LARS LEKSELL BEGAN radiobiological investigations to study the effect of high-dose focused radiation on the central nervous system more than 5 decades ago. Although the effects of radiosurgery on the brain tumor microenvironment are still under investigation, radiosurgery has become a preferred management modality for many intracranial tumors and vascular malformations. The effects and the pathogenesis of biological effects after radiosurgery may be unique. The need for basic research concerning the radiobiological effects of high-dose, single-fraction, ionizing radiation on nervous system tissue is crucial. Information from those studies would be useful in devising strategies to avoid, prevent, or ameliorate damage to normal tissue without compromising treatment efficacy. The development of future applications of radiosurgery will depend on an increase in our understanding of the radiobiology of radiosurgery, which in turn will affect the efficacy of treatment. This article analyzes the current state of radiosurgery research with regard to the nature of central nervous system effects, the techniques developed to increase therapeutic efficacy, investigations into the use of radiosurgery for functional disorders, radiosurgery as a tool for investigations into basic central nervous system biology, and the additional areas that require further investigation.
Radiosurgery is a technique designed to deliver a high dose of focused radiation to a defined target volume to elicit a desired radiobiological response (43). This response depends on the radiation dose, the nature of the tissue, and the time elapsed after radiosurgery. The desired response for a vascular malformation is obliteration of the abnormal vessels, whereas for tumors, it is long-term growth control. For trigeminal neuralgia, the desired response is elimination of face pain; for movement disorders, it is control of abnormal movement. Radiosurgery is a radical departure from the general approach of clinical fractionated radiation therapy, which gradually evolved during the last century. The efficacy of large-field fractionated radiotherapy to treat brain tumors depends on biological differences between normal cells and tumor cells. Large-field radiotherapy is typically fractionated to exploit these differences and thus to limit the risk of normal tissue injury in patients with malignant brain tumors. In this way, fractionated radiotherapy can increase the therapeutic ratio, which is equivalent to the rate of tumor control divided by the rate of complications.
Radiosurgery, in contrast to fractionated radiation therapy, entails the use of a high dose of precisely focused radiation delivered in a single session. Its effect on the surrounding normal tissue is limited by the highly focused nature of radiosurgical beams. Unlike conventional radiotherapy, radiosurgery is used to treat small volumes of tissue with much higher single-fraction doses. Whereas fractionated radiotherapy is generally most effective in killing rapidly dividing cells, radiosurgery is thought to arrest the cell division capability of target cells irrespective of their mitotic activity, oxygenation, and inherent radiosensitivity. In consideration of this, the effects and the pathogenesis of those effects after radiosurgery may be unique. For this reason, it has been important to perform additional animal studies to understand the radiobiological effects of radiosurgery in both normal and abnormal nervous system tissues. Radiosurgery has evolved from a specific treatment for functional disorders, to treatment for brain tumors and vascular malformations, and now to a widely available treatment modality for a variety of intracranial, spinal, and other body targets (12, 13, 54, 61). Further understanding of the radiobiology of high-dose, single-fraction, ionizing radiation is crucial. This article analyzes the current state of radiosurgery research with regard to: 1) the nature of the effects of radiosurgery on the central nervous system (CNS); 2) techniques that have been developed for increasing therapeutic efficacy; 3) explorations into the use of radiosurgery for treatment of functional neurological disorders; 4) analysis of the use of radiosurgery as a tool for investigation into basic CNS biology; and 5) areas that require further investigation.
THE EFFECTS OF CNS RADIOSURGERY: ANIMAL MODELS
The histological response of CNS tissue to conventional radiation therapy has been well characterized and depends on both the radiation dose and the time elapsed after irradiation (10, 15, 62). Radiation response tends to be greater in white matter regions, although all regions of the brain may be affected (9, 65). Between the time of irradiation and the development of tissue injury, there is a dose-related variable latency period that can last from months to years (10, 15, 62). When lower single-fraction doses of 20 Gy are applied to rat brain (small volume), lesions primarily are confined to the vasculature when examined at intervals of more than 12 months (19, 50). Radiation-induced vascular changes in the CNS include perivascular fibrosis, fibrinoid necrosis of vessel walls, hyaline degeneration, edema, telangiectasia, thrombosis, and hemorrhage (55). At higher doses (e.g., 25 Gy), white matter lesions predominate in irradiated rat brain at latencies of 12 months or more. Radiation-induced lesions in the white matter can range from demyelination to complete myelomalacia (55). Primary vascular lesions occur at lower radiation doses after prolonged latency, whereas white matter lesions predominate after higher radiation doses and shorter latencies. These effects also pertain to the spinal cord (62). Vascular damage becomes prevalent once again as the dose is further increased beyond that causing white matter destruction (62).
Results of animal studies involving radiosurgery of normal CNS tissue suggest that the effects of radiosurgery are similar to those of conventional radiation therapy (Table 1) (2, 5, 24, 25, 28, 33, 34, 36, 42, 51). In one of the initial studies, Larsson et al. (33) investigated the use of the high-energy proton beam (185 MeV proton beam from 230-cm synchrocyclotron) as a neurosurgical tool. The early histological results (3rd–8th day) demonstrated complete transection of the rabbit spinal cord by use of 40,000 rad (400 Gy) with a 1.5-mm beam diameter and 20,000 rad (200 Gy) with a 10-mm beam diameter. Using 20,000 rad of stereotactic multiple port proton beam radiation, these investigators documented sharply defined lesions in deep parts of goat brain within 4 to 7 weeks. Rexed et al. (51) studied the long-term effects (2–56 wk) of proton beam irradiation on rabbit brain. With a 1.5 mm collimator, 20,000 rad (200 Gy) was delivered to the anterior part of rabbit brain. Serial histological analysis revealed a well-demarcated lesion in the beam path for up to 3 months. After 3 months, a lesion broader than the beam size was noted. Leksell et al. (34) investigated the features of a radiofrequency lesion in the depth of brain produced by cross-fired irradiation with a narrow, high-energy beam. Their results demonstrated that with 20,000 rad (200 Gy central dose), well-circumscribed intracerebral lesions of appropriate size and shape could be created. Andersson et al. (2) studied the late histological effects of the cross-fired beams of 185 MeV protons on goat brain. No late untoward changes in or around the lesion (e.g., elements resembling neoplasm, hemorrhage, or telangiectasis) were observed 1.5 to 4 years after 20,000 rad (200 Gy) radiosurgery.
Nilsson et al. (42) irradiated (100–300 Gy) the basilar artery of cats by use of stereotactic technique with a 179-source cobalt-60 prototype gamma knife unit. Histological analysis demonstrated vascular lesions such as vacuolization, degeneration, and desquamation of the endothelium and hyalinization and necrosis of the muscular coat. The reparatory reactions were relatively sparse, and thrombosis was completely absent.
Lunsford et al. (36) developed a baboon model to determine the in vivo radiobiological effects of stereotactic radiosurgery with a 201-source cobalt-60 gamma knife. By use of an 8-mm collimator, a dose of 150 Gy (100% isodose) was delivered to the caudate, thalamus, or pons with the gamma knife. There were no changes visible on computed tomographic scans or T1- or T2-weighted, gadolinium-enhanced magnetic resonance imaging (MRI) scans 4 weeks after irradiation. A circumscribed, contrast-enhanced lesion was visible by 6 to 8 weeks, and edema was first evident at 8 weeks (36). The presence of a latency period between irradiation and the onset of tissue damage is consistent with that which occurs after “standard” brain irradiation. When high doses (and high dose fractions) are used in radiosurgery, these changes occur more quickly. Histopathological changes after radiosurgery also were similar to those observed after more typical radiation treatments. At 6 weeks after irradiation, there was a circumscribed lesion at the focus of irradiation, which was characterized by demyelination, microvascular damage and hemorrhage, and astrocytosis. The irradiated region had undergone frank necrosis by 24 weeks.
Subsequent studies of the effects of radiation on normal brain have enhanced our understanding of the radiosurgical dose response. Kondziolka et al. determined the effect of various doses at a fixed time point, 90 days, after irradiation of the rat brain. The frontal lobe was irradiated with maximal doses of 30 to 200 Gy delivered with a 4-mm collimator (28). Doses of 70 Gy or higher were required to induce detectable histological changes, and necrosis was present only in tissues treated with 100 Gy or more. In contrast to the previously cited study, Blatt et al. (5) serially evaluated tissue changes after a constant dose of radiation. The anterior limb of the internal capsule of cats was treated with a linear accelerator (10-mm collimator) at a dose of 125 Gy prescribed to the 84% isodose line. Animals were assessed by use of MRI and histopathological analysis for 1 year starting 3.5 weeks after irradiation. Tissue necrosis was noted in the cat brain by 3.5 weeks and was accompanied by vascular proliferation and edema. Unlike the previously cited baboon study, wherein the focal lesion in one baboon became progressively larger from 8 to 24 weeks after irradiation, the lesions in the cat study were accompanied by increased vascularity, microglial infiltration, and resorption by 12 to 29 weeks. Because of differences in the radiation protocol, dose, species, and time point of evaluation, it is difficult to determine the exact cause of the disparate results. The potential for resolution is important because progressive radionecrosis generally would have severe clinical consequences.
In a study evaluating the effects of radiation dose and time after treatment on the radiosensitivity of brain, rats were irradiated with maximum doses of 50, 75, or 120 Gy and analyzed for histological changes and blood-brain barrier integrity up to 12 months later (23). Whereas 120 Gy induced alterations in astrocytic morphology by just 3 days after treatment, such changes were not observed until 3 months after a 50-Gy dose. Blood-brain barrier breakdown as assessed by Evans blue leakage was evident within 3 weeks after 120-Gy irradiation but was not detected up to 12 months after a dose of 50 Gy. These findings indicate that the latent period between irradiation and detection of pathological alterations depends on both the dose and the biological end point used. Such findings are consistent with the results of studies of a more conventional radiation source (cobalt-60) to irradiate the rat spinal cord. In this model of radiation-induced CNS injury, latency to paralysis after irradiation of an 8- or 16-mm segment of cervical spinal cord decreased as dose increased (20). In addition, after 4 mm of spinal cord irradiation, the effective dose to produce a response in 50% of the subjects for paralysis was 51 Gy, whereas it was only 25.6 Gy for vascular damage.
The impact of dose and biological end points on latency also was reported by Karger et al. (25), who evaluated the rat brain with T1- and T2-weighted, gadolinium-enhanced MRI at 15, 17, or 20 months after treatment with 26 to 50 Gy. A linear accelerator was used to deliver the radiosurgical dose by use of a convergent arc technique; a 3-mm collimator resulted in an 80% isodose distribution of 4.7 mm in diameter. There was no evidence of damage at any time point for doses less than 30 Gy, but the dose response of brain was dependent on the time of evaluation. After a 40-Gy dose, the latency to detectable MRI changes was approximately 19 to 20 weeks, whereas the latency after a 50-Gy dose was 15 to 16 weeks. In addition, T1-weighted changes in the MRI signal had a shorter latency than T2-weighted changes. As changes in T1-weighted images are attributed to leakage of gadolinium-diethylenetriamine penta-acetic acid across the blood-brain barrier, the results of this study point to the probable role of vascular damage in radiation-induced injury.
The importance of the vasculature in radiation-induced brain injury is well recognized; a prevalent hypothesis regarding the pathogenesis after conventional radiotherapy is that damage to capillary endothelium and/or supporting cells ultimately interrupts blood flow resulting in secondary ischemic necrosis. In a report focusing on vascular changes after a maximal dose of 75 Gy delivered to the rat brain with a gamma knife, it was observed that vascular changes, specifically alterations in the basement membrane, preceded changes in necrosis (24). This finding suggests that vascular damage is also an important component in radiation injury after radiosurgery.
Although radiosurgery involves the use of higher single doses and smaller treatment volumes than conventional irradiation, the histological effects of these two methodologies seem similar. The biggest differences are that the latency period after radiosurgery is shorter and the major histological finding is vascular damage. However, studies of more typical irradiation protocols have demonstrated that latency decreases with as dose increases, and that vascular injury is a prominent result of very high radiation doses, so the differences are not unexpected.
STRATEGIES TO ENHANCE THERAPEUTIC EFFICACY
Although radiosurgery is a valuable boost technique (in addition to fractionated radiation therapy) for malignant glial tumors, which are commonly regarded as fast-reacting targets surrounded by late-reacting normal tissue, tumor almost invariably recurs or progresses. Additional strategies are needed to improve cell kill of malignant brain tumors and to protect normal surrounding brain (Table 2) (29, 41, 44, 45). A few strategies for radioprotection of normal tissue and radiosensitization of tumor tissue already have been explored.
Radiation Protection and Repair
Initial strategies included use of various agents to protect normal brain while delivering a high dose to tumor cells. Identification of effective radioprotective agents has been problematic. Oldfield et al. (46) demonstrated that pentobarbital conveyed protection from radiation-induced brain injury. Our studies evaluated the potential radioprotective effect of pentobarbital by intraperitoneal administration of 50 mg/kg of pentobarbital 30 minutes before brain gamma knife radiosurgery in rats (35 Gy to 50% isodose line). This relatively high dose of pentobarbital did not delay or prevent cellular or parenchymal histological changes compared with controls (31).
The 21-aminosteroids (21-AS), also known as lazaroids, have been advocated as a cerebral protective agent in patients with head trauma or subarachnoid hemorrhage (56) and as a potential radioprotective agent. Radiation injury is related to the production of oxygen free radicals, which induce deoxyribonucleic acid strand breaks, initiate lipid peroxidation of vascular membranes, and ultimately lead to membrane lysis and cell death. The basis of the protective effects of 21-AS is probably a result of their antioxidant properties (6). As a lipid antioxidants and free radical scavengers, 21-AS inhibit oxygen radical initiated peroxidation of vascular membrane. They also block the release of free arachidonic acid from cell membranes, thereby inhibiting activation of the pro-inflammatory cyclo-oxygenase pathway. These properties of 21-AS are thought to protect cerebral vessels from injury and prevent cerebral edema.
The effects of 21-AS (U-74389G) have been evaluated in both rat and cat radiation injury models. Bernstein et al. (4) reported that U-74389G reduced brachytherapy-induced brain injury in the rat. Buatti et al. (8) determined that the same agent also protects the cat brain from injury attributable to radiosurgery and was significantly more effective than corticosteroids. In our studies, 15 mg/kg, but not 5 mg/kg, of U-74389G was effective at reducing injury in the rat when administered 1 hour before radiosurgery. U-74389G ameliorated vasculopathy and regional edema and delayed the onset of necrosis, and gliosis was unaffected (32). We then evaluated the effects of 21-AS in the rat C6 glioma model after stereotactic radiosurgery and documented that 21-AS exhibited a radioprotective effect on normal brain tissue but not on tumor (30). Preliminary results suggest that this agent may be acting through reduction of the cytokines induced by brain irradiation.
A second potential strategy for radiation protection aims at use of cells for repair of radiation-induced brain damage. The cellular target primarily responsible for radiation-induced breakdown of normal tissue is unclear. The white matter and the cerebral vasculature seem to be particularly susceptible to radiation. Oligodendrocytes and endothelial cells may be critical targets of radiation. Recent studies also have implicated a potential role for neural progenitors in radiation-induced brain injury (59, 60). We hypothesize that radiation-induced damage to these cell types can be repaired by engrafted neural stem cells. Neural stem and progenitor cells can be isolated from normal adult mammalian brain and passaged long term in culture (16). These cells can be induced to differentiate into neurons or glia (Fig. 1). At our institution, experiments are underway to study the mechanisms of radiation-induced brain damage and the role of neural and endothelial precursors in repairing this damage. If implanted neural stem cells could prevent or repair radiation-induced damage to normal brain, the tumor could be targeted with higher radiosurgery doses. In initial experiments, rodent neural progenitors that had been isolated from adult brain were implanted into normal brain and tumor-bearing brain. In normal brain, the cells were typically located either in perivascular or white matter regions less than 1 mm from the injection site at 24 hours after injection (Fig. 2). Cells that were injected intratumorally were incorporated into the tumor and occasionally within adjacent white matter tracts (data not shown). Additional work will be required to determine whether the transplanted neural progenitors can replace those lost because of irradiation.
We have explored the use of tumor necrosis factor α (TNFα) for enhancing the effect of radiosurgery. TNFα can act as a tumoricidal agent with direct cytotoxicity mediated through binding to its cognate cell-surface receptors and a variety of activities triggering a multifaceted immune attack on tumors (11, 17, 18, 47, 48, 57). Furthermore, it has been reported that locally produced TNFα can enhance the sensitivity of tumors to radiation in nude mice (57). Our study used a replication defective herpes simplex virus (HSV), which was used as a vector to deliver thymidine kinase and/or TNFα genes to the tumors in a U-87 MG nude mouse model. Gene transfer was followed by ganciclovir therapy and gamma knife radiosurgery (15 Gy to the tumor margin and 21.4 Gy to the center). The combination of radiosurgery and HSV-thymidine kinase suicide gene therapy (SGT) with TNFα improved median animal survival; 7 of 9 animals survived beyond 75 days (44). In additional experiments, the connexin 43 gene was added to enhance the formation of gap junctions among tumor cells, which should facilitate the intercellular spread of thymidine kinase-activated ganciclovir therapy from virus-infected cells to noninfected surrounding cells. This creates a bystander effect that can improve tumor cell kill (38). We further tested this strategy in a 9L rat glioma model. Groups of animals were treated with one modality or a combination (SGT, connexin-enhanced SGT, TNFα-enhanced SGT, and radiosurgery). In this experiment, addition of radiosurgery to SGT, SGT combined with connexin 43, or SGT combined with TNFα demonstrated significant improvements. Radiosurgery was a beneficial addition to each of the gene therapy protocols tested. We then combined radiosurgery with TNFα and connexin 43-enhanced SGT and tested this treatment modality in 9L glioma model. This treatment significantly prolonged animal survival (150 d) with 75% animals surviving (45).
The preceding results demonstrate that gene therapy is an effective strategy for enhancing radiosurgery. The exact mechanism of this effect is unclear, however, and must be a topic of future investigations. In other studies, tumor sensitization to radiation apparently was mediated by extracellular TNFα promoting the destruction of tumor vessels, whereas HSV vector-mediated TNFα-enhanced killing of malignant glioma cell cultures is presumably a consequence of an intracellular TNFα activity (17, 40). We have proposed a clinical trial of HSV vector-based multigene therapy and radiosurgery for treatment of recurrent glioblastoma multiforme (Fig. 3).
Radiosurgery is rapidly expanding beyond its use as a treatment of brain tumors and AV malformations. It has been determined to be effective for other neurological disorders such as epilepsy, movement disorder, and trigeminal neuralgia. The promise of “functional” radiosurgery has led to a need for research into its efficacy, limitations, and potential drawbacks (Table 3) (7, 14, 21, 26, 27, 35, 37, 39).
The potential efficacy of radiosurgery for the treatment of epilepsy has been evaluated in rat models. Kainic acid induces epilepsy in reproducible fashion in the rat when injected into the hippocampus. By use of such a model, Mori et al. (39) treated epileptic rats with doses of 20 to 100 Gy delivered with a gamma knife and a 4-mm collimator. The efficacy of the treatment for epilepsy was determined by direct observation and scalp electroencephalography for 42 days. Even 20 Gy significantly reduced the number of seizures, and the efficacy improved with increasing doses. Only doses of 60 Gy or higher induced histological changes. In a similar study, Maesawa et al. (37) treated rats that had kainic acid-induced epilepsy with a single dose of 30 or 60 Gy. Both doses significantly reduced electroencephalography-defined seizures, and this effect occurred somewhat sooner after the higher dose (5–9 wk for 60 Gy versus 7–9 Gy for 30 Gy). In addition, the investigators evaluated learning and memory performance of the rats in the Morris water maze task. Although kainic acid injection alone reduced performance in the water maze task, the performance of rats that treated with radiosurgery after kainic acid administration was not different from that of controls.
To identify potential normal tissue complications and determine dose limits for hippocampal radiosurgery, Liscak et al. (35) evaluated the effects of radiosurgery on normal hippocampus. Four separate 4-mm isocenters were used to irradiate the entire hippocampus with 25 to 100 Gy. Doses lower than 50 Gy caused no perceptible changes in results of histological analysis, MRI, and Morris water maze testing. In contrast, Morris water maze performance of rats that received doses higher than 50 Gy was significantly worse than that of controls. In a recent study, Brisman et al. (7) analyzed histological, electrophysiological, and behavioral effects of proton beam radiosurgery in the rat hippocampus. These investigators observed necrosis after 3 months with 90 cobalt Gray-equivalents or higher doses. Necrotic doses were associated with imaging abnormalities, altered physiology, and adverse behavioral effects. Taken as a whole, the study results described above support the concept that radiosurgery may be an effective method for treating epilepsy, but they also suggest that doses to the hippocampus should be limited to reduce potential effects on learning and memory.
The radiation sensitivity of potential targets for the treatment of movement disorders has been evaluated in a limited number of primates. De Salles et al. (14) used a linear accelerator and 3-mm collimator to deliver a maximal dose of 150 Gy to the subthalamic nucleus of one vervet monkey and to the substantia nigra of another. A 3-mm lesion was formed as revealed by MRI, and the lesion did not increase in size during the course of the study. Kondziolka et al. (26) analyzed the radiosensitivity of the thalamus in the baboon brain, and they reported that a dose of 100 Gy (central dose with a 4-mm collimator) was sufficient to induce contrast enhancement of MRI scans and coagulative necrosis as demonstrated by histological analysis.
Radiosurgery has significant potential as an effective, non-invasive method of treatment for trigeminal neuralgia, and the effect of gamma knife irradiation on the trigeminal nerve has been evaluated in the baboon (27). The proximal trigeminal nerve was irradiated with 80 or 100 Gy via a 4-mm collimator. A 4-mm region of contrast enhancement was visible on MRI scans 6 months after treatment. Both large and small fibers were affected with axonal degeneration occurring after 80 Gy and necrosis after 100 Gy. Neither dose was effective at selectively damaging fibers responsible for transmission of pain while maintaining those responsible for other sensations, which would be optimal for effective treatment of trigeminal neuralgia. Nevertheless, this study demonstrated that it is possible to affect specific nerves noninvasively and precisely with radiosurgery. Whether other dose regimens might cause selective damage to pain fibers requires further investigation.
RADIOSURGERY AS A RESEARCH TOOL
Radiosurgery can serve as a valuable research tool for many studies because it can precisely and noninvasively target selected brain regions. To date, there only few studies have exploited this capability. Recent studies in our laboratory have focused on the role of the subependymal zone (SEZ) in brain repair. The region of the SEZ on the lateral side of the lateral ventricle is thought to be the location of a high concentration of adult neural stem cells. These cells can divide to produce astrocytes, oligodendrocytes, and neurons (1, 3, 22, 49, 52, 53, 58, 63, 64), and it has been speculated that cells of this region may be important for replacing cells lost through damage or normal attrition. The SEZ is highly sensitive to radiation, and irradiation leads to long-term depletion of mitotically active cells within the region (60). In our studies, rats are being irradiated with a dose of 40 Gy at the 90% isodose line with a 4-mm collimator. The dose is centered 3 mm lateral to bregma. Histological analysis 1 week after treatment invariably reveals that the SEZ within the irradiated hemisphere is dramatically altered. When compared with the contralateral hemisphere, the SEZ within the targeted area has reduced cellularity and thickness, and there is dramatically reduced labeling with bromodeoxyuridine, an agent used to label mitotically active cells (Fig. 4). The effect of such cellular loss on brain function has not been established yet and is currently under intense investigation. Nonetheless, these results demonstrate the ability of radiosurgery to target brain regions precisely, even in an animal as small as the rat.
There have been rapid developments in neuroimaging, stereotactic techniques, and robotic technology during the last decade. These advancements have led to improved results and wider applications of radiosurgery. Radiosurgery is an optimal modality for managing many intracranial tumors such as schwannomas, meningiomas, and metastatic tumors, especially in the absence of local mass effect. Although there are benefits to the use of radiosurgery in the treatment of diffuse malignant brain tumors, cure still is not possible. Such tumors almost always recur because of microscopic infiltration into the surrounding functional brain. For this reason, strategies are needed to target tumor cells specifically while preserving normal tissue. One possible approach is to use gene transfer to sensitize malignant tumor cells to radiosurgery. Similarly, radioprotective agents could prevent damage to surrounding normal tissue.
Although the nature of brain injury after radiosurgery seems similar to that after conventional radiation treatments, the pathogenesis of such effects after both forms of radiotherapy still is not entirely clear. Further research to establish the pathogenesis is needed to maximize the effectiveness of treatment on target regions and minimize injury to other areas. One observation that requires further investigation is that sensitivity of the CNS to radiation damage increases as treatment volume increases. This property, commonly referred to as the volume effect, significantly limits the radiation dose used in treating large intracranial tumors. This effect may be a result of the production of various injury cytokines after irradiation of a critical volume of tissue. Alternatively, larger treatment volumes may limit the ability of cells to migrate into damaged areas and to effect repair. Another unanswered question concerns the primary target of vascular injury and the subsequent pathogenesis of radionecrosis. Both endothelial and smooth muscle cells could be primary facilitators of vascular injury. Although loss of blood flow could have severe consequences for tissue, the effect of breakdown of the blood-brain barrier and leakage of blood components after irradiation is unknown.
A defining characteristic of radiosurgery is that there is a rapid falloff in radiation so that the center of the field receives a dose much higher than that required for necrosis. A significant question is whether the intensity of the dose at the isocenter might increase the radiosensitivity of surrounding tissue. Other important questions include whether various regions of the CNS differ in their radiation sensitivity and whether damage to particular areas, such as the SEZ, has a greater impact on the sensitivity of surrounding tissue. If particular regions of brain are less sensitive, then lesions in those regions might be treated more aggressively. Although radiation necrosis probably is caused by killing cells that are capable of mitosis, such as endothelial cells and oligodendrocytes, the effect of radiation on postmitotic cells, such as neurons, is not well established. Although neurons may survive treatment, their functional capabilities may be impaired. Increasing our understanding of the factors that effect tissue sensitivity will enhance the development of future radiosurgical applications.
LDL and DK are consultants for Elekta Instruments AB.
No financial support was received in conjunction with the generation of this submission.
The authors of this article state the case for “basic research” on the radiobiology of high-dose single fractions of radiation on the brain for the purpose of improving the outcome of radiosurgical treatments in patients. What this article reviews cannot really be called basic research. This is a review of studies from various centers in which the central nervous system, primarily of rodents, is irradiated and then pathological and functional studies are performed. Because rodent brain is relatively radioresistant, for any effect to be seen, the studies were performed primarily at supramaximal doses. Paralysis, discrete lesions, white matter necrosis, vascular lesions, hemorrhage, and thrombosis were documented in the normal brain of various models, as expected. No mechanistic insights are forthcoming. The authors further review studies showing regression of tumors, the amelioration of epilepsy, and trigeminal lesions in appropriate models.
Most of the experiments reviewed in this article seem to be set up to demonstrate what is already known rather than to test hypotheses. Because of this, it is difficult to point to any particular advances in treatment that have arisen from or will arise from the experimental studies of radiosurgery performed to date.
Philip H. Gutin
New York, New York
Niranjan et al. have reviewed the experimental radiosurgical literature. They contend that further major advances in radiosurgical treatments will more likely require significant improvements in our understanding of the basic radiobiology involved rather than evolutions in hardware or computer technology. Our hopes for the future of this field will indeed involve marrying a much deeper understanding of the basic molecular events involved and our current technology. What genes are up-regulated or down-regulated in glial and vascular cells after radiosurgery? How can we manipulate this response on a molecular level to achieve a breakthrough in either brain protection or tumor kill?
William A. Friedman
Understanding the radiobiological underpinnings of radio-surgery is vitally important to future progress in this field. In this report, Niranjan et al. have reviewed the current state of knowledge, citing the most important animal studies, many of which were published by their own group. Although the radiobiological concepts of radiosensitivity, effects of oxygenation, and dose-volume effects are reasonably well characterized for conventional fractionated radiation therapy, it is understood to a much lesser degree for radiosurgery. The authors appropriately and comprehensively discuss the major differences between the radiobiological effects seen with these two approaches to radiation delivery, e.g., the latency of radiation injury and the importance of vascular effects.
Whether delivered as a single large dose or with a hypofractionated schedule, the ever-increasing application of radiosurgery to lesions everywhere in the central nervous system necessitates further study into the underlying basic radiobiology. Eventually, these investigations may help to challenge the widely believed dogmas surrounding the tolerance to radiation dose of certain critical structures as well as what we believe about the essential radiosensitivity of various pathological conditions. For example, our group has recently published our results using hypofractionated CyberKnife radiosurgery to treat perioptic lesions (1); this clinical experience clearly demonstrated that visual preservation can be achieved in almost all cases even after delivery of doses to the optic apparatus that exceed the “established” dose tolerance. Although our approach in this recent report was largely retrospective, we agree that the most compelling radiobiological evidence comes from well-organized animal and human studies.
Iris C. Gibbs
John R. Adler, Jr.
This review provides a comprehensive assessment of the radiobiology of radiosurgery (high-dose, single-fraction radiation). Not surprisingly, the work of Drs. Lunsford, Kondziolka, and colleagues from the University of Pittsburgh are frequently referenced, because this center has been instrumental to our understanding of this topic. At this point, it is clear that our knowledge of this subject lags behind our clinical application of this technology. Still, I completely agree with the authors’ contention that continued research will someday provide us with a greater ability to enhance the therapeutic applications of radiation. This is a timely piece, and I look forward to the authors’ future work on this critical aspect of radiosurgery.
Bruce E. Pollock