Stereotactic Radiosurgery: Gliomas

Introduction

Stereotactic radiosurgery for the treatment of primary malignant gliomas is counterintuitive. Radiosurgery provides a high dose of ionizing radiation to a small, well-defined volume of tissue, whereas gliomas tend to be large, diffuse, and infiltrating. Nevertheless, standard therapeutic approaches in the management of gliomas yield discouraging results, with the majority of patients diagnosed with glioblastoma multiforme (GBM) suffering local recurrence and dying within a year of diagnosis, and with survival beyond 2 years being rare. Nearly 80% of GBM recurrences occur within 2 cm of the primary site after conventional therapies [1]. In this context, either by retarding or preventing recurrence, dose escalation in areas with greatest tumor cell density may offer significant benefit for the individual patient while sparing normal functioning brain tissue lying on the periphery.

Management of Primary Gliomas Using Stereotactic Radiation

Clinical experiences with focal boost techniques include trials of brachy-therapy, stereotactic radiosurgery, and fractionated proton radiotherapy, each designed to escalate dose to well-defined volumes within the target tumor tissue. Major series are summarized in Table 1.

Stereotactic Brachytherapy

During the 1980s, stereotactic brachytherapy yielded encouraging results as a treatment for selected patients with recurrent disease. These results led to the use of iodine-125 (I-125) brachytherapy as boost technique for primary glioblastoma, with early results suggesting a substantial increase in survival when compared with historical controls [2]. Larger series also demonstrated encouraging results, with patients receiving additional median boost doses between 50 and 60 Gy after surgical biopsy or resection and external beam irradiation of 60 Gy in 30 to 33 fractions (Table 1).

A large series from the University of California at San Francisco (UCSF) demonstrated a 3-year survival rate of 22% within median survival of 20 months for 106 patients with primary glioblastoma [3]. A subgroup of these patients was enrolled in the Northern California Oncology Group (NCOG) Study 6G-82-2. These patients received hydroxyurea during the external beam irradiation, and thereafter, adjuvant procarbazine, lomustine (CCNU), and vincristine chemotherapy (PCV) for 1 year [4]. Although 30 of the 64 GBM patients on this trial were excluded from receiving brachy-therapy, mostly because of intercurrent death or tumor progression, for all enrolled patients the median survival was 67 weeks, with nine patients alive after 2 years’ follow-up, and three alive after 3 years. Among the 34 GBM patients who did receive brachytherapy, the median survival was 88 weeks.

Similarly, at the Joint Center for Radiation Therapy, 56 patients were treated with surgery, limited field external beam radiotherapy to 60 Gy, and brachytherapy for an additional 50 Gy. These were compared to a set of 40 historical controls with similar clinical and radiologic characteristics [5]. Median survival for the brachytherapy group was 18 months, compared with 11 months for the historical control group.

One randomized trial addressed the value of I-125 brachytherapy as a boost for treatment for high-grade glioma. The Brain Tumor Cooperative Group Trial 87-01, published in abstract form, demonstrated an improvement in median survival for eligible patients who underwent stereotactic implant boost [6]. More than 250 patients (87% with the diagnosis of GBM) were randomized to receive 60 Gy brachytherapy or to be observed, after external beam irradiation and carmustine chemotherapy (BCNU). The median survival for those randomized to receive I-125 brachytherapy was 16 months versus 13 months for the control group, with the reoperation rates similar for the two groups (50% and 42%, respectively).

Table 1 Stereotactic Radiation Boost for Primary Glioblastoma Multiforme

Institution	Median minimum boost	Diameter size limit (cm)	Median boost volume (cm3)	Patient number	Median survival (months)	Survival at 2 years (%)	Reoperation rate (%)
Brachytherapy
UCSF (3)	52.9 Gy	< 6 cm	N/S	106	20	39%	38%
JCRT (5)	50.0 Gy	< 5 cm	22 cm3	56	18	34%	64%
BTCG (6)	60.0 Gy	N/S	N/S	125	16	N/S	50%
Sterotactic radiosurgery
JCRT, Wisconsin, and Florida (14)	12.0 Gy	< 4 cm	10.0 cm3	96	21	38%	29%
Pittsburgh (15)	15.5 Gy	< 3.5 cm	6.5 cm3	45	20	41%	19%
Harvard (16)	12 Gy	< 4 cm	9.4 cm3	78	20	36%	50%
Fractionated proton radiotherapy (10 fractions/ week for 5 weeks) MGH-HCL (19)	33.5 CGE (total dose 93.5 CGE)	< 5 cm	36 cm3	23	20	34%	57%

N/S, not stated.

Stereotactic Radiosurgery

The encouraging results of stereotactic brachytherapy, which enabled the escalation of radiation dose within a well-defined volume beyond the capabilities of conventional external-beam techniques, led to several centers applying techniques of stereotactic radiosurgery (SRS) to the management of patients presenting with GBM tumors. Stereotactic radiosurgery provided another radiation therapy technique for marked dose escalation while avoiding some of the potential risks of the brachytherapy implant procedure in patients with serious coexisting morbidities or tumors located in relatively inaccessible or eloquent regions of the brain. This early experience with SRS for GBM tumors suggested benefits comparable to those seen with stereotactic brachytherapy, with several GBM patients surviving beyond 2 years at rates higher than what would be expected for conventional therapies [7-9]. However, other data suggested the observed survival benefit accruing from SRS for glioma patients might be ascribable to selection factors, particularly a smaller total target volume for the SRS boost, rather than to improved outcomes after radiosurgery [10-12].

In the early 1990s, recursive partitioning analysis of several Radiation Therapy Oncology Group (RTOG) studies yielded detailed information about selection factors that influenced the prognosis for glioma patients (Table 2) [13]. Given the importance of these selection factors in the survival of conventionally treated patients—including age, performance status, and extent of resection—subsequent studies of the efficacy of SRS boost treatments controlled for these prognostic factors (Table 3) [14-16].

Using the RTOG analysis, one series from the Joint Center for Radiation Therapy (JCRT), the University of Wisconsin, and the University of Florida examined a combined group of 96 patients with GBM, along with an additional 19 patients with anaplastic astrocytoma, partitioned into prognostic classes III through VI [14]. As shown in Table 3, the relative risk of death for SRS-treated patients was about half that of the RTOG patients, for prognostic groups III through V.

Similarly, 65 patients at the University of Pittsburgh underwent SRS as part of their initial management plans for histologically proven anaplastic astrocytoma or glioblastoma [15]. The patients who were included had a contrast-enhancing tumor diameter size (the disease targeted for SRS boost) that was limited to less than 3.5 cm, although the study did include tumors in sensitive locations such as the diencephalon and brainstem. Like the JCRT/Wisconsin/Florida study, for patients with RTOG groups III, IV, and V, SRS yielded about a doubling of the survival rate at 2 years (Table 3). Although neither RTOG class nor extent of resection appeared predictive in this study, multivariate analysis did show age, Karnofsky score of 70 or higher, and histology to be important predictors of survival.

Table 2 RTOG Definitions of Prognostic Classes for Malignant Glioma Patients with Estimated Survival Rates Using Standard Therapies (13)

Prognostic class	GBM patients	Anaplastic astrocytoma patients	Survival at two years
I	N/A		76%
II	N/A		68%
III			35%
IV			15%
V			6%
VI			4%

RTOG, Radiation Therapy Oncology Group; KPS, Karnovsky Performance Status; GBM, glioblastoma multiforme.

Table 3 Comparison of 2-Year Survival Rates (%), Stratified by RTOG-Defined Risk Groups: III, IV, V, and VI

Series stratified by risk group	III	IV	V	VI
Historical controls—RTOG standard risk groups (13)	35% (175)	15% (457)	6% (395)	4% (263)
SRS—combined experience of Wisconsin, JCRT, and Gainesville (14)	75% (24)	34% (35)	21% (43)	(3 pts included in class V)
SRS—Univ. of Pittsburgh (15)	73% (13)	24% (11)	26% (24)	0% (2)
SRS—JCRT (16)	58% (27)	23% (29)	23% (22)	N/A
Fractionated treatment—MGH-HCL (19)	57% (7)	43% (7)	22% (9)	N/A

Numbers of patients in each group at time of diagnosis are shown in parentheses. RTOG, Radiation Therapy Oncology Group; SRS, Stereostatic Radiotherapy; JCRT, Joint Center for Radiation Therapy; MGH-HCL, Massachusetts General Hospital-Harvard Cyclotron Laboratory.

Shrieve and coworkers found similar results for 78 patients with GBM tumors treated with SRS boost after attempted surgical resection and standard postoperative radiation therapy [16]. Patients were eligible for SRS treatment if tumors measured no more than 4 cm in diameter with contrast enhancement, excluding edema. Seven patients showed no enhancing tumor on the postoperative imaging studies, in which cases the surgical cavity was treated with a 5-mm margin. Again, similar to the results from Pittsburgh, SRS showed a substantial improvement in 2-year survival rates, when compared with the RTOG historical control groups (Table 3).

Fractionated Stereotactic Radiotherapy

Although several studies have investigated alternative fractionation schemes as attempts to achieve radiobiological advantages with high doses in the treatment of glial neoplasms [17,18], only one study has investigated the use of fractionated radiotherapy to boost gliomas beyond more conventional dose levels to the higher levels that may be biologically comparable to stereotactic radiosurgery or brachytherapy. A phase I/II protocol at the Massachusetts General Hospital (MGH) studied 23 GBM patients treated with a mixture of photons and protons to doses above 90 cobalt gray equivalents (CGE), treating two fractions a day with a minimum 6-hour interfraction interval, for a total of 5 weeks [19]. Conventional volumes received a median dose of 64.8 CGE, and volumes considered at highest risk for harboring residual disease were boosted to a median total dose of 93.5 CGE. Although 10 patients showed no residual gadolinium enhancement after resection and another eight had residual enhancing volumes ranging from 0.1 to 1.0 cm3, across all patients the median volume receiving the high-dose boost was 36 cm3. This high-dose boost volume encompassed the remaining surgical cavity on the earliest postoperative imaging study, as well as any remaining gadolinium-enhancing tissue.

Stratifying by RTOG prognostic group, the results of this study were roughly comparable to the various SRS series (Table 3). The reoperation rate, however, was relatively high (13/23 = 57%). Of these 13 patients five underwent biopsies subsequent to radiation therapy, five underwent one resection, whereas three underwent multiple resections. Among the 15 patients with pathological material available for analysis subsequent to radiation therapy, all showed evidence of extensive tumor necrosis, but 60% also showed evidence of tumor persistence or recurrence. The median survival for patients showing only tumor necrosis was 29 months, as compared with 16 months for those also with pathological evidence of tumor recurrence (P = 0.01). In only one of the nine pathologically documented recurrences was tumor found within the 90-CGE volume, although among all 23 patients, 18 (78%) developed new enhancement on MR imaging within the high-dose target volume, suggesting that most of these imaging changes represented tissue necrosis rather than tumor recurrence.

Indications for SRS in the Treatment of Primary High-Grand Gliomas

In summary, in comparison with conventional radiation therapy, following maximal surgical debulking with SRS provides a survival benefit for appropriately selected patients. Major series demonstrate survival rates that appear almost double those achieved with conventional radiation therapy. This likely benefit of SRS has persuaded the neuro-oncology community to pursue a phase-III trial, RTOG 93-05, randomizing GBM patients with less than 4-cm tumors (all of whom receive BCNU chemotherapy) between (1) standard radiation therapy versus (2) radiosurgery followed by standard radiation therapy. To date, more than 250 patients have been randomized, and the results are pending.

Some commentators suggest that within each RTOG risk group, there may be further patient selection effects that account for the observed differences in survival rates, particularly for patients with residual contrast-enhancing tumor volumes that are small or non-existent. However, the results of the proton-dose escalation study argue against this objection. Among these 23 patients 10 (43%) had no postoperative gadolinium enhancement. Therefore, in terms of residual enhancing tumor, this study represents a subgroup more favorable than those patients in the study by Shrieve et al., in which only 7 of 78 patients had no postoperative gadolinium tumor enhancement [16], and more favorable than those in the study of Kondziolka et al., in which (the authors imply that) all GBM patients evinced some enhancement [15]. There was only one documented recurrence in the high-dose volumes treated with protons. These treatment volumes were substantially larger than the SRS median boost volumes in the two other studies, yet the survival benefits by RTOG risk classification were comparable to those achieved with SRS (Table 3). Thus, the clinical benefits that could be achieved with 90 CGE using fractionated therapy to larger volumes with smaller tumors— benefits that were confirmed on pathological review—were comparable to the clinical benefits achieved with SRS boosts to smaller volumes with larger tumors. Together, these points argue against the suggestion that the benefits to SRS boost treatment seen in Table 3 are ascribable only to selection effects attributable to target tumor volume.

For very large targets with postoperative enhancement beyond 3.5 to 4.0 cm in diameter, the likelihood of radiosurgical complications increases. For these cases, however, the MGH proton study demonstrates that fractionated conformal irradiation can be clinically delivered to larger volumes while maintaining high pathologic complete response rates in the high-dose volume. Further investigations are indicated in this regard, such as RTOG 98-03, a phase I/II radiation study investigating fractionated dose escalation from 66 Gy up to 82 Gy, applying conformal radiation technologies to the treatment of large supratentorial GBM tumors.

THE MANAGEMENT OF RECURRENT GLIOMAS

There are several treatment options for patients with recurrent malignant gliomas, including reoperation, radiosurgery, interstitial implantation, and chemotherapy [20]. Historically, chemotherapy has been the standard therapy for recurrent gliomas, yet results are discouraging. For example, Levin et al. reported a 55% response rate to combination chemotherapy for recurrent GBM disease with a median time to progression of 23 weeks [21]. Overall, tumor location, size, the patient’s overall condition, and the prior therapeutic history each plays a role in the choice of appropriate modality.

Surgery for Recurrent Gliomas

One series from UCSF showed that in younger patients with higher Kar-nofsky scores and with large, surgically accessible, recurrent tumors that caused deficits from compression rather than infiltration, reoperation contributed to high quality postoperative survival [Karnovsky Performance Status (KPS > 70) as well as to overall survival, with a median overall survival of 36 weeks after reoperation [22]. A more recent series confirmed a median survival of 36 weeks for patients selected for reoperation, along with a median high-quality survival period of 18 weeks, compared with total median survival of 23 weeks after first tumor progression for patients not undergoing reoperations [23]. Postoperative improvements in KPS scores (28% of patients) were slightly more likely than declines in KPS scores (23% of patients), with these improvements most likely in those patients who had symptomatic recurrences.

Stereotactic Radiosurgery for Recurrent Disease

Stereotactic radiosurgery also has a role in the management of recurrent glial tumors after standard therapeutic approaches. First presented in 1995, the results of RTOG 90-05 demonstrated that the incidence of severe central nervous system (CNS) toxicity in previously irradiated brain tissues subsequently treated with SRS for recurrence was a function of both the target volume and the prescribed dose. Nevertheless, the incidence of complications could be maintained at an acceptable, low level while providing clinically meaningful doses to malignant tissues. There were chronic, severe CNS toxicities in 14% of patients who had 3.1 to 4.0-cm tumors treated to 15.0 Gy, in 20% of patients who had 2.1 to 3.0-cm tumors treated to 18.0 Gy, and in 10% of patients who had up to 2.0-cm tumors treated to 24.0 Gy [24].

Similar to results from the reoperation series, studies of SRS for recurrent gliomas have shown median survivals after SRS treatment across all treated patients that range from 7 to 10 months [25-29]. In a series from Boston, comparison of patients undergoing SRS and those receiving stere-otactic brachytherapy suggested the two modalities had similar survival benefits [27]. For SRS, the median actuarial survival from time of treatment for recurrence was 10.2 months, whereas for brachytherapy the median actuarial survival after treatment was 11.5 months. Patients receiving SRS had somewhat smaller tumor volumes compared with brachytherapy (median 10.1 cm3 vs 29 cm3). Among the SRS patients, younger age and a tumor volume less than 10.1 cm3 were predictive of better outcome; however, for brachytherapy patients, patient age was predictive of outcome, whereas tumor volume, interval from initial diagnosis, and tumor dose were not. Of 86 patients treated with SRS, 19 (22%) required subsequent reoperation, whereas 14 of 32 patients (44%) required reoperation after brachytherapy; furthermore, the outcomes after SRS were independent of a need for reoperation. This comparison suggests that, for patients qualifying for SRS at time of recurrence, and particularly for younger patients with limited-volume tumor recurrences, SRS is the preferred therapeutic option. For larger tumors or irregularly shaped volumes, other modalities may be more appropriate.

Fractionated Stereotactic Radiation Therapy for Recurrent Gliomas

Some recent data suggest that fractionated stereotactic radiation therapy (SRT) may be of benefit for patients otherwise unsuitable for SRS. Cho and colleagues evaluated 71 patients with recurrent high-grade gliomas: 46 patients received single-fraction SRS (median 17 Gy to the 50% isodose surface), and 25 received fractionated SRT (37.5 Gy in 15 fractions to the 85% isodose surface) [30]. Patients in the SRS group had more favorable prognostic factors than those in the SRT group, including median age (48 vs 53 years), median KPS (70 vs 60), and median tumor volume (10 vs 25 cm3), but median survival times were comparable for the two groups: 11 months for the SRS group and 12 months for the SRT group. Late complications developed in 14 (30%) of the 46 SRS patients but in only 2 (8%) of the 25 SRT patients, suggesting the SRT dose-fractionation schemes were less toxic than the SRS plans.

There may be a role for chemotherapy, in combination with fractionated stereotactic radiation therapy (SRT), in the treatment of recurrent glial tumors. One pilot study treated 14 patients with recurrent glioblastoma that had median tumor volumes of 15.7 cm3, using fractionated stereotactic radiation therapy along with Taxol as a radiation sensitizer [31]. Taxol was given once per week for 4 weeks, with an SRT treatment delivered immediately after each Taxol infusion. The median radiation dose per week was 6.0 Gy at the 90% isodose line, for a median total dose of 24 Gy in four fractions. The median survival from time of treatment for recurrence was 14.2 months, but with a short minimum follow-up of 10 months. The fractionated radiation dose appeared well-tolerated, with only four patients undergoing reoperation. These data suggest that, for large volume recurrences not surgically accessible or amenable to SRS, there may be a role for fractionated SRT, perhaps in conjunction with systemic chemotherapy. Further studies are required.

CONCLUSION

Stereotactic radiosurgery is effective in the treatment of selected primary and recurrent glial neoplasms. After maximal tumor resection, in conjunction with a standard course of radiation therapy, SRS boost likely improves survival for patients in RTOG risk classes III, IV, and V. We anticipate RTOG 93-05 will confirm this survival benefit. For primary tumors with anatomically amenable, well-defined postoperative residual volumes less than 4 cm in diameter, SRS is the preferred radiation boost technique, whether using Linac radiosurgery, the Leksell Gamma Knife, or proton radiosurgery. For larger lesions, irregular volumes, or difficult anatomical constraints, other boost techniques may be considered, including brachytherapy, fractionated stereotactic irradiation, or proton radiotherapy. For focally recurrent GBM disease, patients with small (less than 3 cm in diameter), radiographically distinct lesions may benefit from SRS. Larger lesions, especially those adjacent to eloquent cortex or critical white matter pathways, must be evaluated with caution. Although SRS offers another tool in the treatment of high-grade gliomas, these tumors continue to present a serious therapeutic challenge, and overall results are still dismal. Further innovations in dose-delivery, targeting, and adjuvant treatments are required.