Crossfire Proton-beam Treatments

In 1946, Wilson first proposed the clincial use of charged-particle beams because of their unique characteristics.¹⁴ Lars Leksell adressed the theoretical and many practical aspects of stereotactic radiosurgery in 1951.⁹ Using the Uppsala University cyclotron Leksell and Borje Larsson, a radiobiologist, used a cross fired proton beam in intial experiments in animals and in the first treatments of human patients.⁸ These treatments used the plataeu ionization portion of the beam's energy rather than the focal Bragg peak at its end.

Early Bragg-peak Proton Radiosurgery

In 1954, John Lawrence began to use the Berkely cyclotron's Bragg peak to irradiate the pituitaries of patients with metastatic breast cancer for hormonal suppression.¹² The first thirty patients were treated with protons and thereafter helium ions were used.

In 1961 Raymond Kjellberg began treating patients using the Bragg peak of protons from the Harvard Cyclotron Laboratory.⁷ This was soon followed by similar efforts led by V.S. Koroshkov in Moscow.

Pituitary lesioning and subsequently treatment of adenomas were the first successful applications of radiosurgery because of the ability to localize the sella turcica on plane radiographs. The main risks of such treatment was injury to the cranial nerves.¹³ Late hypopituitarism is also confirmed as an expected result of successful control of secretory and non-functioning adenomas.

Ateriovenous malformations were the first parenchymal lesions on which radiosurgery was extensively evaluated. Development of single-dose radiation for this type of lesion required determination of the tolerance of normal brain and of the brain surrounding AVMs to radiosurgical doses. Using a combination of clinical and experimental observations, Kjellberg proposed the standard dose effect curves for radiation necrosis in proton therapy of the brain.⁶ It is of particular note that Kjellberg's one percent dose-diameter line for radiation necrosis also serves as the basis for gamma knife and linear accelerator dosimetry.^{4, 11}

Initial attempts at proton radiosurgery were limited by neuroradiologic techniques which prevented successful three dimensional treatment planning. These limitations were first overcome for proton radiosurgery of the pituitary because of its midline symmetry and because of the presence of reliable bony landmarks visible on conventional radiographs. Stereotactic treatment of arteriovenous malformations began in 1963 and was based on a stereotactic guidance device and angiograms.⁶ Some tumors including skull base lesions could be adequately localized by pneumoencephelography. Leksell performed the first such treatment, radiating a vestibular schwannoma in 1969.¹⁰ Treatment of the majority of intracranial tumors required the ability to image three dimensionally and awaited the widespread availability of computed tomography and magnetic resonance imaging.

Intital efforts at beam delivery used conventional radiographs and stereotactic immobilization to identify targets. Three-dimensional stereotactic techniques were then applied to radiosurgery, but required continuous immobilization of the patient in the sterotactic apparatus from the time of imaging to the completion of the treatment. Transposition of three-dimensional imaging information to conventional X-ray stereotactic space was possible but somewhat inconvenient and occassionally inaccurate.3 More recently, implanted skull fiducials have been employed to allow reproducible correlation of conventional radiographs with three-dimensional imaging.^{1, 5} This makes fractionated proton therapy practicle and may allow a further increase in the risk-to-benefit ratio of particle beam radiosurgery.

Beam scanning is another technique under development to allow optimization of delivery. Current proposals involve using electromagnetic beam modulators to move the single Bragg peak through the entire treament volume rather than using a fixed number of static beams.

A patient positioning system known as STAR (stereotactic alignment for radiosurgery) is currently in use at the Harvard Cyclotron.² It uses the target-centered principle, allowing complete rotational freedom once the linear coordinates of the target have been defined. It is compatible with any orthogonal or radial stereotactic coordinate system and accepts targets obtained directly from computed tomography, magnetic resonance imaging, and angiography. This arrangement is required to allow the implementation of line-of-sight treament planning because the Harvard beam is limited to the horizontal position. Another solution to this challenge used by some new medically-dedicated particle beams are designs that allow protons to be delivery from arbitrary angles rather than from a horizontal beam.