Radiosurgery and fractionated radiation therapy: comparison of different techniques in an in vivo rat glioma model

Transcription

1 J Neurosurg 84: , 1996 Radiosurgery and fractionated radiation therapy: comparison of different techniques in an in vivo rat glioma model DOUGLAS KONDZIOLKA, M.D., M.SC., F.R.C.S.(C), SALVADOR SOMAZA, M.D., CHRISTOPHER COMEY, M.D., L. DADE LUNSFORD, M.D., DIANA CLAASSEN, M.D., SUDHA PANDALAI, B.SC., ANN MAITZ, M.SC., AND JOHN C. FLICKINGER, M.D. Departments of Neurological Surgery, Radiation Oncology, and Pathology, University of Pittsburgh; and the Center for Image-Guided Neurosurgery, Presbyterian University Hospital, Pittsburgh, Pennsylvania To identify histological changes and effects on survival in rats harboring C6 gliomas, the authors compared radiosurgery to different fractionated radiation therapy regimens including doses of calculated biological equivalence. Rats were randomized to control (54 animals) or treatment groups after implantation of C6 glioma cells into the right frontal brain region. At 14 days, treated rats underwent stereotactic radiosurgery (35 Gy to tumor margin; 22 animals), wholebrain radiation therapy (WBRT) (20 Gy in five fractions; 18 animals), radiosurgery plus WBRT (13 animals), hemibrain radiation therapy (85 Gy in 10 fractions; 16 animals) or single-fraction hemibrain irradiation (35 Gy; 10 animals). When compared to the control group (median survival 22 days), prolonged survival was identified after radiosurgery (p ), radiosurgery plus WBRT (p ), WBRT alone (p = ), hemibrain radiation therapy to 85 Gy (p ), and 35-Gy hemibrain single-fraction irradiation (p = 0.004). Compared to the control group (mean tumor diameter, 6.8 mm), the tumor size was reduced in all treatment groups except WBRT alone. Reduced tumor cell density was exhibited in rats that underwent radiosurgery (p = 0.006) and radiosurgery plus WBRT (p = 0.009) when compared with rats in the control group, a finding not observed after any fractionated regimen. Increased intratumoral edema was identified after radiosurgery (p = 0.03) and combined treatment (p = 0.05), but not after fractionated radiation therapy or 35-Gy single-fraction hemibrain irradiation. In this animal model, the addition of radiosurgery significantly increased tumor cytotoxicity, potentially at the expense of radiation effects to regional brain. We found no difference in survival benefit or tumor diameter in animals that underwent radiosurgery compared to the calculated biologically equivalent regimen of 10-fraction radiation therapy to 85 Gy. The histological responses after radiosurgery were generally greater than those achieved with biologically equivalent doses of fractionated radiation therapy. KEY WORDS radiosurgery brain neoplasm glioma radiation therapy T HE increased use of stereotactic radiosurgery or stereotactic fractionated irradiation as an addition or alternative to conventional therapy for malignant brain tumors mandates investigation into the relative effects of these approaches. Use of the C6 rat glioma model is a valuable technique to evaluate the response of different radiation modalities on tumor biology, to study chemotherapeutic agents, and to analyze tumor growth patterns. We previously tested the effects of stereotactic radiosurgery alone using a frontal lobe C6 rat glioma model. 7 In that initial study we examined animal survival after radiosurgery and studied the histological effects of various radiosurgical doses. Although we detected no statistical increase in animal survival after radiosurgery, treated tumors significantly regressed and had reduced cellular density. In the present study, we sought to evaluate the clinical and histological response of a growing J. Neurosurg. / Volume 84 / June, 1996 malignant glioma after either high-dose stereotactic radiosurgery (margin dose 35 Gy), fractionated hemibrain radiation therapy to 85 Gy in 10 fractions (calculated to be biologically equivalent to radiosurgery), a single-fraction hemibrain regimen of 35 Gy, lower-dose fractionated whole-brain radiation therapy (WBRT) (20 Gy in five fractions), radiosurgery plus WBRT, or an untreated control group. We hypothesized that radiosurgery alone or in combination with WBRT would increase animal survival in comparison to no treatment or WBRT alone and that a histological correlate could be defined with the survival response. The treatment arms were chosen to compare the effects of radiosurgery and doses of fractionated radiation therapy of approximate biological equivalence for malignant tumors and to study the effects of focused (radiosurgery) versus large-field irradiation at the same 35 Gy margin dose. 1033

2 D. Kondziolka, et al. Materials and Methods All experiments were performed with the approval of the Animal Care and Use Committee of the University of Pittsburgh. Male Sprague Dawley tumor implanted rats (average weight, 300 g) were randomized to a variety of groups: control (no irradiation), stereotactic radiosurgery (35 Gy to 50% isodose; maximum dose of 70 Gy), radiosurgery (35 Gy) plus WBRT to 20 Gy in five fractions, WBRT to 20 Gy in five fractions, large-field (hemibrain) radiotherapy to 85 Gy in 10 fractions, and single-fraction nonstereotactic hemibrain irradiation to 35 Gy. Tumors were grown as previously reported. 7 We used the C6 glioma 2 in this and prior experiments because of its widespread use and utility, 14 its consistency in tumor histology, and because it produced a fairly well-circumscribed tumor suitable for focused radiosurgical targeting. 7 We implanted C6 glioma cells into the rat frontal brain region at passage number 40 to 42. Cells were pelleted by centrifugation and resuspended to a final concentration of 10 6 cells/10 l. For implantation, the anesthetized rat was placed in a stereotactic head frame (David Kopf Instruments, Tujunga, CA) and a small right frontal craniectomy was drilled 2 mm from the midline and 1 to 2 mm anterior to the coronal suture. Glioma cells were implanted stereotactically to a depth of 4 mm below the craniectomy. A 1-mm segment of a 25- gauge needle was inserted into the craniectomy site for later stereotactic targeting. The craniectomy was sealed with bone wax and the scalp closed. Any unused cells were recultured to determine postimplantation survival and, in each case, normal growth in culture was exhibited. In all groups, rats received intramuscular dexamethasone 0.25 mg once daily from Day 20 to Day 45 postimplantation; this regimen was constant for all animals. We administered this daily course of corticosteroids to prevent early animal demise from tumor-related brain swelling. A total of 189 rats were implanted with tumor and randomly placed into treatment or control groups. Because the implantation success rate was 68% in our initial study, we censored any animal that had no evidence of tumor at autopsy. In this study, as in our prior work, the percentage of animals with no tumor did not differ between control and treatment groups, even with doses up to 100 Gy. We therefore rejected the concept that treatment would eradicate all histological evidence of tumor. After autopsy, we found that 133 rats (70%) had successfully grafted tumors. These animals either served as controls (54 animals), or had radiosurgery (22 animals), radiosurgery plus WBRT (13 animals), WBRT alone (18 animals), hemibrain 10-fraction radiation therapy (16 animals), or hemibrain single-fraction irradiation (10 animals). Radiosurgical Technique For radiosurgery, rats were placed onto our modified small-animal stereotactic frame and underwent lateral and anteroposterior plain radiographic targeting. We selected a brain target 4 mm perpendicular to an extradural needle fragment in the right frontal brain region, which corresponded to the center of cell injection. The 4- mm collimator of the 201-source cobalt 60 gamma unit (Elekta Instruments, Atlanta, GA) was used for radiosurgery. The 50% isodose line (35 Gy) matched the approximate tumor size at 14 days postimplantation, 7 the time at which radiosurgery was performed. All animals that underwent radiosurgery received a maximum dose of 70 Gy. Prior work has shown us that this dose does not cause necrosis in the normal rat brain within 90 days. 6 Although we have evaluated the response to 100 Gy in animals observed up to 1 year after radiosurgery, we have not studied the longer term effects of a 70 Gy maximum dose in the rat. 8 Radiation Therapy Technique Fractionated WBRT was performed on Days 14 to 18 postimplantation. Because the animals that underwent radiosurgery were treated on Day 14, those that received additional radiotherapy did so 4 to 6 hours later the same day. Whole-brain radiation therapy was administered to a dose of 20 Gy in five fractions prescribed to the midplane of the brain. Both right and left lateral fields were treated each day using a 6-mV linear accelerator beam with 1 cm of tissueequivalent bolus material. TABLE 1 Length of survival after application of different radiation techniques for controlling C6 glioma in rats* Median 90-Day p Value Survival 95% CI Survival (no. vs. Group (days) (days) of animals) Control control of 54 radiosurgery (35 Gy) of 22 < radiosurgery (35 Gy) of 13 < WBRT (20 Gy in five fractions) WBRT (20 Gy in five of fractions) fractionated hemibrain RT of 16 < (85 Gy in 10 fractions) single fraction hemibrain of RT (35 Gy) * Abbreviations: CI = confidence interval; RT = radiation therapy; WBRT = whole-brain radiation therapy. Single- or 10-fraction hemibrain irradiation was initiated on Day 14 using a mm anteroposterior field to administer either 35 Gy in a single-fraction dose or 85 Gy in 10 daily fractions over 11 elapsed days. The dose of 85 Gy was chosen because theoretically it has a biologically equivalent effect for early responding tissues as calculated with the linear quadratic model assuming an alpha/beta ratio of 10 and no time factor. 9 Posttreatment Assessments After tumor implantation, the rats were examined daily to monitor their external appearance, feeding behavior, and mobility (ability to walk 50 cm in 10 seconds). 7 We did not use magnetic resonance imaging to follow the tumor response, although this would have provided a better assessment of temporal changes in the tumor (the meaning of such changes might not be clear); we chose instead to rely on the histological appearance of the tumor. The contralateral (left) forelimb was observed daily for development of paresis both passively and actively. The rat was killed when it could not walk the 50-cm length or when it had survived the full 90-day observation period. For whole or hemibrain irradiation, we similarly used this assessment of mobility to determine neurological demise. We increased our prior observation period of 65 days 7 to 90 days to evaluate potentially longer survivals with the different treatment regimens. Tissue Preparation The anesthetized animal was exsanguinated and perfused with 10% formalin. The brain was removed and fixed in formalin. Coronal sections were made at the level of the implanted extradural needle fragment to mark the center of the tumor. Serial sections of cerebral hemisphere and midbrain were examined in each animal. The section with the maximum tumor diameter was used to measure the tumor in two planes and a mean diameter was calculated by the neuropathologist. Because these tumors typically grew in a spherical fashion, we believe that mean diameter was a reliable estimate of tumor volume. Tissue blocks were embedded in paraffin, sectioned at 7 m and stained with hematoxylin and eosin. Two observers (one neurosurgeon, and one neuropathologist blinded to the study groups) examined each specimen for tumor size, tumor cell density, and edema. Tumor edema was considered present when more than a 50% increase in tumor intercellular spaces was observed. Tumors were defined as hypocellular (low cell density) if more than 50% of the tumor area (excluding necrosis) appeared pale because of increased separation between adjacent viable cell nuclei, usually due to individual cell death or edema. Intratumoral hemorrhage was defined as present if at least 10% of the tumor volume was composed of blood clot. We did not use a grading scale because 1034 J. Neurosurg. / Volume 84 / June, 1996

3 Radiation strategies tested in rat glioma FIG. 1. Graph displaying Kaplan Meier survival curves for rats implanted with C6 glioma that underwent radiosurgery using 35 Gy to the tumor margin (R), 10-fraction hemibrain radiation therapy to 85 Gy (X), or no treatment (C) control. The two treatment arms were calculated to be of biological equivalence. of the subjective nature of these findings. Rather we coded a characteristic as being present or absent according to the descriptions noted above. Our evaluations of these parameters proved reproducible when specimens were reexamined (no finding was changed in any animal on microscopic reassessment and all histological preparations were examined twice). Statistical Methods Survival data were evaluated using both a log-rank test and a two-tailed t-test for comparing independent mean values. To compare histological features we used Yates corrected chi-square test for column tabled data, or Fisher s exact test for smaller samples. A probability level equivalent to 0.05 was used to determine statistical significance. Results Animal Survival Table 1 details the survival results for animals in each treatment group using the log rank survival statistic. When compared to the control group (median survival, 22 days), a significant survival benefit was found after radiosurgery (median survival, 45 days; p ), radiosurgery plus WBRT (median, 45 days; p ), and WBRT alone (median survival, 40 days; p = ). Similarly, a survival benefit was found after 10-fraction radiotherapy to 85 Gy (median survival, 49 days; p ) and after single fraction hemibrain irradiation to 35 Gy (median survival, 38 days; p = 0.005). In the survival analysis, we could not identify a difference between the results after radiosurgery alone and those after radiosurgery plus WBRT (p = 0.72), between radiosurgery alone and WBRT alone (p = 0.12), or between radiosurgery plus WBRT and WBRT alone (p = 0.19). There was no difference between the biologically equivalent groups of radiosurgery and 10-fraction radiotherapy (p = 0.45) or between radiosurgery and single-fraction nonstereotactic irradiation at the same 35-Gy margin dose (p = 0.8). Similar statistical results were obtained using a two-tailed t-test. Kaplan Meier survival curves are shown in Fig. 1. J. Neurosurg. / Volume 84 / June, 1996 FIG. 2. Enlarged photographs showing whole-mount coronal sections through the implanted rat glioma. a: Control, 15-day survival. b: Radiosurgery alone, 90-day survival. c: Wholebrain irradiation (20 Gy in five fractions), 44-day survival. Tumor Size When compared with the control group (mean tumor diameter, 6.8 mm 1.3 mm), a significant reduction in tumor size was identified after radiosurgery alone (p = ) and after radiosurgery plus WBRT (p = 0.003) (Fig. 2). After radiosurgery, the mean tumor diameter was reduced to 5.2 mm 2.6 mm and after radiosurgery plus WBRT, 5.3 mm 2.3 mm. No significant reduction in tumor size was identified after 20-Gy WBRT alone (mean diameter 6.3 mm 1.2 mm) (p = 0.20). However, at the higher fractionated dose of 85 Gy, a significant reduction in tumor size was found (mean diameter 5.4 mm 1.6 mm) (p = 0.004), which was not different from the radiosurgery arm. Similarly, single-fraction irradiation to 35 Gy led to a reduction in tumor size compared with control (mean diameter 5.0 mm 1.0 mm) (p = 0.001). 1035

4 D. Kondziolka, et al. TABLE 2 Histological findings after radiation treatment of C6 glioma in rats* 5-Fraction 10-Fraction Single-Fraction Radiosurgery Radiosurgery WBRT Hemibrain RT Hemibrain RT Control 35 Gy + WBRT 20 Gy 85 Gy 35 Gy Parameter (54 animals) (22 animals) (13 animals) (18 animals) (16 animals) (10 animals) hypocellular tumor 0 (0%) 4 (18%) 4 (31%) 0 (0%) 0 (0%) 0 (0%) intratumoral hemorrhage 17 (32%) 8 (36%) 0 (0%) 4 (22%) 4 (25%) 0 (0%) tumor edema 7 (13%) 8 (36%) 5 (38%) 0 (0%) 4 (25%) 2 (20%) * RT = radiation therapy; WBRT = whole-brain radiation therapy. Histological Changes Table 2 details the individual histological changes identified among the different groups. A significant reduction in tumor cell density was identified after radiosurgery (p = 0.006) and after radiosurgery plus WBRT (p = 0.009) when compared with the control group. No significant reduction in cell density was identified after either 20- Gy WBRT alone, the 85-Gy 10-fraction regimen, or the 35-Gy single-fraction regimen. Between groups, radiosurgery plus WBRT was associated with reduced cell density when compared to any of the other nonradiosurgical groups (p = 0.02) (Fig. 3). We compared the presence of intratumoral hemorrhage (defined as greater than 10% of tumor volume) between groups. In the group that underwent radiosurgery plus WBRT, fewer animals experienced tumor hemorrhage than in the control group (p = 0.03). We found no other significant relations among the groups. We compared the presence or absence of cellular edema in each group after treatment. A significant increase in the number of animals with intratumoral edema was identified after radiosurgery (p = 0.03), and after radiosurgery plus WBRT (p = 0.05) when compared with control. No significant difference in the presence of edema was identified after WBRT alone (p = 0.18), 10-fraction radiotherapy (p = 0.45), or single-fraction hemibrain irradiation (p = 0.62) when compared with control. When compared with all other groups, both radiosurgery and radiosurgery plus WBRT were associated with a significant increase in the number of animals that had tumor edema. We found no significant interrelationships among cell density, the presence of hemorrhage, or the degree of tumor edema. Discussion Comparison of Radiosurgery and Fractionated Radiotherapy Most clinicians do not have a sound understanding of the magnitude of radiosurgical effects, especially when described in terms of conventional radiation therapy doses. One goal of this report was to begin to study the effects of radiosurgery and radiotherapy as distinct treatment modalities in the same model and to evaluate responses at doses that were biologically equivalent. Such a comparison is clinically relevant because there is an increasing use of stereotactic radiation therapy approaches for brain tumors (as an alternative to radiosurgery) despite limited knowledge regarding the number of fractions or dose necessary to be comparable to radiosurgery. The rat C6 glioma used in this study is not a human tumor, and the radiation response of the rat brain is not equivalent to that of the human brain. We realize this limitation of using the C6 glioma line, which some investigators believe behaves more like a metastatic tumor than an intrinsic glioma; however, we chose this cell line because of its utility in the rat model and its fairly uniform growth rate. 3 Despite these limitations, because there is increasing clinical interest not only in radiosurgery but in the use of small fraction numbers (less than 10) for stereotactic fractionated radiotherapy, we believe this experimental analysis is pertinent within the limitations of the model, which is not meant to duplicate the human setting. Although it is difficult to select a dose for each group that is radiobiologically equivalent, use of information provided by Larson, et al., 9 would be one starting point. In their report, for early-responding tissue such as a malignant neoplasm (alpha/beta = 10), a radiosurgery dose of 20 Gy was hypothesized to equal a fractionated dose of 50 Gy. Calculated alpha/beta ratios for different malignant glioma cell lines showed a mean value of ,18,20 Thus, we realized that an alpha/beta ratio of 10 was an assumption and may not exactly reflect the tumor studied. Using this model, 35-Gy radiosurgery would be biologically equivalent to using 85 Gy in 10 fractions. For the radiosurgery group, the average dose (between 35 and 70 Gy) was 52.5 Gy. The calculated 10-fraction biologically equivalent dose would be 138 Gy. We chose to study margin dose rather than average dose in this experiment, because that is the dose prescribed in clinical radiosurgery. Experimental Design Previous investigators showed that single-fraction, WBRT of an in vivo rat glioma led to improved animal survival compared to that of a control group. 1,4,7,16,17,19 We selected a maximum radiosurgical dose of 70 Gy (margin dose 35 Gy) because it represented a middle-range dose from our prior study, in which we had tested doses of 30, 40, 50, 70, and 100 Gy. Previously, we identified no specific dose response relationship and thus chose a single dose for this study to provide uniformity within the radiosurgery cohort. In our prior study, 7 we found no survival benefit to radiosurgery over control (p = 0.07), which perhaps was related to the small sample size examined. Schwachenwald and colleagues 15 found that single-fraction irradiation using the gamma unit at doses of 25 and 45 Gy was effective in limiting growth of the D-54 Mg glioma cell line in culture. Our selection of 35 Gy for the 1036 J. Neurosurg. / Volume 84 / June, 1996

5 Radiation strategies tested in rat glioma FIG. 3. Photomicrographs showing the tumor brain interface. a: Control; tumor is densely hypercellular. A similar appearance was found in rats that had fractionated radiation therapy. b: Radiosurgery; the entire tumor is small, necrotic, and appears to contain nonviable cells. c: Radiosurgery plus whole-brain radiotherapy; marked necrosis and a hypocellular appearance at the tumor margin is shown. H & E, original magnification 25. minimum dose received by the tumor was in accordance with their findings, although different cell lines were used. A WBRT regimen of 20 Gy administered in five fractions was selected as a dose within cerebral tolerance for the rat brain; this scheme is a clinically relevant one used by radiation oncologists. Rogers, et al., 13 found that early radiation-induced histological changes included trigeminal and facial nerve necrosis, whereas others found central and hemispheric white matter necrosis characterized as both early 5 and late effects. 13 Prior studies that investigated animal survival (with 9L tumors) at different WBRT doses showed that single-fraction doses above 21 Gy led to the deaths of all animals in 8 to 12 days. 4 In normal animals, a 10-fraction regimen of 21 Gy did not lead to animal death until at least Day 200. The authors suggested a conversion factor of 1.9 for whole-brain radiobiological equivalence between single-fraction and 10-fraction radiotherapy. 4 Because we hoped to evaluate animals for as long as 90 days, we wanted to avoid treatment-related morbidity within that period. Our previous normal brain studies of different radiosurgery doses showed no 90-day animal morbidity with doses as high as 200 Gy. 6 Radiosurgery in Addition to Radiation Therapy J. Neurosurg. / Volume 84 / June, 1996 Because most investigators agree that fractionated radiation is useful and beneficial in the clinical management of malignant gliomas, the possibility that radiosurgery alone might be of similar clinical benefit was not examined. However, a study of independent radiosurgical effects is important to evaluate its histological correlate and its potential clinical benefits when it might be used as an isolated treatment for recurrent tumors. At our institution, we currently use radiosurgery as a boost to fractionated radiotherapy (60 Gy) for the contrast-enhancing volume of anaplastic astrocytomas or glioblastomas multiforme less than 3 cm in diameter. Thus, the present animal study was relevant to the radiobiological evaluation of this clinical approach. Wheeler and Kaufman 19 identified an improvement in animal survival using a 26-Gy fractionated regimen in their 9L rat model. Barker and colleagues 1 found a survival benefit after 20-Gy single-fraction WBRT (administered Day 16 after implantation) in a similar model. Although those authors categorized animals without any evidence of tumor cells as cured, we excluded all such animals from our analysis in the belief that failure of initial tumor grafting was a serious, potentially confounding variable in such studies. Radiobiological Interpretation We detected no significant difference in median survival after radiosurgery, fractionated radiotherapy (wholebrain or hemibrain), or a combination of both. Thus, we did not define a dose response relationship for survival. This may be related to continued growth of persistently viable cells, and thus an inability to attain a mean survival longer than 50 days despite any treatment regimen. In addition, irradiating a larger volume, as used in all the nonradiosurgery groups, did not impact survival benefit. However, the cytotoxic effects between groups were different, as reflected by changes in tumor size, cellularity and edema. The exception to this was that radiosurgery and 85-Gy radiation therapy showed similar reductions in tumor size. It seems clear that the observed effects within this early 90-day period are primarily direct cellular responses, and not indirect vascular responses (most of which occur later), due to the absence of vessel thrombosis or infarction. Thus, the degree of decreased tumor cellularity was proportional to the observed edema (or expansion of extracellular spaces); these changes were noted only in the radiosurgery groups. This result was consistent with our prior work

6 D. Kondziolka, et al. Although we found more animals with tumoral edema in the control group than in the WBRT group, this was not statistically significant. One could hypothesize that WBRT led to a reduction in cytotoxic edema compared with control; however, further studies using brain weight measurements would be necessary to confirm this. Animals that underwent radiosurgery showed a decrease in tumor size that was not observed after the 5-fraction regimen or after the 35-Gy single-fraction irradiation. It is possible that a nonedematous tumor in the radiosurgery arms might have been smaller, indicating an even greater cytotoxic effect (a conclusion not reached according to the numbers of animals in this experiment). It would be logical to expect a greater decrease in tumor size or even less tumor cellularity after combined radiosurgery/radiotherapy than after the other therapies because more radiation was administered. However, in the biologically equivalent groups of 35-Gy radiosurgery and 10-fraction 85-Gy radiotherapy, even when compared to the single-fraction 35-Gy arm, radiosurgery led to greater cytotoxic effects, as noted by a greater reduction in cellular density. This was most likely due to the spectrum of dose delivered across the tumor in radiosurgery (35 Gy at the margin, increasing to 70 Gy at the center) versus a much more uniform dose delivered across the tumor in the other regimens. This finding is an example of the potential benefit of dose heterogeneity within solid neoplasms and may indicate not only the importance of the minimum dose, but also of isodose and the corresponding maximum dose. Such higher central doses, as well as a delayed vascular response from vessels irradiated at the tumor periphery, are probably responsible for the marked loss of central contrast enhancement often found after human acoustic tumor radiosurgery. 10,12 Fractionated radiotherapy to the 100% isodose line did not appear to be as efficacious as radiosurgery to the 50% isodose line. This finding might be clinically relevant to those now using stereotactic fractionated techniques, in which both the benefits of precise targeting as well as the use of less homogeneous doses (at or below the 80% isodose) are being explored. Future Experiments Because the goal of radiosurgery is tumor control rather than tumor elimination, we next plan to investigate the in vitro behavior of irradiated cells and the molecular correlates of tumor dormancy. First, irradiated tumor cells within the rat brain will be removed and regrown in culture. This will allow quantification of their in vitro growth rate (in comparison to control) as well as an analysis of tumor-regulated growth factor production after radiosurgery. Similar experiments are planned for our acoustic tumor xenograft model 9 to study such responses after benign tumor radiosurgery. Acknowledgments The authors would like to thank Wendy Fellows-Mayle for excellent technical support and Kim Tonet for manuscript preparation. References 1. Barker M, Deen DF, Baker DG: BCNU and x-ray therapy of intracerebral 9L rat tumors. Int J Radiat Oncol Biol Phys 5: , Benda P, Lightbody J, Sato G, et al: Differentiated rat glial cell strain in tissue culture. Science 161: , Cohen JD, Robins HI, Javid MJ: Radiosensitization of C6 glioma by thymidine and 41.8 C hyperthermia. J Neurosurg 72: , Henderson SD, Kimler BF, Morantz RA: Radiation therapy of 9L rat brain tumors. Int J Radiat Oncol Biol Phys 7: , Hornsey S, Morris CC, Myers R: The relationship between fractionation and total dose for x ray induced brain damage. Int J Radiat Oncol Biol Phys 7: , Kondziolka D, Lunsford LD, Claassen D, et al: Radiobiology of radiosurgery: Part I. The normal rat brain model. Neurosurgery 31: , Kondziolka D, Lunsford LD, Claassen D, et al: Radiobiology of radiosurgery: Part II. The rat C6 glioma model. Neurosurgery 31: , Kondziolka D, Somaza S, Flickinger JC, et al: Cerebral radioprotective effects of high-dose pentobarbital evaluated in an animal radiosurgery model. Neurol Res 16: , Larson DA, Flickinger JC, Loeffler JS: The radiobiology of radiosurgery. Int J Radiat Oncol Biol Phys 25: , Linskey ME, Martinez AJ, Kondziolka D, et al: The radiobiology of human acoustic schwannoma xenografts after stereotactic radiosurgery evaluated in the subrenal capsule of athymic mice. J Neurosurg 78: , Malaise EP, Fertil B, Chavaudra N, et al: Distribution of radiation sensitivities for human tumor cells of specific histological types: comparison of in vitro to in vivo data. Int J Radiat Oncol Biol Phys 12: , Pollock BE, Lunsford LD, Kondziolka D, et al: Outcome analysis of acoustic neuroma management: a comparison of microsurgery and stereotactic radiosurgery. Neurosurgery 36: , Rogers M, Myers R, Jenkinson T, et al: Histology of the irradiated rat brain in the first post-irradiation year. Br J Radiol 55: , San-Galli F, Vrignaud P, Robert J, et al: Assessment of the experimental model of transplanted C6 glioblastoma in Wistar rats. J Neurooncol 7: , Schwachenwald R, Engebråten O, Valen H, et al: A technique for studying single-dose radiation effects on glioma invasiveness in tissue culture a pilot study, in Steiner L (ed): Radiosurgery. Baseline and Trends. New York: Raven Press, 1992, pp Spence AM, Geraci JP: Combined cyclotron fast-neutron and BCNU therapy in a rat brain-tumor model. J Neurosurg 54: , Steinbok P, Mahaley MS Jr, U R, et al: Treatment of autochthonous rat brain tumors with fractionated radiotherapy. The effects of graded radiation doses and of combined therapy with BCNU or steroids. J Neurosurg 53:68 72, Stuschke M, Budach V, Sack H: Radioresponsiveness of human glioma, sarcoma, and breast cancer spheroids depends on tumor differentiation. Int J Radiat Oncol Biol Phys 27: , Wheeler KT, Kaufman K: Influence of fractionation schedules on the response of a rat brain tumor to therapy with BCNU and radiation. Int J Radiat Oncol Biol Phys 6: , Yang X, Darling JL, McMillan TJ, et al: Radiosensitivity, recovery and dose-rate effect in three human glioma cell lines. Radiother Oncol 19:49 56, 1990 Manuscript received September 14, Accepted in final form January 15, Dr. Kondziolka is supported by Grant No. 1 K08 NS01723 from the National Institutes of Health. Address reprint requests to: Douglas Kondziolka, M.D., Department of Neurological Surgery, Suite B-400, Presbyterian University Hospital, 200 Lothrop Street, Pittsburgh, Pennsylvania J. Neurosurg. / Volume 84 / June, 1996