Stereotactic cranial radiosurgery

INTRODUCTION — The potential utility of ionizing radiation to treat cancer was recognized shortly after the discovery of X-rays, which can cause extensive damage to DNA.

Prior to the development of stereotactic techniques, radiation was delivered to the cancer and a significant volume of surrounding normal tissue. Therapeutic efficacy was based upon the increased DNA-repair capacity of normal cells compared with tumor cells following radiation exposure. The total dose of radiation was delivered in small fractions over a period of weeks to allow normal tissue to preferentially recover or heal itself between doses compared with the tumor so that the cumulative damage to tumor cells was much greater. This fractionated approach is known as radiation therapy (RT) (figure 1).

Major advances in stereotactic localization, noninvasive neuroimaging, and radiation physics made it possible to selectively irradiate a sharply defined target, largely sparing the surrounding normal tissue. This approach, called stereotactic radiosurgery (SRS), is achieved by focusing multiple radiation beams on the tumor tissue from different directions (figure 1).

The biologic differences between fractionated RT and SRS and the technology of administering SRS are reviewed here. The applications of SRS for various tumors (both malignant and nonmalignant) are discussed under separate headings. Similarly the potential complications of cranial SRS and the application of SRS to extracranial sites are discussed separately. (See "Radiation therapy techniques in cancer treatment", section on 'Stereotactic radiation therapy techniques' and "Acute complications of cranial irradiation" and "Delayed complications of cranial irradiation".)

DEFINITIONS

Fractionated (conventional) RT — Fractionated or conventional radiation therapy (RT) refers to the repeated administration of small doses of radiation to a relatively large target, as in whole-brain RT or focal (involved-field) RT. Fractionation of the total dose minimizes damage to normal tissues by allowing time for repair of damage to DNA and maximizes the killing of tumor cells. Conventional dose fractionation schemes for intracranial lesions typically consist of 1.8 to 2 Gy in daily sessions with cumulative doses of 30 to 60 Gy [1].

SRS — Stereotactic radiosurgery (SRS) refers to the delivery of a single or very limited number of relatively large doses of radiation administered to a small, precisely-defined target. This is achieved by using multiple, non-parallel radiation beams that converge on the target lesion (figure 1) [2]. The full therapeutic dose is limited to the area where all of the beams overlap, while non-target areas receive much smaller doses from one or a limited number of the radiation beams. SRS thus requires accurate localization of the lesion and patient positioning during treatment.

The SRS dose is defined by the amount of radiation given at the margin of the targeted lesion; this is referred to as the prescription dose. Typically, the marginal dose delivered in a single session ranges from 11 Gy (for the treatment of benign lesions) to as much as 70 Gy (for thalamotomy, in the treatment of movement disorders) [3,4]. The amount of radiation absorbed by tissues in adjacent nontarget areas decreases rapidly with increasing distance from the target.

In some instances, the SRS dose is divided into a very limited number of fractions, up to a maximum of five. Although this was referred to as hypofractionated RT at one time, this technique is now generally included in the definition of SRS [5]. The term stereotactic RT is now reserved for highly fractionated schedules that rely upon the differential sensitivity of tumor and normal tissue to radiation.

RADIOBIOLOGY

Fractionated RT — Small, daily fractions of radiation cumulatively damage rapidly proliferating tumor cells more than normal tissues. The theoretical framework for dose fractionation was established in the 1960s [1,6]. The basic principles are often referred to as the four R's:

●Repair – Small doses of radiation cause sublethal levels of DNA damage, and normal tissues repair this damage more effectively than tumor cells do.

●Reoxygenation – Tumors often contain areas of hypoxia that cause resistance to radiation-induced cell killing. Dose fractionation can improve circulation and oxygenation within the tumor, thereby maximizing the effects of radiation.

●Redistribution and Repopulation – Tumor cells proceed through the cell cycle differently from normal cells following irradiation. Most normal cells linger in the S phase of the cell cycle after exposure to radiation; during the S phase, cells are highly resistant to further damage [7]. By contrast, tumor cells continue to proceed through the cell cycle (redistribute) and proliferate (repopulate) despite radiation exposure. Thus, the tumor cells move out of the radioresistant phase of the cell cycle and are more radiosensitive than normal cells during subsequent treatments.

SRS — Stereotactic radiosurgery (SRS) delivers a single or very limited number of fractions of high dose of radiation to a well-defined tumor volume, rather than repeated small fractions to both cancer and normal cells. Protection of normal tissue is achieved through a progressive, steep decline in the dose outside the treatment target.

The focusing of radiation on the target results from the convergence of multiple, non-parallel beams of radiation (figure 1) [2]. The dose of radiation received by the normal tissue in each path of the beam is small except where the beams converge. As an analogy, if everyone in an unlit, crowded stadium focused a flashlight on a single point, the point of convergence would be bright while the rest of the stadium remained relatively dark.

The prerequisites for SRS include precise delineation of the target using neuroimaging, an understanding of the neuroanatomy of the brain surrounding the lesion, and the technology required to deliver radiation reliably and precisely to the lesion. The planning and administration of SRS thus requires close collaboration between radiation oncologists, neurosurgeons, and medical physicists.

The cellular processes triggered by single or a few high-dose radiation fractions are poorly understood but appear to differ from those of much smaller fractionated doses of radiation. Impairment of DNA repair, redistribution, repopulation, and reoxygenation are less important with SRS than with fractionated radiation therapy (RT). (See 'Fractionated RT' above.)

This is illustrated by the clinical results of treating brain metastases using either of the two methods. Whereas whole-brain RT is only marginally effective for radioresistant tumors such as melanoma and renal cell carcinoma, SRS is as effective for these tumors as it is for radiosensitive tumors such as breast and lung. (See "Overview of the treatment of brain metastases", section on 'Efficacy of SRS alone'.)

The effectiveness of SRS cannot be explained entirely by the large dose of radiation simply killing all the tumor cells. Several observations support a more complex mechanism of action, possibly involving the host immune response [8]:

●The dose used to control cerebral metastasis (14 to 20 Gy) does not sterilize tumor cell lines in vitro. Following such irradiation in vitro, 0.01 to 1 percent of tumor cells survive, and the remaining tumor cells repopulate the tissue culture plate within days to weeks.

●Biopsies after SRS typically reveal tumor cells, although active tumor growth is rarely observed.

●Complete tissue destruction in vivo requires radiation doses significantly higher than those required for local tumor control. As an example, more than 70 Gy of radiation is required to achieve complete cell killing in thalamotomy for the treatment of movement disorders [4], whereas only 14 to 20 Gy of radiation is required to achieve local control in the treatment of cerebral metastasis [3,9].

●In patients with cerebral metastases that are surgically removed because of progressive growth more than five months after SRS, pathology shows a moderate to intense inflammatory cell response [10]. In contrast, when progression occurs less than five months after SRS, such a response is absent or limited.

Dose homogeneity — With fractionated RT, delivery of a homogeneous dose of radiation within the target volume is highly desirable. Uneven distribution of radiation doses may increase killing of tumor cells, but also may increase destruction of normal tissue [11-13]. Since the goal of SRS is to treat well-circumscribed lesions with little or no normal parenchyma in the target, dose homogeneity is less critical.

SRS VERSUS CONVENTIONAL RT — Multiple factors influence the decision of whether to use stereotactic radiosurgery (SRS) or conventional radiation therapy (RT). These include the volume of the target lesion, its proximity to cranial nerves such as the optic nerves, and the specific area of the brain to be irradiated.

Tumor volume — As the size of the target lesion for SRS increases, incidental irradiation of surrounding normal tissue also increases. This may be important since a much higher dose of irradiation per fraction is administered with SRS compared with conventional RT.

A dose-escalation study conducted by the Radiation Therapy Oncology Group defined the maximally tolerated SRS dose in the treatment of cerebral metastasis as a function of tumor size [14]. The recommended marginal doses of SRS were 24, 18, and 15 Gy for lesions ≤2 cm, 2 to 3 cm, and 3 to 4 cm in the largest diameter. SRS was not recommended for lesions >4 cm because adequate control could not be achieved without an unacceptable level of radiation toxicity to surrounding normal tissue.

Given these restraints, larger intracranial lesions that require radiation doses exceeding that which can be considered safe are increasingly treated with hypofractionated SRS or with staged SRS targeting distinct geographic regions of the lesion.

Proximity to cranial nerves — The close proximity of a target to cranial nerves can cause radiation neurotoxicity, despite the steep decrease in dose outside the intended target. Factors that increase the risk of damage to cranial nerves include previous surgery or radiation, a large volume irradiated, and a high total radiation dose [15,16]. Fractionated RT should be considered when SRS may jeopardize cranial nerve function.

Cranial nerves II and VIII are more sensitive to radiation injury than the other cranial nerves. SRS is generally avoided if the maximal dose delivered to the optic nerve exceeds 10 Gy [3,17-19]. Assessment of damage to cranial nerve VIII with SRS is difficult since it is often radiated during treatment of vestibular schwannoma (acoustic neuroma). In this situation, damage can be due to the tumor and/or treatment.

The proximity of target lesions to cranial nerves other than II and VIII is not a major factor in deciding between SRS and RT since other cranial nerves are more resistant to damage:

●Neuropathies of cranial nerves III, IV, and VI have not been reported for doses <15 Gy delivered to the cavernous sinus, while doses of 15 to 40 Gy are associated with injury in 10 to 15 percent of cases [20].

●Cranial nerve VII routinely receives doses of 11 to 15 Gy with SRS for vestibular schwannoma, but facial neuropathy is rare (less than 1 percent in a series of 829 patients followed for 10 years) [21].

●SRS for jugular foramen schwannomas or skull base meningiomas has been associated with less than a 2 percent incidence of neuropathy involving cranial nerves IX, X, or XI using doses of 8 to 12 Gy [22-25].

Location of the lesion — The risk of developing permanent damage following SRS varies dramatically with the location of the lesion in the brain. This was illustrated by a series of 422 patients treated with SRS for arteriovenous malformations, 85 of whom (20 percent) developed significant toxicity [26]. In a multivariate analysis, the maximum risk of neurologic complications was seen when the lesions were located in the deep gray matter (thalamus, basal ganglia) or brainstem (pons, midbrain), while complications were least likely with lesions in the frontal and temporal lobes. For this reason, fractionated RT is often preferred to SRS for the treatment of lesions in the deep gray matter or the brainstem.

Neurocognitive function — A major consideration in the selection of SRS versus conventional RT is the impact of RT on neurocognitive function (NCF). (See "Delayed complications of cranial irradiation", section on 'Neurocognitive effects'.)

The general consensus is that administration of whole-brain RT is associated with an increased risk of NCF decline compared with SRS. Thus, SRS is generally preferred in the treatment of focal intracranial lesions that are limited in size and number and located in a favorable location. The potential value of SRS is illustrated by two randomized trials that directly compare these modalities:

●In one trial, 213 adult patients with brain metastasis were randomly assigned to SRS or SRS plus whole-brain RT [27]. The primary endpoint was cognitive deterioration, defined as a decline of greater than 1 standard deviation from baseline on at least one of seven NCF tests three months after radiation administration. The results indicate that NCF decline was less likely in the SRS cohort relative to the SRS plus whole-brain RT cohort for memory as well as executive functions.

●In another trial, 58 adult patients with brain metastasis were randomized to SRS or SRS plus whole-brain RT [28]. The primary endpoint of this study was NCF assessment using seven NCF tests four months after radiation administration. While the trial endpoint required 90 patients, the trial was stopped early due to the observation that patients assigned to the whole-brain RT plus SRS group were significantly more likely to exhibit a decline in memory.

The effect of whole-brain RT on NCF is further examined by a review of randomized controlled studies that applied standardized NCF tests to lung cancer patients randomized to prophylactic whole-brain RT versus no prophylaxis [29]. Two pertinent studies were identified in this regard:

●In one trial, 356 patients with stage IIIA/B non-small cell lung cancer were randomized to 30 Gy (over 15 fractions) or to no prophylaxis. NCF was a secondary endpoint. There was a near quadrupling in the risk of NCF decline for the radiated patients compared with those not receiving RT (22 versus 6 percent) [30].

●In another trial, 264 small cell lung cancer patients were randomized to 25 Gy (in 10 fractions) or 36 Gy (in 18 fractions). NCF was the primary endpoint. There was a significant increase in the proportion of patients with NCF decline in the 36 Gy cohort, suggesting a dose response in terms of the effect of whole-brain RT on NCF [31].

These studies suggest that irradiation of normal cerebrum significantly impacts NCF in a dose-dependent manner. While SRS affords conformal radiation delivery, "spillage" of radiation dose to the surrounding normal brain tissue inevitably occurs. For patients who undergo a single round of SRS for smaller brain metastasis, the calculated "spillage" dose is sufficiently low that the risk of NCF decline is likely negligible. However, in patients who undergo SRS for larger lesions or are treated with multiple rounds of SRS, the "spillage" dose may accumulate and trigger deleterious NCF effects [32].

Another consideration in patients who undergo immunotherapy is that such therapy exaggerates radiation-associated inflammatory responses, which potentially affects NCF [33]. Unfortunately, there are little robust clinical data to address these risk considerations at the present time. While the available data suggest that NCF of long-term cancer survivors with brain metastasis tends to be poor [34], it is difficult to tease out the relative contributions of "chemo-brain," radiation effect, and delayed physiologic consequences of injury associated with brain metastases [35].

TECHNIQUE

Stereotactic guidance — Precision in target localization is a prerequisite for successful SRS.

Historically, this has been accomplished by the application of a stereotactic head frame using four pins that attach to the outer table of the skull (figure 2). The head frame placement is done with local anesthesia, although sedation may be required. This method has typically been used for photon-based SRS (X-ray and gamma ray radiation).

Modern Linac-based radiosurgical systems now regularly employ online cone beam computed tomography (CT) scanning for precision localization, which eliminates the need for skeletal fixation of the patient's head. Other frameless, image-guided stereotactic systems have also been developed for use with X-ray radiation sources. These include the CyberKnife and the ExacTrac X-Ray 6D systems, obviating the need for cranial fiducials (see 'CyberKnife' below and 'ExacTrac Robotic system' below). With all of these noninvasive systems, motion is minimized by application of an individualized frame or mask.

At present, proton-beam radiosurgery employs metallic fiducials that are percutaneously implanted in the outer table of the skull. Used in conjunction with a noninvasive frame and orthogonal X-rays at the time of treatment, these provide the highly precise targeting necessary for proton-beam radiosurgery [36]. Online cone beam CT scanning, as described above, is anticipated to supplant the use of imbedded cranial fiducials for proton-beam radiosurgery as well.

Radiation delivery — Both gamma rays and X-rays are photon radiation. As photons penetrate tissue, energy deposition decreases exponentially with depth below the surface and the radiation passes entirely through the tissue (figure 3 and image 1).

Several different systems are available for photon-based SRS. The most widely used are the Gamma Knife and Linac, which have similar efficacy. This was illustrated in a multicenter clinical trial that combined SRS with whole-brain radiation therapy (RT) for the treatment of brain metastases; no differences were observed in either efficacy or toxicity in patients treated with the two systems [37].

Gamma Knife — The Gamma Knife system consists an array of more than 200 cobalt-60 sources surrounded by an 18,000 kg shield (figure 4). The sources are oriented such that all the beams converge at a single point termed the isocenter. This array produces a target accuracy between 0.1 and 1 mm, which is at least as good as the best possible lesion delineation with current imaging technology [38].

During treatment, the patient is positioned so that the target coincides with the isocenter of the Gamma Knife unit. Using techniques of beam blocking, multiple or overlapping isocenters, and differential isocenter weighting, the radiation volume is approximated to that of the target lesion (figure 4).

Historically, Gamma Knife was administered as a single-session procedure because of the system-required frame placement. However, the latest iteration of the Gamma Knife, the ICON, incorporated frameless stereotaxy and face-mask immobilization, thereby rendering hypofractionated SRS feasible.

Linac — The principles of a Linac (linear accelerator) are identical to those of the Gamma Knife. Instead of using an array of cobalt sources, Linac SRS utilizes multiple non-coplanar arcs of radiation that intersect at the target volume. As a result, radiation received by normal tissue in each beam path is minimal relative to the point of beam convergence. Linac-based devices also achieve target accuracy between 0.1 and 1 mm [38]. The radiation volume is carefully matched to the lesion [3].

CyberKnife — The CyberKnife device combines a mobile linear accelerator with an image-guided robotic system (figure 5) [39]. The mobility of the device, combined with real-time imaging, obviates the need for an invasive stereotactic head frame.

ExacTrac Robotic system — The ExacTrac X-Ray 6D consists of two infrared (IR) cameras for patient tracking, two floor-mounted kilovolt X-ray tubes, and two ceiling-mounted detectors. X-ray images of the cranial skeletal anatomy are fused to the digital reconstructed radiographs derived from the treatment planning CT scan to facilitate patient positioning. IR fiducial markers attached to the patient allow precise tracking of patient's motion by the IR camera. This information is transmitted to an integrated computer system that corrects for any motion by adjusting the position of the treatment couch prior to the radiation delivery. Target accuracy of approximately 1 mm is achieved.

Prior to treatment, CT images are used to define the spatial relationship between the patient's bony anatomy and the target volume. During the actual treatment, patient movement is monitored with minimal time lag by the system's low dose X-ray cameras. These images are compared with radiographs derived from the pretreatment CT scan. Based upon these comparisons, the computer-controlled robotic arm adjusts the mobile linear accelerator in response to changes in patient position. A target accuracy of less than 1 mm is achieved.

Zap-X gyroscopic radiosurgery system — The ZAP-X is a radiosurgery platform where a mobile linear accelerator is mounted onto a gyroscopic mechanization that rotates around the patient to allow convergence of nonparallel radiation beams at the tumor target. The ZAP-X platform is unique relative to other radiosurgery platforms in that the platform harbors a built-in radiation shelter [40]. This built-in shelter eliminates the need for a radiation vault, which is required for all other radiosurgery platforms.

Tomotherapy — Tomotherapy (also known as helical tomotherapy) is a form of intensity-modified RT where the radiation is delivered by a linear accelerator that rotates around the patient. Comparisons between tomotherapy and Gamma Knife or dedicated Linac SRS platforms as treatment for brain metastasis have shown that the latter platforms are superior in terms of conformal delivery and sparing of radiation to the normal brain tissue [41]. In this context, tomotherapy is not routinely used as a delivery platform for SRS.

PROTON-BEAM SRS — A proton beam is generated by stripping a hydrogen atom of its electron and accelerating the residual proton in a magnetic field [42]. Charged protons have biologic properties that are different from a photon beam and offer advantages in selected situations. (See "Radiation therapy techniques in cancer treatment", section on 'Particle therapy'.)

The dose distribution of an unmodulated proton beam consists of an entrance region through which there is a slowly increasing dose, followed by a rapid rise to a maximum (the Bragg peak), and a fall to near zero (figure 3 and image 1). In contrast, the dose administered to tissue with a photon beam undergoes exponential decay with increasing tissue penetration.

The depth of the Bragg peak is a function of the energy of the proton beam and is modulated so that the depth of penetration corresponds to the depth of the target beneath the surface. Because there is no energy deposition beyond the Bragg peak, tissues beyond the target receive no radiation exposure [43]. The Bragg peak itself is only millimeters in depth. In order to cover the entire axial diameter of a tumor, the energy of the beam is precisely varied as it is delivered to cover the entire lesion from front to back. This is known as the spread-out Bragg peak (figure 3 and image 1).

Technique — In most cases, two to six beams are aimed at the target from different directions in sequence. This minimizes the dose of radiation received by normal tissue along the beam path from the surface to the intended target in each direction. Each of the beams is additionally shaped by brass collimators (apertures) that correspond to the cross-sectional silhouette of the tumor along the axis of that particular beam (figure 6A-B). Additionally each beam passes through a custom shaped bolus of absorbing material that modifies the energy of the protons across the face of the beam. This controls the energies of the individual protons so that the back edge of the Bragg peak conforms optimally to the three-dimensional shape of the target at its deep surface (figure 6A-B). All of these maneuvers together maximize the conformality of the radiation treatment volume in relation to the volume of the target being irradiated and absolutely minimize the dose of radiation to nearby normal tissues. This constitutes the physical advantage of proton radiation in contrast to photons (gamma or X-rays).

An automated device, known as Stereotactic Alignment for Radiosurgery (STAR), was developed to facilitate versatile and precise patient positioning (figure 7). The patient is placed in an immobilizing head frame attached to a couch apparatus that can be rotated relative to a fixed beam portal. Further refinements have utilized a mobile proton-beam source (picture 1).

Same-day treatments for proton SRS are not feasible because of the time required for preparation of the necessary equipment that must be fabricated for each patient and for quality assurance and control. To allow for this, we implant three small stainless steel beads into the outer surface of the skull through a needle. This 15 minute procedure using local anesthesia is done prior to a high resolution CT scan that identifies each of the markers in relation to the target to be treated. This information allows one to precisely localize the target at the time of proton-beam treatment by using conventional X-rays for patient alignment. The patient immobilization during proton SRS is achieved using a modified Gill-Thomas-Cosman frame (picture 2).

Availability — While there is growing availability of proton-beam systems for radiation therapy, proton-beam SRS requires dedicated expertise that is not commonly available. In contrast, photon-based SRS (Gamma Knife, Linac, CyberKnife) is widely available [44].

Because of limited availability, patient selection for proton SRS is restricted to carefully defined indications. In general, small (<10 cm³), spherically-shaped lesions do not require proton SRS if they are not located close to critical anatomic structures, in eloquent regions, in deep subcortical areas, or in previously irradiated volumes. In these cases, equally effective results can typically be attained with photon SRS. Patients with a limited expected survival are also unlikely to derive additional benefit from proton SRS, since delayed radiation toxicity from photon irradiation does not develop until years after treatment. In contrast, the advantage of protons with regard to conformality of the radiation dose (ie, minimizing radiation to normal tissues) becomes increasingly relevant with larger, irregularly shaped targets such as tumors and arteriovenous malformations, especially if they are intimately related to critical brain structures.

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.

●Basics topics (see "Patient education: Arteriovenous malformations in the brain (The Basics)")

SUMMARY

●Stereotactic radiosurgery (SRS) has emerged as an important option based upon advances in neuroimaging, medical physics, stereotactic neurosurgery, and radiation oncology. Its successful clinical application requires a team-based approach with active collaboration between specialists from each of these fields.

●Careful consideration must be given by the team members regarding the relative merits of SRS, conventional fractionated radiation therapy, surgery, and observation. In cases where SRS is indicated, additional consideration must be given to whether proton-beam SRS, if available, offers additional clinical benefits.

ACKNOWLEDGMENT — We are saddened by the death of Jay Loeffler, MD, who passed away in June 2023. UpToDate acknowledges Dr. Loeffler's past work as an author for this topic.