Systematic Review and Meta-Analysis of Best Available Photon, Proton, and Carbon-Ion Radiotherapy in Various Head and Neck Cancers

Reviewer: Christine Hill-Kayser
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: May 28, 2010

Presenter: Dr. Janneke Grutters for Dr. Bram Ramaekers
Presenter's Affiliation: Maastricht Radiation Oncology (MAASTRO), Maastricht, The Netherlands


  • Cancers of the head and neck represent a heterogeneous group of malignancies. Although these cancers can and do metastasize to distant sites, local control has been demonstrated numerous times to be essential to overall survival.
  • For this reason, all patients who are being treated with curative intent undergo definitive surgical resection, radiotherapy, or a combination of the two, aimed to achieve durable local control.
  • In many cases, chemotherapy is added to these modalities; the addition of chemotherapy to concurrent radiotherapy has been demonstrated to improve both local control and overall survival in squamous cell carcinoma of the head and neck (Pignon, 2009).
  • Various radiotherapy regimens have also been explored; the most notable of these is alteration of fractionation (Fu, 2000). This type of treatment has been demonstrated to provide a local control benefit, even in the absence of chemotherapy.
  • Despite the essential role of radiotherapy in treatment of many cancers of the head and neck, it is associated with significant morbidity. The close proximity of vital structures within the head and neck, including the parotid glands, muscles of swallowing and speaking, brain, auditory/visual structures, and spinal cord, as well as reactivity of the mucosa to radiation, result in moderate-severe acute toxicity to most patients treated with conventional photon radiation.
  • This acute toxicity may translate to long-term toxicity, including xerostomia that may lead to severe dental injury, swallowing dysfunction that may result in feeding tube-dependence, speaking difficulties, and poor cosmetic outcomes.
  • The unique dosimetric properties of particle therapy are attractive for treatment of sites within the head and neck because of their associated increased conformality compared to photon-based radiation. This improved conformality may provide decreased dose to organs at risk while allowing delivery of adequate dose to tumor and areas of microscopic disease.
  • Despite these theoretic advantages, though, the efficacy of protons and carbon ions in treatment of head and neck cancers has not been demonstrated; this is largely due to lack of large, randomized, controlled trials comparing these treatment types to conventional photon therapy and/or intensity modulated radiation therapy using photons.
  • Although such studies might provide valuable answers in the future, they are not currently available, and may never be.
  • In order to address this need within the literature, the authors performed a systematic review and meta-analysis of the current data available regarding proton, carbon ion, and photon radiation for treatment of head and neck cancer, with a goal of considering the effectiveness of these three radiation modalities.


  • The authors performed a systematic review of the literature, including clinical and dosimetric (in silico) studies evaluating proton, carbon ion, and photon radiation in treatment of head and neck cancer.
  • Inclusion criteria included:
    • Published in or after 1990
    • Written in the English language
    • Evaluating efficacy of carbon ion, proton, or photon radiotherapy for head and neck cancer
    • Considering at least 10 patients in the case of clinical studies, and at least 5 patients in the case of dosimetric studies.
  • Outcomes considered were local control (LC), disease-free survival (DFS), overall survival (OS) and toxicity.
  • Patients were considered in subgroups:
    • Nasopharyngeal cancer
    • Oropharyngeal cancer
    • Paranasal/sinonasal cancer
    • Malignant mucosal melanoma
    • Adenoid cystic carcinoma
  • A meta-regression analysis was performed to compare LC, DFS, OS, and toxicity of particle therapy compared to the best available photon radiation (conformal or IMRT).
    • Patients were stratified according to tumor location and histology.
    • Corrections were performed to account for tumor stage within a given study, as well as the proportion of patients undergoing surgery.
    • P values < 0.05 were considered significant.


  • A total of 323 studies were initially identified evaluating treatment with proton or carbon ion therapy as treatment for head and neck cancer. Based on exclusion criteria, a total of 11 clinical studies (7 proton, 4 carbon ion) and 9 dosimetric studies were considered in this analysis. These were considered in subgroups based on tumor histology and site.
  • Over 1200 photon clinical studies were identified, and ultimately 70 of them were included in this analysis. Again, patients were considered in subgroups.
  • Outcomes compared for each subgroup of patients depended on the data available as follows:
  • Nasopharyngeal Cancer:
    • Protons (n = 17) were compared to IMRT (n = 1168)
    • 2 year LC was 92% with protons, and 95% with photons (p = NS)
    • 3 year LC was 74% with protons and 90% with photons (p = NS)
  • Oropharyngeal Cancer:
    • Protons (n = 29) were compared to IMRT (n = 297)
    • 2 year OS was 96% with protons and 95% with photons (p = NS)
    • 2 year DFS was 81% with protons and 87% with photons (p = NS)
  • Paranasal/Sinonasal Cancer
    • Carbon ions (n = 10), protons (n = 116), and IMRT (n = 109) were compared to one another.
    • 5-year LC was 49% with carbon ions, 88% with protons, and 66% with IMRT (p = 0.035, favoring protons).
    • 5-year DFS was 61% with protons and 54% with IMRT (p = NS). These data were not available after treatment with carbon ions.
    • 5-year OS was 71% with protons and 52% with IMRT (p = NS). These data were not available after treatment with carbon ions.
  • Malignant Mucosal Melanoma
    • Carbon ions (n = 157) were compared to conformal photon radiotherapy (n = 696).
    • 5 year overall survival was 44% with carbon ions versus 25% with photons (p = 0.007).
  • Adenoid Cystic Carcinoma
    • Carbon ions (n = 166), protons (n = 23), and conformal photon radiotherapy (n = 853) were compared.
    • 5 year LC was 69% with carbon ions, 93% with protons, and 75% with photons (p = NS).
    • 5 year OS was 70% with carbon ions, 89% with protons, and 73% with photons (p = NS).
  • The authors also considered toxicity according to the above subgroups:
  • In oropharynx/ nasopharynx, no xerostomia or dysphagia data was available with particle therapy.
  • In paranasal/sinonasal carcinoma, incidence of dysphagia after particle therapy was 0% compared to 9% with IMRT.
  • In malignant mucosal melanoma, no skin or mucosal toxicity reports were available.
  • In adenoid cystic carcinoma, the incidence of all low grade toxicity was comparable for carbon ions and conformal photon radiotherapy (63% vs. 64%). The incidence of high grade toxicity was increased with photon radiation (4-11%) compared to carbon ion therapy (0%).
  • The authors report that the dosimetric studies included evaluated conformality and dose to organs at risk with particle therapy versus photon therapy. All found decreased dose to organs at risk with similar target coverage with particle therapy.

Authors' Conclusions

  • The authors conclude that studies evaluating particle therapy in treatment of head and neck cancer are scarce, with limited patient numbers.
  • They conclude that particle therapy seems to be at least as good as photon radiation in the sites evaluated here, and appears to provide increased local control in the setting of paranasal/sinonasal tumors, and improved overall survival in the setting of malignant mucosal melanoma. They also describe consistent demonstration in dosimetric studies of improved sparing of organs at risk with adequate target coverage.
  • They note that all of the results presented here should be interpreted with caution do to small patient numbers and the lack of direct comparison. They suggest that an international patient registry tracking outcomes and toxicity would likely be to mutual benefit.

Scientific/Clinical Implications

  • Despite the limitations of the study presented here, the authors have performed a needed and useful exercise. As such, they present well synthesized, systematic data that may help guide clinical decision-making in the absence of definitive comparison data.
  • The authors' findings regarding the use of carbon ion therapy in treatment of malignant mucosal melanoma are perhaps of the most interest. They document significant improvement in overall survival using carbon ions, with relatively large patient numbers (157 and 696 treated with carbon ions and photons, respectively). From a radiobiologic standpoint, this improvement in overall survival with higher LET radiation is logical given our understanding of the radiobiology of melanoma cells, and this data may be practice-guiding. More information on sites of mucosal disease and regional nodes included in target volumes would be of interest.
  • The majority of findings in the other sites investigated did not demonstrate significant difference in outcome between particle therapy and photon therapy. This may be in part due to small patient numbers. For example, in the nasopharynx subgroup, 3 year local control was 90% with IMRT and 74% with protons, with 1168 and 17 patients in the two groups, respectively. We cannot conclude much from this data because of the very small number of patients in the proton arm; however, were this comparison better powered, a benefit of IMRT over protons may have been demonstrated.
  • Very little meaningful toxicity data is available in the analysis presented here, although reductions in dysphagia during sinus treatment, and high-grade toxicity after treatment for adenoid cystic carcinoma, with particle therapy are not surprising.
  • The authors have drawn appropriate conclusions from their data, recognizing the limitations of the project and data available. In the absence of stronger data, the authors have made a valuable contribution to the literature.
  • As they point out, all of the data presented here should be interpreted with caution, and considered hypothesis-generating. Ideally, further studies, which may include collaborative work as the authors suggest, will provide more comprehensive and conclusive data.

Stay informed with the latest information from OncoLink!   Subscribe to OncoLink eNews
View our newsletter archives