Module 5: Clinical Outcomes by Disease Site - The use of proton therapy in the treatment of cancers of the head and neck

Eric Shinohara MD, MSCI
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 15, 2009

The morbidity associated with the treatment of head and neck cancer with protons and conventional photons has been reviewed at various institutions. Specifically, cancers of the paranasal sinuses, tonsillar region, and nasopharynx have been evaluated. In each of these cancers, proton therapy should result in an improvement of local control with a reduction in the morbidity associated with conventional photon treatment. There has been a significant reduction in the rates of blindness seen in the treatment of paranasal sinus tumors. Also, comparative plans for the treatment of tonsillar and nasopharyngeal cancer revealed proton beam therapy can deliver higher doses of to the tumor volumes with significantly reduced radiation to the salivary glands and mandible than can photon beam irradiation. This results in a decreased incidence of xerostomia and radionecrosis of the mandible.

It should be noted that essentially 100% of all patients treated for head and neck cancer with x-rays will experience severe xerostomia (dry mouth), which although it may not be life-threatening, severely impairs quality of life. Many of these patients are unable to eat in a restaurant since they may require their food to be pureed or specially prepared for them to be able to eat. It is these sorts of poor quality of life outcomes that are very inadequately measured in current cancer statistics where the only measure of outcome is survival. Patients may be alive, but at considerable personal cost. This complication, xerostomia, is the sort of complication that significant with standard X-ray and IMRT treatment. This is based on the through and through penetrating nature requiring us to treat both parotid glands even for well lateralized lesions, and which can be totally avoided with protons because of their lack of an "exit" dose. Given a choice of cure with or without xerostomia patients will make compelling argument for exploring the use of protons over conventional X-rays.

There are several potential advantages to proton therapy compared with IMRT in the treatment of patients with head and neck cancers. As discussed above, the major advantage is likely to be a decrease in the late side effects associated with radiation therapy. Head and neck cancers are already taken to a high dose, and escalation beyond what is already being used has the potential to increase late toxicities. Hence, protons are most likely to improve our ability to get adequate dose to difficult-to-treat areas (i.e. base of skull), and this may be where the greatest benefit is seen.

There are potential disadvantages for proton use in the head and neck, including the proton’s range dependence on the daily changes in air cavity density (which must be accounted for); increased skin reaction; and the potential to actually be too conformal. It is possible that there may be more marginal misses with the greater conformality afforded by proton therapy, and it is very difficult to salvage such patients. Hence, defining new and adequate margins in proton therapy will be critical.

The specific head and neck sites that may benefit the most include the nasopharynx, oropharynx, minor salivary glands, nasal cavity and paranasal sinuses. In the oropharynx, the greatest benefit is likely to be decreased xerostomia due to greater sparing of the contralateral parotid. A proton boost could potentially decrease the mean contralateral parotid dose to less than 26 Gy which should decrease xerostomia. Additionally, in tumors of the oropharynx, it may be possible to decrease dose to the brainstem and spinal cord substantially.

There have been numerous dosimetric studies which have demonstrated better conformality with protons, but what are the true long term outcomes and side effects? There are only limited long term data available for head and neck proton therapy and that these studies are limited by a number of factors including small patient numbers and variable sites included in the studies.

Proton data by disease site with comparisons with patients treated with photons is presented below (Mendenhall WM PTCOG 2008):

Cancers of the Nasal cavity and Paranasal Sinuses

University of Florida (Photon data): From 1964-2005, 109 patients were treated for nasal cavity or paranasal sinus tumors. Of these patients, 56 patients were treated with definitive radiation therapy. 96 patients (88%) were treated using altered fractionation. The median follow up was 9.4 years. 5-year data are shown below:

T1-T3 tumors

T4 tumors


Local Control












Overall Survival




In the definitive radiation group 16% of patients had severe complications (the majority involved loss of vision) where as 25% of patients treated with surgery and radiation had severe complications.

Massachusetts General Hospital (Proton data): 91 patients with paranasal sinus tumors were treated to a median dose of 73.6 CGE (mixture of protons and photons; median proportion of protons was 49%) with a similar proportion of patients treated with altered fractionation compared with the University of Florida data above (~88%). All patients had carcinomas (82) or sarcomas (9). 87% were treated with accelerated hyperfractionation and 35% received adjuvant chemotherapy. Median follow up was 45 months. 5-year outcomes are as follows: Local control was 82%, disease free survival 52%, and overall survival was 58%. There were 4 patients who had changes on brain MRI. Two patients developed soft tissue necrosis and another developed bone necrosis. 5 year outcomes:

Chiba University (Proton data): 14 patients with Esthesioneuroblastoma were treated from 1999-2005 with protons to 65 CGE in 2.5 CGE fractions. Median follow-up was 40 months. 5-year outcomes: Local control was 84%, progression free survival 71% and overall survival 93%. There was one patient with apparent bone necrosis, but there were no other grade 3-4 complications.

When the proton results are compared with the prior Florida photon data it appears that local control rates were similar or better while the late complication rates were lower for the proton based therapy.

Adenoid Cyst Carcinoma (ACC)

University of Florida (Photon Data): 101 patients with a de novo diagnosis of ACC were treated from 1966-2001. 57 patients had T1-T3 disease and 44 patients had T4 disease. Median follow up was 6.6 years. Thelocal control rate at 5 years was approximately 77%. Local control for T1-T2 lesions was 92% versus 64% for T3-T4 lesions. Local control for T4 lesions treated with surgery alone was 44% as compared to surgery plus radiation where the local control was 93%. Six patients developed ipsilateral blindness, three developed osteoradionecrosis which required surgery, one patient required permanent PEG, and one patient developed and oral-antral fistula. One patient developed fatal meningitis after salvage surgery and one developed fatal hemorrhage after reconstruction of the trachea.

Example of a proton dose distribution for a maxillary sinus tumor - Mendenhall WM, PTCOG 2008

Massachusetts General Hospital (Proton data): 23 patients with the de novo diagnosis of skull base ACC were accrued from 1991-2002. 48% had biopsy alone, 39% had subtotal resection and 13% had a gross total resection. Median follow up was 64 months and a median dose of 75.9 CGE was delivered. The local control rate at 5 years was 93%, distant metastasis free survival was 62% and overall survival was 77%. There were no grade 5 visual complications and one grade 4 retinopathy. Seven patients developed a chronic seizure disorder after therapy, which was controlled with medication. One patient developed a fistula with a CSF leak and meningitis.


Loma Linda (Proton data): From 1991-2002, 29 patients with Stage II-IV disease were treated. Patients were treated to 75.9 CGE in 45 fractions over the course of five and a half weeks. Follow up ranged from 2-90 months. At five years, local control was 88% and neck control was 96%. Loco-regional control was 84% and late grade 3 toxicity occurred in 10% of patients (3/29). There were no cases of osteoradionecrosis. . Shown in the table below are the Loma Linda proton data compared with photon data from the University of Florida:

Number of Patients

% T4 patients

% Stage IV

5 year Local Control

Late Toxicity

Loma Linda






University of Florida (tonsil)






University of Florida (base of tongue)






The Loma Linda Data appeared to show less late toxicity with greater local control.

Overall protons may prove to be most useful in tumors of the base of skull to reduce visual complications and CNS toxicity. It may be possible to improve local control with dose escalation. Protons may also be useful in oropharyngeal cancers to reduce late effects, specifically xerostomia.

Links to reviews of recent abstracts and presentations regarding proton therapy for cancers of the head and neck:

« Previous Module | Next Module »