Efficacy and Safety of Spot Scanning Proton Radiation Therapy for Skull Base Tumors: First Long Term Report of the PSI Experience

Reviewer: Eric Shinohara MD, MSCI
Abramson Cancer Center of the University of Pennsylvania
Last Modified: May 23, 2008

Presenter: Carmen Ares
Presenter's Affiliation: Paul Scherrer Institute
Type of Session: Scientific


Chordomas (C) are rare, slow growing tumors which are thought to arise from remnants of the notochord. Though slow growing, they can be locally aggressive and tend to involve the axial skeleton, often involving the base of skull. Chondrosarcomas (CS) are malignant cartilaginous tumors which account for approximately 6% of base of skull tumors. 

Surgical resection is generally done initially for debulking and to obtain a diagnosis. Complete resection is preferable, however, for the reasons mentioned above, this is often difficult. As there is often residual disease post operative radiation therapy is often used.  Photon therapy has not resulted in good local control rates in C or CS. Due to the involvement of the base of skull, it can be difficult to get adequate dose to the tumor while still sparing critical structures such as the brainstem, optic nerves and chiasm. 
Charged particle therapy has the potential to improve the local control in C and CS due to their greater conformality, which may allow higher radiation doses to be delivered safely. The present study investigates the safety and efficacy of a spot scanning proton beam in the treatment of C and CS of the base of skull. 

Materials and Methods

The present study is a retrospective single institution study of the use of spot scanning proton therapy in the treatment of C and low and intermediate grade CS of the base of skull.

The present study examined outcomes from 64 patients, 42 patients with low or intermediate grade C and 22 patients with CS of the base of skull treated at the Paul Scherrer Institute (PSI). Patients were treated with a degraded 590 MeV proton beam using a spot scanning technique. All patients had gross residual disease at the time of radiation. C were treated to a median dose of 73.8 Gy (relative biological effect (RBE)) with a range from 67.6 to 74 Gy (RBE). CS were treated to a median dose of 68.4 Gy (RBE) with a range from 63-74 Gy (RBE). Volumes were defined using both CT and MRI scans. Five patients were treated with a mix of protons and photons. Patients were treated in 2 Gy fractions. Constraints used were 64 Gy for the surface of the midbrain, 53 Gy for the center of the midbrain, and 60 Gy for the optic nerves and chiasm.     
Kaplan-Meier analysis was used to calculate local control (LC), disease specific survival (DSS), and overall survival (OS). Toxicity was assessed using the Common Terminology Criteria (CTCAE v.3.0).



LC: 87% at three years and 81% at 5 years
There were 5 local failures in the C group, one male and four females.   No regional or distant failures were noted.   
DSS: 90% at three years and 81% at 5 years
            Four children died of brainstem compression
            OS: 62% at five years.
LC: 95% at three years and 94% at 5 years
There was 1 local failure noted in the CS group.   No regional or distant failures were noted.   
DSS: 100% at three years and 100% at 5 years
            OS: 91% at five years.

All Patients:
Median Age for all patients: 44 years

Median Follow up: 38 months (minimum follow up of 1 year; range 14-92 months)
Prognosis: Patients with a tumor volume of greater than 25 ml (p=0.03) and those with brainstem compression at the time of radiation (p=0.007) had significantly worse local control rates.
Five patients had low grade toxicity (changes on MRI without symptoms)
Four patients had high grade neurotoxicity, 2 with optic nerve toxicity (one with grade 3 and one with grade 4) and 2 with radiation necrosis. The two patients who had grade 2 temporal lobe radiation necrosis responded well symptomatically to steroids. No patients had brainstem toxicity.   There was a 94% freedom from toxicity at 5 years. No predictors for toxicity were found. No secondary malignancies have been noted as of last follow up.

Author's Conclusions

1)      The toxicity seen with passive scatter proton treatment of C and CS of the base of skull is comparable to that seen with spot scanning beam proton therapy when the same target and constraints are used.
2)      Most toxicity related to the treatment is seen within two years of the treatment
3)      This study demonstrated the feasibility of spot scanning beam protons in the treatment of C and CS of the base of skull and should prompt further studies of its use in these tumors.

Clinical/Scientific Implications

The present study presents the basis for further studies investigating the use of scanning proton beams in the treatment of C and CS of the base of skull. Given the location of these tumors, in close proximity to a number of critical structures, it is imperative to find ways to improve conformality and limit toxicity. Proton beam therapy allows for greater conformality compared with photon based therapy. There should be a further improvement in conformality with spot scanning beam plans compared with passive scatter proton beams. This could allow further dose escalation to further improve local control. Furthermore, spot scanning proton beams may reduce the neutron dose to normal tissues reducing the risk of secondary malignancy. It is important to note that the range of proton beams do have a level of uncertainty which must be accounted for in treatment planning to avoid inadvertent overdosing of critical structures.