most common grade 3 or higher toxicities were hearing loss and weight loss in 29 percent and 38 percent of patients, respectively, with no reports of grade 3 xerostomia. (18)
- Holliday et al. reported a lower G-tube placement frequency with IMPT vs. IMRT (20 percent vs. 65 percent, p = 0.02), likely due to the lower mean radiation dose to the oral cavity. (21)
- In a retrospective study of patients treated with proton therapy vs. IMRT for nasopharyngeal, nasal cavity, or paranasal sinus cancers, McDonald and colleagues found lower opioid use for pain control and a lower G-tube dependence rate, likely tracing to the lower mean dose to the oral cavity and esophagus. (22)
- At MSK, we recently conducted one of the largest studies to date comparing results for 28 patients with definitive nasopharyngeal cancer treated with IMPT with a matched IMRT cohort of 49 patients. The IMPT group received significantly lower doses to organs at risk and experienced fewer acute toxicities, such as fatigue, dysphagia, mucositis, and weight loss. We are currently preparing the manuscript for this study.
- Chan et al. reported late toxicities in a study of 17 patients treated with proton therapy for T4 nasopharyngeal cancer. There was only one local failure at a median follow-up of 43 months. Late toxicities included five patients with radiographic temporal lobe changes, one with endocrine dysfunction, and one with mandibular osteoradionecrosis. (23)
Trials in progress: Longer follow-up of existing series (24), (25), (26) and updated prospective data may confirm the promising potential of proton therapy to improve oncologic outcomes and reduce late toxicities. The ongoing NGR Oncology HN001 study, a phase II/III prospective randomized trial using plasma Epstein-Barr Virus DNA as a biomarker to direct therapy allows proton therapy and is collecting quality of life data.
Sinonasal cancers include squamous cell carcinomas, adenoid cystic carcinoma, and olfactory neuroblastoma. Most of these cancers are treated by surgical resection followed by radiation with or without chemotherapy. Complex anatomy involving multiple paranasal sinuses and the nasal cavity, and proximity to the skull base and other critical structures makes treatment exceptionally challenging with suboptimal outcomes. (27)
Studies comparing dosimetry have shown that IMPT may provide better avoidance of critical structures than IMRT, which typically irradiates ipsilateral optic structures beyond acceptable dose constraints. (28), (29), (30)
- Two studies have shown late toxicities more than five years after receiving proton therapy. Zenda and colleagues reported late toxicities in 90 patients with sinonasal cancers, with a median follow-up to 57.5 months. They observed grade 3 late toxicities in 17 patients (19 percent) and grade 4 toxicities in 6 patients (7 percent), including four instances of optic nerve disorder. (35) Russo et al. reported their experience over two decades treating patients with locally advanced sinonasal squamous cell carcinoma. Among 54 patients with stage III or IV disease, local control rates at two and five years were both 80 percent, while two-year OS and five-year OS were 67 percent and 47 percent, respectively. Nine grade 3 and six grade 4 toxicities were reported; most were wound toxicities. (36)
- Yu and colleagues recently reported a multi-institutional experience from a registry study of the Proton Collaborative Group. Sixty-nine patients with sinonasal cancer were treated with curative intent IMPT, including 27 patients treated with reirradiation. Among 42 patients who received initial proton radiation, the three-year freedom from locoregional recurrence rate and OS were 93 percent and 100 percent, respectively, greatly exceeding any previously reported outcome in other series. No patients developed vision loss or symptomatic brain necrosis. Grade 3 or higher toxicities occurred in only 15 percent of patients, including patients treated with reirradiation. (37)
- At MSK, we recently reported results from our experience treating 86 consecutive patients with sinonasal cancer with proton therapy. IMPT significantly improved local control compared to 3D conformal proton techniques (91 percent vs. 72 percent, p < 0.01). At a median follow-up of 23 months, two-year local control, DFS, and OS were 83 percent, 74 percent, and 81 percent, respectively, for radiation-naïve patients, and 77 percent, 54 percent, and 66 percent for patients undergoing reirradiation. Grade 3 toxicities were lower than historical controls. (38)
- Finally, a systematic review and metaanalysis of charged particle therapy vs. proton therapy for paranasal sinus and nasal cavity cancers by Patel et al. found that OS was significantly improved at five years with charged particle therapy. Notably, their subgroup analysis found improved local control and higher DFS at five years for patients who received proton therapy vs. IMRT, (39) providing high-level evidence that proton therapy can improve survival in select patients.
Local or regional recurrence of head and neck cancer requires salvage treatment since uncontrolled disease significantly impacts the quality of life. Salvage therapy may include surgery followed by reirradiation or reirradiation alone.
Radiation oncologists must balance increasing doses for disease control while avoiding nearby local tissues that have likely already received significant radiation in the initial treatment. IMRT has typically been the most common technique used for reirradiation. Our experience at MSK among 257 patients treated with IMRT showed a two-year locoregional control of 47 percent and an OS of 43 percent. (40) An increased dose, greater than 50 Gray (Gy), was independently associated with improved locoregional control. The one- to two-year locoregional control rates in the available literature range from about 50 to 60 percent with IMRT. (41)
Proton therapy has an acceptable toxicity profile in the reirradiation setting for recurrent or new primary head and neck cancer, allowing radiologists to escalate doses to achieve higher locoregional control while avoiding previously treated tissue. Recent clinical evidence is as follows:
- Romesser and colleagues reported a one-year locoregional failure rate of 25.1 percent and OS of 65.2 percent in a multi-institutional study. These proton therapy results were comparable to IMRT reirradiation yet with lower grade 3 or 4 toxicities: Among 92 patients, only six patients experienced grade 3 or 4 skin complications (8.7 percent), four patients had grade 3 dysphagia (7.1 percent), and there were two cases of grade 5 bleeding (2.9 percent). (42)
- In a retrospective analysis by Phan et al. among 60 patients treated with reirradiation proton therapy, locoregional control was 68 percent at one year, with 30 percent of patients experiencing grade 3 toxicities, including 22 percent who required a feeding tube. However, late grade 3 toxicities decreased to 16.7 percent, with only two percent of patients remaining tube-dependent, and three patients may have died from reirradiation toxicities. (43)
- McDonald and colleagues reported a series of 61 patients treated with curative intent proton beam reirradiation for skull base cancers. The two-year local failure rate was 19.7 percent, with 14.7 percent acute and 24.6 percent late grade 3 or higher toxicities, including three treatment-related deaths. The three to five percent death rates underscore the need for careful patient selection and balancing the need for disease control while mitigating toxicity, especially in patients with larger retreatment volumes. (44)
- A few smaller series of retrospective studies have shown that proton reirradiation can provide favorable locoregional control and OS at acceptable rates of both acute and late toxicities in nasopharyngeal and oral cavity cancer. (45), (46)
- For patients with locally recurrent head and neck cancer who are ineligible for reirradiation as salvage therapy, options are extremely limited. A recent study at MSK investigated last-line treatment with the Quad Shot (QS) regimen, which involves treatment with 3.7 Gy, twice daily over two consecutive days at four-week intervals per cycle, for up to four cycles. Among 166 patients, the overall palliative response rate was 66 percent, and symptoms improved in 60 percent. Proton therapy, a Karnofsky Performance Status Score greater than 70, palliative response, and receiving three to four QS cycles were associated with improved local PFS and OS. The grade 3 toxicity rate was 10.8 percent, and no grade 4 or 5 toxicities occurred. (47) Palliative QS appears to be an effective last-line therapy for patients with previously irradiated head and neck cancer.
Research in progress: At MSK, we have treated nearly 350 patients with proton reirradiation for head and neck malignancies, including 300 squamous cell carcinomas. Our investigators will report on our patient experience soon.
Despite proton therapy having a dosimetry advantage to photons for sparing normal structures outside of the treatment field, toxicities may still occur within or very close to the target. Multiple studies have shown progressively higher relative biological effectiveness at the distal edge of the proton Bragg peak, (48), (49) which may land accidentally in critical organs due to the combination of suboptimal beam arrangement and range uncertainty.
As proton therapy is used for reirradiation, there are higher risks of normal tissue injury due to significant previous radiation contributing to potential subacute and late toxicities. Central nervous system (CNS) toxicities can develop years later, be quite symptomatic, and significantly affect long term survivors’ quality of life. Temporal lobe necrosis is often seen after head and neck irradiation to the skull base, nasopharynx, and paranasal sinuses. Overall, there is an estimated five percent risk of developing late CNS toxicities after proton therapy for head and neck cancer. Notable clinical evidence of CNS toxicities is as follows:
- McDonald et al. found that any grade of temporal lobe necrosis approached 12.7 percent at three years, with grade 2 or higher radiation necrosis occurring in 5.7 percent of patients. (50)
- A recent report by Kitpanit et al. found a two-year temporal lobe necrosis rate of 4.6 percent among 234 patients with a median time to necrosis of 20.9 months. (51)
Patients treated for skull base tumors may experience brain stem necrosis. Investigators have collected substantial dosimetry data for brainstem injury after proton therapy in pediatric patients and contributed to guidelines defining parameters for safe delivery of proton radiation near the brainstem. (52)
Radiation-induced optic neuropathy (RION) can have a significant impact on patients’ quality of life. Consider the following evidence:
- Kountouri and colleagues reported a 6.5 percent risk of any grade of RION in a study of 216 patients with head and neck cancers treated with pencil beam scanning proton therapy. (53)
- In a large series of 514 patients treated over three decades at a single institution, the RION incidence was one percent for patients who received less than 60 Gy compared to 5.8 percent of patients who received more than 60 Gy to optic pathways. (54)
Non-CNS late toxicities after proton therapy have also been studied:
- Zhang et al. retrospectively reviewed more than 500 patients with oropharyngeal cancer treated at MD Anderson. They found lower mandibular doses for 50 patients treated with IMPT (mean 25.6 vs. 41.2 Gy; p = 0.001), and lower osteoradionecrosis rates with IMPT; (two percent vs. 7.7 percent). (55)
- A series of patients re-irradiated with particle therapy, including proton therapy, reported a carotid blowout rate of 2.7 percent, similar to the photon series rate. (56)
Advancing Proton Therapy Research for Head and Neck Cancer
Radiation oncologists, physicists, and biologists continue to collaborate to improve proton therapy. Based on current evidence and promising trends, we anticipate that proton beam therapy will become more effective and less toxic.
Multiple institutions are studying outcomes for patients with head and neck cancer treated with proton therapy. Table 1 in our paper summarizes the trial names, inclusions, treatments, and primary endpoints for 15 ongoing clinical trials. Three of the studies led by MSK are:
- A clinical trial of endoscopic surgery followed by chemotherapy and proton radiation for the treatment of tumors in the sinus and nasal passages (NCT03274414)
- A phase II study of proton re-irradiation for recurrent head and neck cancer (NCT03217188)
A phase II randomized study of proton beam versus photon beam radiotherapy in the treatment of unilateral head and neck cancer, in collaboration with the Mayo Clinic and Mount Sinai Hospital (NCT02923570).